Merge pull request #11 from jepler/merge-upstream-4

Merge upstream 4
Merge remote-tracking branch 'upstream/master' into merge-upstream-4
2020-02-27 14:09:18 -06:00 · 2020-02-27 14:08:15 -06:00 · 2020-02-27 14:07:04 -06:00 · 2020-02-27 13:58:25 -06:00 · 2020-02-27 13:58:14 -06:00 · 2020-02-27 13:56:09 -06:00
37 changed files with 2307 additions and 9702 deletions
--- a/.github/workflows/build.yml
+++ b/.github/workflows/build.yml
@ -0,0 +1,58 @@
+name: Build CI
+
+on:
+  push:
+  pull_request:
+  release:
+    types: [published]
+  check_suite:
+    types: [rerequested]
+
+jobs:
+  test:
+    runs-on: ubuntu-16.04
+    steps:
+    - name: Dump GitHub context
+      env:
+        GITHUB_CONTEXT: ${{ toJson(github) }}
+      run: echo "$GITHUB_CONTEXT"
+    - name: Set up Python 3.5
+      uses: actions/setup-python@v1
+      with:
+        python-version: 3.5
+
+    - name: Versions
+      run: |
+        gcc --version
+        python3 --version
+    - name: Checkout ulab
+      uses: actions/checkout@v1
+
+    - name: Checkout micropython repo
+      uses: actions/checkout@v2
+      with:
+        repository: micropython/micropython
+        path: micropython
+
+    - name: Checkout micropython submodules
+      run: (cd micropython && git submodule update --init)
+
+    - name: Build mpy-cross
+      run: make -C micropython/mpy-cross -j2
+
+    - name: Build micropython unix port
+      run: |
+        make -C micropython/ports/unix -j2 deplibs
+        make -C micropython/ports/unix -j2 USER_C_MODULES=$(readlink -f .)
+
+    - name: Run tests
+      run: env MICROPYTHON_CPYTHON3=python3.5 MICROPY_MICROPYTHON=micropython/ports/unix/micropython micropython/tests/run-tests -d tests
+    - name: Print failure info
+      run: |
+        for exp in *.exp;
+        do testbase=$(basename $exp .exp);
+        echo -e "\nFAILURE $testbase";
+        diff -u $testbase.exp $testbase.out;
+        done
+      if: failure()
+
--- a/.gitignore
+++ b/.gitignore
@ -0,0 +1,3 @@
+/micropython
+/*.exp
+/*.out
--- a/README.md
+++ b/README.md
@ -1,62 +1,12 @@
-# micropython-ulab
+# circuitpython-ulab

 ulab is a numpy-like array manipulation library for micropython. 
 The module is written in C, defines compact containers for numerical 
-data, and is fast. 
+data, and is fast.

-Documentation can be found under https://micropython-ulab.readthedocs.io/en/latest/
-The source for the manual is in https://github.com/v923z/micropython-ulab/blob/master/docs/ulab-manual.ipynb,
-while developer help is in https://github.com/v923z/micropython-ulab/blob/master/docs/ulab.ipynb.
+ulab will be incorporated in builds of most CircuitPython supported
+devices, so there's usually no need to use the files here directly.
+If you've encountered a problem with circuitpython-ulab, please
+file an issue [in the circuitpython issue tracker](https://github.com/adafruit/circuitpython).

-# Firmware
-
-Firmware for pyboard.v.1.1, and PYBD_SF6 is updated once in a while, and can be downloaded 
-from https://github.com/v923z/micropython-ulab/releases.
-
-## Compiling
-
-If you want to try the latest version of `ulab`, or your hardware is 
-different to pyboard.v.1.1, or PYBD_SF6, the firmware can be compiled 
-from the source by following these steps:
-
-First, you have to clone the micropython repository by running 
-
-```
-git clone https://github.com/micropython/micropython.git
-```
-on the command line. This will create a new repository with the name `micropython`. Staying there, clone the `ulab` repository with 
-
-```
-git clone https://github.com/v923z/micropython-ulab.git ulab
-```
-
-Then you have to include `ulab` in the compilation process by editing `mpconfigport.h` of the directory of the port for which you want to compile, so, still on the command line, navigate to `micropython/ports/unix`, or `micropython/ports/stm32`, or whichever port is your favourite, and edit the `mpconfigport.h` file there. All you have to do is add a single line at the end: 
-
-```
-#define MODULE_ULAB_ENABLED (1)
-```
-
-This line will inform the compiler that you want `ulab` in the resulting firmware. If you don't have the cross-compiler installed, your might want to do that now, for instance on Linux by executing 
-
-```
-sudo apt-get install gcc-arm-none-eabi
-```
-If that was successful, you can try to run the make command in the port's directory as 
-```
-make BOARD=PYBV11 USER_C_MODULES=../../../ulab all
-```
-which will prepare the firmware for pyboard.v.11. Similarly, 
-```
-make BOARD=PYBD_SF6 USER_C_MODULES=../../../ulab all
-```
-will compile for the SF6 member of the PYBD series. Provided that you managed to compile the firmware, you would upload that by running
-either
-```
-dfu-util --alt 0 -D firmware.dfu
-```
-or 
-```
-python pydfu.py -u firmware.dfu
-```
-
-In case you got stuck somewhere in the process, a bit more detailed instructions can be found under https://github.com/micropython/micropython/wiki/Getting-Started, and https://github.com/micropython/micropython/wiki/Pyboard-Firmware-Update.
+circuitpython-ulab is based on [micropython-ulab](https://github.com/v923z/micropython-ulab).
--- a/build.sh
+++ b/build.sh
@ -0,0 +1,17 @@
+#!/bin/sh
+set -e
+HERE="$(dirname -- "$(readlink -f -- "${0}")" )"
+[ -e micropython/py/py.mk ] || git clone https://github.com/micropython/micropython
+[ -e micropython/lib/libffi/autogen.sh ] || (cd micropython && git submodule update --init lib/libffi )
+#git clone https://github.com/micropython/micropython
+make -C micropython/mpy-cross -j$(nproc)
+make -C micropython/ports/unix -j$(nproc) deplibs
+make -C micropython/ports/unix -j$(nproc) USER_C_MODULES="${HERE}"
+
+if ! env MICROPY_MICROPYTHON=micropython/ports/unix/micropython micropython/tests/run-tests -d tests; then
+    for exp in *.exp; do
+        testbase=$(basename $exp .exp);
+        echo -e "\nFAILURE $testbase";
+        diff -u $testbase.exp $testbase.out;
+    done
+fi
--- a/code/create.c
+++ b/code/create.c
@ -0,0 +1,155 @@
+/*
+ * This file is part of the micropython-ulab project,
+ *
+ * https://github.com/v923z/micropython-ulab
+ *
+ * The MIT License (MIT)
+ *
+ * Copyright (c) 2020 Jeff Epler for Adafruit Industries
+ * 				 2019-2020 Zoltán Vörös
+*/
+
+
+#include "py/obj.h"
+#include "py/runtime.h"
+#include "create.h"
+
+mp_obj_t create_zeros_ones(size_t n_args, const mp_obj_t *pos_args, mp_map_t *kw_args, uint8_t kind) {
+    static const mp_arg_t allowed_args[] = {
+        { MP_QSTR_, MP_ARG_REQUIRED | MP_ARG_OBJ, {.u_obj = MP_OBJ_NULL} } ,
+        { MP_QSTR_dtype, MP_ARG_KW_ONLY | MP_ARG_INT, {.u_int = NDARRAY_FLOAT} },
+    };
+
+    mp_arg_val_t args[MP_ARRAY_SIZE(allowed_args)];
+    mp_arg_parse_all(n_args, pos_args, kw_args, MP_ARRAY_SIZE(allowed_args), allowed_args, args);
+    
+    uint8_t dtype = args[1].u_int;
+    if(!MP_OBJ_IS_INT(args[0].u_obj) && !MP_OBJ_IS_TYPE(args[0].u_obj, &mp_type_tuple)) {
+        mp_raise_TypeError(translate("input argument must be an integer or a 2-tuple"));
+    }
+    ndarray_obj_t *ndarray = NULL;
+    if(MP_OBJ_IS_INT(args[0].u_obj)) {
+        size_t n = mp_obj_get_int(args[0].u_obj);
+        ndarray = create_new_ndarray(1, n, dtype);
+    } else if(MP_OBJ_IS_TYPE(args[0].u_obj, &mp_type_tuple)) {
+        mp_obj_tuple_t *tuple = MP_OBJ_TO_PTR(args[0].u_obj);
+        if(tuple->len != 2) {
+            mp_raise_TypeError(translate("input argument must be an integer or a 2-tuple"));
+        }
+        ndarray = create_new_ndarray(mp_obj_get_int(tuple->items[0]), 
+                                                  mp_obj_get_int(tuple->items[1]), dtype);
+    }
+    if(kind == 1) {
+        mp_obj_t one = mp_obj_new_int(1);
+        for(size_t i=0; i < ndarray->array->len; i++) {
+            mp_binary_set_val_array(dtype, ndarray->array->items, i, one);
+        }
+    }
+    return MP_OBJ_FROM_PTR(ndarray);
+}
+
+mp_obj_t create_zeros(size_t n_args, const mp_obj_t *pos_args, mp_map_t *kw_args) {
+    return create_zeros_ones(n_args, pos_args, kw_args, 0);
+}
+
+MP_DEFINE_CONST_FUN_OBJ_KW(create_zeros_obj, 0, create_zeros);
+
+mp_obj_t create_ones(size_t n_args, const mp_obj_t *pos_args, mp_map_t *kw_args) {
+    return create_zeros_ones(n_args, pos_args, kw_args, 1);
+}
+
+MP_DEFINE_CONST_FUN_OBJ_KW(create_ones_obj, 0, create_ones);
+
+mp_obj_t create_eye(size_t n_args, const mp_obj_t *pos_args, mp_map_t *kw_args) {
+    static const mp_arg_t allowed_args[] = {
+        { MP_QSTR_, MP_ARG_REQUIRED | MP_ARG_INT, {.u_int = 0} },
+        { MP_QSTR_M, MP_ARG_KW_ONLY | MP_ARG_OBJ, {.u_rom_obj = mp_const_none } },
+        { MP_QSTR_k, MP_ARG_KW_ONLY | MP_ARG_INT, {.u_int = 0} },        
+        { MP_QSTR_dtype, MP_ARG_KW_ONLY | MP_ARG_INT, {.u_int = NDARRAY_FLOAT} },
+    };
+
+    mp_arg_val_t args[MP_ARRAY_SIZE(allowed_args)];
+    mp_arg_parse_all(n_args, pos_args, kw_args, MP_ARRAY_SIZE(allowed_args), allowed_args, args);
+
+    size_t n = args[0].u_int, m;
+    int16_t k = args[2].u_int;
+    uint8_t dtype = args[3].u_int;
+    if(args[1].u_rom_obj == mp_const_none) {
+        m = n;
+    } else {
+        m = mp_obj_get_int(args[1].u_rom_obj);
+    }
+    
+    ndarray_obj_t *ndarray = create_new_ndarray(m, n, dtype);
+    mp_obj_t one = mp_obj_new_int(1);
+    size_t i = 0;
+    if((k >= 0) && (k < n)) {
+        while(k < n) {
+            mp_binary_set_val_array(dtype, ndarray->array->items, i*n+k, one);
+            k++;
+            i++;
+        }
+    } else if((k < 0) && (-k < m)) {
+        k = -k;
+        i = 0;
+        while(k < m) {
+            mp_binary_set_val_array(dtype, ndarray->array->items, k*n+i, one);
+            k++;
+            i++;
+        }
+    }
+    return MP_OBJ_FROM_PTR(ndarray);
+}
+
+MP_DEFINE_CONST_FUN_OBJ_KW(create_eye_obj, 0, create_eye);
+
+mp_obj_t create_linspace(size_t n_args, const mp_obj_t *pos_args, mp_map_t *kw_args) {
+    static const mp_arg_t allowed_args[] = {
+        { MP_QSTR_, MP_ARG_REQUIRED | MP_ARG_OBJ, {.u_rom_obj = mp_const_none } },
+        { MP_QSTR_, MP_ARG_REQUIRED | MP_ARG_OBJ, {.u_rom_obj = mp_const_none } },
+        { MP_QSTR_num, MP_ARG_INT, {.u_int = 50} },
+        { MP_QSTR_endpoint, MP_ARG_KW_ONLY | MP_ARG_OBJ, {.u_rom_obj = mp_const_true} },
+        { MP_QSTR_retstep, MP_ARG_KW_ONLY | MP_ARG_OBJ, {.u_rom_obj = mp_const_false} },
+        { MP_QSTR_dtype, MP_ARG_KW_ONLY | MP_ARG_INT, {.u_int = NDARRAY_FLOAT} },
+    };
+
+    mp_arg_val_t args[MP_ARRAY_SIZE(allowed_args)];
+    mp_arg_parse_all(2, pos_args, kw_args, MP_ARRAY_SIZE(allowed_args), allowed_args, args);
+
+    uint16_t len = args[2].u_int;
+    if(len < 2) {
+        mp_raise_ValueError(translate("number of points must be at least 2"));
+    }
+    mp_float_t value, step;
+    value = mp_obj_get_float(args[0].u_obj);
+    uint8_t typecode = args[5].u_int;
+    if(args[3].u_obj == mp_const_true) step = (mp_obj_get_float(args[1].u_obj)-value)/(len-1);
+    else step = (mp_obj_get_float(args[1].u_obj)-value)/len;
+    ndarray_obj_t *ndarray = create_new_ndarray(1, len, typecode);
+    if(typecode == NDARRAY_UINT8) {
+        uint8_t *array = (uint8_t *)ndarray->array->items;
+        for(size_t i=0; i < len; i++, value += step) array[i] = (uint8_t)value;
+    } else if(typecode == NDARRAY_INT8) {
+        int8_t *array = (int8_t *)ndarray->array->items;
+        for(size_t i=0; i < len; i++, value += step) array[i] = (int8_t)value;
+    } else if(typecode == NDARRAY_UINT16) {
+        uint16_t *array = (uint16_t *)ndarray->array->items;
+        for(size_t i=0; i < len; i++, value += step) array[i] = (uint16_t)value;
+    } else if(typecode == NDARRAY_INT16) {
+        int16_t *array = (int16_t *)ndarray->array->items;
+        for(size_t i=0; i < len; i++, value += step) array[i] = (int16_t)value;
+    } else {
+        mp_float_t *array = (mp_float_t *)ndarray->array->items;
+        for(size_t i=0; i < len; i++, value += step) array[i] = value;
+    }
+    if(args[4].u_obj == mp_const_false) {
+        return MP_OBJ_FROM_PTR(ndarray);
+    } else {
+        mp_obj_t tuple[2];
+        tuple[0] = ndarray;
+        tuple[1] = mp_obj_new_float(step);
+        return mp_obj_new_tuple(2, tuple);
+    }
+}
+
+MP_DEFINE_CONST_FUN_OBJ_KW(create_linspace_obj, 2, create_linspace);
--- a/code/create.h
+++ b/code/create.h
@ -0,0 +1,29 @@
+/*
+ * This file is part of the micropython-ulab project,
+ *
+ * https://github.com/v923z/micropython-ulab
+ *
+ * The MIT License (MIT)
+ *
+ * Copyright (c) 2020 Jeff Epler for Adafruit Industries
+ * 				 2019-2020 Zoltán Vörös
+*/
+
+#ifndef _CREATE_
+#define _CREATE_
+
+#include "ulab.h"
+#include "ndarray.h"
+
+/*
+mp_obj_t create_zeros(size_t , const mp_obj_t *, mp_map_t *);
+mp_obj_t create_ones(size_t , const mp_obj_t *, mp_map_t *);
+mp_obj_t create_eye(size_t , const mp_obj_t *, mp_map_t *);
+mp_obj_t create_linspace(size_t , const mp_obj_t *, mp_map_t *);
+*/
+MP_DECLARE_CONST_FUN_OBJ_KW(create_ones_obj);
+MP_DECLARE_CONST_FUN_OBJ_KW(create_zeros_obj);
+MP_DECLARE_CONST_FUN_OBJ_KW(create_eye_obj);
+MP_DECLARE_CONST_FUN_OBJ_KW(create_linspace_obj);
+
+#endif
--- a/code/extras.c
+++ b/code/extras.c
@ -0,0 +1,33 @@
+
+/*
+ * This file is part of the micropython-ulab project,
+ *
+ * https://github.com/v923z/micropython-ulab
+ *
+ * The MIT License (MIT)
+ *
+ * Copyright (c) 2020 Zoltán Vörös
+*/
+
+#include <math.h>
+#include <stdlib.h>
+#include <string.h>
+#include "py/obj.h"
+#include "py/runtime.h"
+#include "py/misc.h"
+#include "extras.h"
+
+#if ULAB_EXTRAS_MODULE
+
+STATIC const mp_rom_map_elem_t ulab_filter_globals_table[] = {
+    { MP_OBJ_NEW_QSTR(MP_QSTR___name__), MP_OBJ_NEW_QSTR(MP_QSTR_extras) },
+};
+
+STATIC MP_DEFINE_CONST_DICT(mp_module_ulab_extras_globals, ulab_extras_globals_table);
+
+mp_obj_module_t ulab_filter_module = {
+    .base = { &mp_type_module },
+    .globals = (mp_obj_dict_t*)&mp_module_ulab_extras_globals,
+};
+
+#endif
--- a/code/extras.h
+++ b/code/extras.h
@ -0,0 +1,23 @@
+
+/*
+ * This file is part of the micropython-ulab project,
+ *
+ * https://github.com/v923z/micropython-ulab
+ *
+ * The MIT License (MIT)
+ *
+ * Copyright (c) 2020 Zoltán Vörös
+*/
+
+#ifndef _EXTRA_
+#define _EXTRA_
+
+#include "ulab.h"
+#include "ndarray.h"
+
+#if ULAB_EXTRAS_MODULE
+
+mp_obj_module_t ulab_extras_module;
+
+#endif
+#endif
--- a/code/fft.c
+++ b/code/fft.c
@ -1,3 +1,4 @@
+
 /*
 * This file is part of the micropython-ulab project, 
 *
@ -5,20 +6,23 @@
 *
 * The MIT License (MIT)
 *
- * Copyright (c) 2019 Zoltán Vörös
+ * Copyright (c) 2019-2020 Zoltán Vörös
 */
-    
+
 #include <math.h>
 #include <stdio.h>
 #include <stdlib.h>
 #include <string.h>
 #include "py/runtime.h"
+#include "py/builtin.h"
 #include "py/binary.h"
 #include "py/obj.h"
 #include "py/objarray.h"
 #include "ndarray.h"
 #include "fft.h"

+#if ULAB_FFT_MODULE
+
 enum FFT_TYPE {
    FFT_FFT,
    FFT_IFFT,
@ -76,18 +80,18 @@ void fft_kernel(mp_float_t *real, mp_float_t *imag, int n, int isign) {

 mp_obj_t fft_fft_ifft_spectrum(size_t n_args, mp_obj_t arg_re, mp_obj_t arg_im, uint8_t type) {
    if(!MP_OBJ_IS_TYPE(arg_re, &ulab_ndarray_type)) {
-        mp_raise_NotImplementedError("FFT is defined for ndarrays only");
+        mp_raise_NotImplementedError(translate("FFT is defined for ndarrays only"));
    } 
    if(n_args == 2) {
        if(!MP_OBJ_IS_TYPE(arg_im, &ulab_ndarray_type)) {
-            mp_raise_NotImplementedError("FFT is defined for ndarrays only");
+            mp_raise_NotImplementedError(translate("FFT is defined for ndarrays only"));
        }
    }
    // Check if input is of length of power of 2
    ndarray_obj_t *re = MP_OBJ_TO_PTR(arg_re);
    uint16_t len = re->array->len;
    if((len & (len-1)) != 0) {
-        mp_raise_ValueError("input array length must be power of 2");
+        mp_raise_ValueError(translate("input array length must be power of 2"));
    }
    
    ndarray_obj_t *out_re = create_new_ndarray(1, len, NDARRAY_FLOAT);
@ -99,8 +103,9 @@ mp_obj_t fft_fft_ifft_spectrum(size_t n_args, mp_obj_t arg_re, mp_obj_t arg_im,
        memcpy((mp_float_t *)out_re->array->items, (mp_float_t *)re->array->items, re->bytes);
    } else {
        for(size_t i=0; i < len; i++) {
-            data_re[i] = ndarray_get_float_value(re->array->items, re->array->typecode, i);
+            *data_re++ = ndarray_get_float_value(re->array->items, re->array->typecode, i);
        }
+        data_re -= len;
    }
    ndarray_obj_t *out_im = create_new_ndarray(1, len, NDARRAY_FLOAT);
    mp_float_t *data_im = (mp_float_t *)out_im->array->items;
@ -108,29 +113,33 @@ mp_obj_t fft_fft_ifft_spectrum(size_t n_args, mp_obj_t arg_re, mp_obj_t arg_im,
    if(n_args == 2) {
        ndarray_obj_t *im = MP_OBJ_TO_PTR(arg_im);
        if (re->array->len != im->array->len) {
-            mp_raise_ValueError("real and imaginary parts must be of equal length");
+            mp_raise_ValueError(translate("real and imaginary parts must be of equal length"));
        }
        if(im->array->typecode == NDARRAY_FLOAT) {
            memcpy((mp_float_t *)out_im->array->items, (mp_float_t *)im->array->items, im->bytes);
        } else {
            for(size_t i=0; i < len; i++) {
-                data_im[i] = ndarray_get_float_value(im->array->items, im->array->typecode, i);
+               *data_im++ = ndarray_get_float_value(im->array->items, im->array->typecode, i);
            }
+            data_im -= len;
        }
    }
+
    if((type == FFT_FFT) || (type == FFT_SPECTRUM)) {
        fft_kernel(data_re, data_im, len, 1);
        if(type == FFT_SPECTRUM) {
            for(size_t i=0; i < len; i++) {
-                data_re[i] = MICROPY_FLOAT_C_FUN(sqrt)(data_re[i]*data_re[i] + data_im[i]*data_im[i]);
+                *data_re = MICROPY_FLOAT_C_FUN(sqrt)(*data_re * *data_re + *data_im * *data_im);
+                data_re++;
+                data_im++;
            }
        }
    } else { // inverse transform
        fft_kernel(data_re, data_im, len, -1);
        // TODO: numpy accepts the norm keyword argument
        for(size_t i=0; i < len; i++) {
-            data_re[i] /= len;
-            data_im[i] /= len;
+            *data_re++ /= len;
+            *data_im++ /= len;
        }
    }
    if(type == FFT_SPECTRUM) {
@ -151,6 +160,8 @@ mp_obj_t fft_fft(size_t n_args, const mp_obj_t *args) {
    }
 }

+MP_DEFINE_CONST_FUN_OBJ_VAR_BETWEEN(fft_fft_obj, 1, 2, fft_fft);
+
 mp_obj_t fft_ifft(size_t n_args, const mp_obj_t *args) {
    if(n_args == 2) {
        return fft_fft_ifft_spectrum(n_args, args[0], args[1], FFT_IFFT);
@ -159,6 +170,8 @@ mp_obj_t fft_ifft(size_t n_args, const mp_obj_t *args) {
    }
 }

+MP_DEFINE_CONST_FUN_OBJ_VAR_BETWEEN(fft_ifft_obj, 1, 2, fft_ifft);
+
 mp_obj_t fft_spectrum(size_t n_args, const mp_obj_t *args) {
    if(n_args == 2) {
        return fft_fft_ifft_spectrum(n_args, args[0], args[1], FFT_SPECTRUM);
@ -166,3 +179,21 @@ mp_obj_t fft_spectrum(size_t n_args, const mp_obj_t *args) {
        return fft_fft_ifft_spectrum(n_args, args[0], mp_const_none, FFT_SPECTRUM);
    }
 }
+
+MP_DEFINE_CONST_FUN_OBJ_VAR_BETWEEN(fft_spectrum_obj, 1, 2, fft_spectrum);
+
+STATIC const mp_rom_map_elem_t ulab_fft_globals_table[] = {
+    { MP_OBJ_NEW_QSTR(MP_QSTR___name__), MP_OBJ_NEW_QSTR(MP_QSTR_fft) },
+    { MP_OBJ_NEW_QSTR(MP_QSTR_fft), (mp_obj_t)&fft_fft_obj },
+    { MP_OBJ_NEW_QSTR(MP_QSTR_ifft), (mp_obj_t)&fft_ifft_obj },
+    { MP_OBJ_NEW_QSTR(MP_QSTR_spectrum), (mp_obj_t)&fft_spectrum_obj },
+};
+
+STATIC MP_DEFINE_CONST_DICT(mp_module_ulab_fft_globals, ulab_fft_globals_table);
+
+mp_obj_module_t ulab_fft_module = {
+    .base = { &mp_type_module },
+    .globals = (mp_obj_dict_t*)&mp_module_ulab_fft_globals,
+};
+
+#endif
--- a/code/fft.h
+++ b/code/fft.h
@ -1,3 +1,4 @@
+
 /*
 * This file is part of the micropython-ulab project, 
 *
@ -5,11 +6,12 @@
 *
 * The MIT License (MIT)
 *
- * Copyright (c) 2019 Zoltán Vörös
+ * Copyright (c) 2019-2020 Zoltán Vörös
 */
-    
+
 #ifndef _FFT_
 #define _FFT_
+#include "ulab.h"

 #ifndef MP_PI
 #define MP_PI MICROPY_FLOAT_CONST(3.14159265358979323846)
@ -17,7 +19,13 @@

 #define SWAP(t, a, b) { t tmp = a; a = b; b = tmp; }

-mp_obj_t fft_fft(size_t , const mp_obj_t *);
-mp_obj_t fft_ifft(size_t , const mp_obj_t *);
-mp_obj_t fft_spectrum(size_t , const mp_obj_t *);
+#if ULAB_FFT_MODULE
+
+extern mp_obj_module_t ulab_fft_module;
+
+MP_DECLARE_CONST_FUN_OBJ_VAR_BETWEEN(fft_fft_obj);
+MP_DECLARE_CONST_FUN_OBJ_VAR_BETWEEN(fft_ifft_obj);
+MP_DECLARE_CONST_FUN_OBJ_VAR_BETWEEN(fft_spectrum_obj);
+
+#endif
 #endif
--- a/code/filter.c
+++ b/code/filter.c
@ -0,0 +1,99 @@
+
+/*
+ * This file is part of the micropython-ulab project,
+ *
+ * https://github.com/v923z/micropython-ulab
+ *
+ * The MIT License (MIT)
+ *
+ * Copyright (c) 2020 Jeff Epler for Adafruit Industries
+*/
+
+#include <math.h>
+#include <stdlib.h>
+#include <string.h>
+#include "py/obj.h"
+#include "py/runtime.h"
+#include "py/misc.h"
+#include "filter.h"
+
+#if ULAB_FILTER_MODULE
+mp_obj_t filter_convolve(size_t n_args, const mp_obj_t *pos_args, mp_map_t *kw_args) {
+    static const mp_arg_t allowed_args[] = {
+        { MP_QSTR_a, MP_ARG_REQUIRED | MP_ARG_OBJ, {.u_rom_obj = mp_const_none } },
+        { MP_QSTR_v, MP_ARG_REQUIRED | MP_ARG_OBJ, {.u_rom_obj = mp_const_none } },
+    };
+
+    mp_arg_val_t args[MP_ARRAY_SIZE(allowed_args)];
+    mp_arg_parse_all(2, pos_args, kw_args, MP_ARRAY_SIZE(allowed_args), allowed_args, args);
+
+    if(!MP_OBJ_IS_TYPE(args[0].u_obj, &ulab_ndarray_type) || !MP_OBJ_IS_TYPE(args[1].u_obj, &ulab_ndarray_type)) {
+        mp_raise_TypeError(translate("convolve arguments must be ndarrays"));
+    }
+
+    ndarray_obj_t *a = MP_OBJ_TO_PTR(args[0].u_obj);
+    ndarray_obj_t *c = MP_OBJ_TO_PTR(args[1].u_obj);
+    int len_a = a->array->len;
+    int len_c = c->array->len;
+    // deal with linear arrays only
+    if(a->m*a->n != len_a || c->m*c->n != len_c) {
+        mp_raise_TypeError(translate("convolve arguments must be linear arrays"));
+    }
+    if(len_a == 0 || len_c == 0) {
+        mp_raise_TypeError(translate("convolve arguments must not be empty"));
+    }
+
+    int len = len_a + len_c - 1; // convolve mode "full"
+    ndarray_obj_t *out = create_new_ndarray(1, len, NDARRAY_FLOAT);
+    mp_float_t *outptr = out->array->items;
+    int off = len_c-1;
+
+    if(a->array->typecode == NDARRAY_FLOAT && c->array->typecode == NDARRAY_FLOAT) {
+        mp_float_t* a_items = (mp_float_t*)a->array->items;
+        mp_float_t* c_items = (mp_float_t*)c->array->items;
+        for(int k=-off; k<len-off; k++) {
+            mp_float_t accum = (mp_float_t)0;
+            int top_n = MIN(len_c, len_a - k);
+            int bot_n = MAX(-k, 0);
+            mp_float_t* a_ptr = a_items + bot_n + k;
+            mp_float_t* a_end = a_ptr + (top_n - bot_n);
+            mp_float_t* c_ptr = c_items + len_c - bot_n - 1;
+            for(; a_ptr != a_end;) {
+                accum += *a_ptr++ * *c_ptr--;
+            }
+            *outptr++ = accum;
+        }
+    } else {
+        for(int k=-off; k<len-off; k++) {
+            mp_float_t accum = (mp_float_t)0;
+            int top_n = MIN(len_c, len_a - k);
+            int bot_n = MAX(-k, 0);
+            for(int n=bot_n; n<top_n; n++) {
+                int idx_c = len_c - n - 1;
+                int idx_a = n+k;
+                mp_float_t ai = ndarray_get_float_value(a->array->items, a->array->typecode, idx_a);
+                mp_float_t ci = ndarray_get_float_value(c->array->items, c->array->typecode, idx_c);
+                accum += ai * ci;
+            }
+            *outptr++ = accum;
+        }
+    }
+
+    return out;
+}
+
+MP_DEFINE_CONST_FUN_OBJ_KW(filter_convolve_obj, 2, filter_convolve);
+
+STATIC const mp_rom_map_elem_t ulab_filter_globals_table[] = {
+    { MP_OBJ_NEW_QSTR(MP_QSTR___name__), MP_OBJ_NEW_QSTR(MP_QSTR_filter) },
+    { MP_OBJ_NEW_QSTR(MP_QSTR_convolve), (mp_obj_t)&filter_convolve_obj },
+};
+
+STATIC MP_DEFINE_CONST_DICT(mp_module_ulab_filter_globals, ulab_filter_globals_table);
+
+mp_obj_module_t ulab_filter_module = {
+    .base = { &mp_type_module },
+    .globals = (mp_obj_dict_t*)&mp_module_ulab_filter_globals,
+};
+
+#endif
--- a/code/filter.h
+++ b/code/filter.h
@ -0,0 +1,25 @@
+
+/*
+ * This file is part of the micropython-ulab project,
+ *
+ * https://github.com/v923z/micropython-ulab
+ *
+ * The MIT License (MIT)
+ *
+ * Copyright (c) 2020 Jeff Epler for Adafruit Industries
+*/
+
+#ifndef _FILTER_
+#define _FILTER_
+
+#include "ulab.h"
+#include "ndarray.h"
+
+#if ULAB_FILTER_MODULE
+
+extern mp_obj_module_t ulab_filter_module;
+
+MP_DECLARE_CONST_FUN_OBJ_KW(filter_convolve_obj);
+
+#endif
+#endif
--- a/code/linalg.c
+++ b/code/linalg.c
@ -1,3 +1,4 @@
+
 /*
 * This file is part of the micropython-ulab project, 
 *
@ -5,9 +6,9 @@
 *
 * The MIT License (MIT)
 *
- * Copyright (c) 2019 Zoltán Vörös
+ * Copyright (c) 2019-2020 Zoltán Vörös
 */
-    
+
 #include <stdlib.h>
 #include <string.h>
 #include <math.h>
@ -16,76 +17,24 @@
 #include "py/misc.h"
 #include "linalg.h"

-mp_obj_t linalg_transpose(mp_obj_t self_in) {
-    ndarray_obj_t *self = MP_OBJ_TO_PTR(self_in);
-    // the size of a single item in the array
-    uint8_t _sizeof = mp_binary_get_size('@', self->array->typecode, NULL);
-    
-    // NOTE: 
-    // if the matrices are square, we can simply swap items, but 
-    // generic matrices can't be transposed in place, so we have to 
-    // declare a temporary variable
-    
-    // NOTE: 
-    //  In the old matrix, the coordinate (m, n) is m*self->n + n
-    //  We have to assign this to the coordinate (n, m) in the new 
-    //  matrix, i.e., to n*self->m + m (since the new matrix has self->m columns)
-    
-    // one-dimensional arrays can be transposed by simply swapping the dimensions
-    if((self->m != 1) && (self->n != 1)) {
-        uint8_t *c = (uint8_t *)self->array->items;
-        // self->bytes is the size of the bytearray, irrespective of the typecode
-        uint8_t *tmp = m_new(uint8_t, self->bytes);
-        for(size_t m=0; m < self->m; m++) {
-            for(size_t n=0; n < self->n; n++) {
-                memcpy(tmp+_sizeof*(n*self->m + m), c+_sizeof*(m*self->n + n), _sizeof);
-            }
-        }
-        memcpy(self->array->items, tmp, self->bytes);
-        m_del(uint8_t, tmp, self->bytes);
-    } 
-    SWAP(size_t, self->m, self->n);
-    return mp_const_none;
-}
-
-mp_obj_t linalg_reshape(mp_obj_t self_in, mp_obj_t shape) {
-    ndarray_obj_t *self = MP_OBJ_TO_PTR(self_in);
-    if(!MP_OBJ_IS_TYPE(shape, &mp_type_tuple) || (MP_OBJ_SMALL_INT_VALUE(mp_obj_len_maybe(shape)) != 2)) {
-        mp_raise_ValueError("shape must be a 2-tuple");
-    }
-
-    mp_obj_iter_buf_t iter_buf;
-    mp_obj_t item, iterable = mp_getiter(shape, &iter_buf);
-    uint16_t m, n;
-    item = mp_iternext(iterable);
-    m = mp_obj_get_int(item);
-    item = mp_iternext(iterable);
-    n = mp_obj_get_int(item);
-    if(m*n != self->m*self->n) {
-        // TODO: the proper error message would be "cannot reshape array of size %d into shape (%d, %d)"
-        mp_raise_ValueError("cannot reshape array (incompatible input/output shape)");
-    }
-    self->m = m;
-    self->n = n;
-    return MP_OBJ_FROM_PTR(self);
-}
+#if ULAB_LINALG_MODULE

 mp_obj_t linalg_size(size_t n_args, const mp_obj_t *pos_args, mp_map_t *kw_args) {
    static const mp_arg_t allowed_args[] = {
-        { MP_QSTR_, MP_ARG_REQUIRED | MP_ARG_OBJ, {.u_rom_obj = MP_ROM_PTR(&mp_const_none_obj) } },
-        { MP_QSTR_axis, MP_ARG_KW_ONLY | MP_ARG_OBJ, {.u_rom_obj = MP_ROM_PTR(&mp_const_none_obj)} },
+        { MP_QSTR_, MP_ARG_REQUIRED | MP_ARG_OBJ, {.u_rom_obj = mp_const_none } },
+        { MP_QSTR_axis, MP_ARG_KW_ONLY | MP_ARG_OBJ, {.u_rom_obj = mp_const_none } },
    };

    mp_arg_val_t args[MP_ARRAY_SIZE(allowed_args)];
    mp_arg_parse_all(1, pos_args, kw_args, MP_ARRAY_SIZE(allowed_args), allowed_args, args);

-    if(!mp_obj_is_type(args[0].u_obj, &ulab_ndarray_type)) {
-        mp_raise_TypeError("size is defined for ndarrays only");
+    if(!MP_OBJ_IS_TYPE(args[0].u_obj, &ulab_ndarray_type)) {
+        mp_raise_TypeError(translate("size is defined for ndarrays only"));
    } else {
        ndarray_obj_t *ndarray = MP_OBJ_TO_PTR(args[0].u_obj);
        if(args[1].u_obj == mp_const_none) {
            return mp_obj_new_int(ndarray->array->len);
-        } else if(mp_obj_is_int(args[1].u_obj)) {
+        } else if(MP_OBJ_IS_INT(args[1].u_obj)) {
            uint8_t ax = mp_obj_get_int(args[1].u_obj);
            if(ax == 0) {
                if(ndarray->m == 1) {
@ -95,19 +44,21 @@ mp_obj_t linalg_size(size_t n_args, const mp_obj_t *pos_args, mp_map_t *kw_args)
                }
            } else if(ax == 1) {
                if(ndarray->m == 1) {
-                    mp_raise_ValueError("tuple index out of range");
+                    mp_raise_ValueError(translate("tuple index out of range"));
                } else {
                    return mp_obj_new_int(ndarray->n);
                }
            } else {
-                    mp_raise_ValueError("tuple index out of range");                
+                    mp_raise_ValueError(translate("tuple index out of range"));
            }
        } else {
-            mp_raise_TypeError("wrong argument type");
+            mp_raise_TypeError(translate("wrong argument type"));
        }
    }
 }

+MP_DEFINE_CONST_FUN_OBJ_KW(linalg_size_obj, 1, linalg_size);
+
 bool linalg_invert_matrix(mp_float_t *data, size_t N) {
    // returns true, of the inversion was successful, 
    // false, if the matrix is singular
@ -153,14 +104,14 @@ bool linalg_invert_matrix(mp_float_t *data, size_t N) {
 mp_obj_t linalg_inv(mp_obj_t o_in) {
    // since inv is not a class method, we have to inspect the input argument first
    if(!MP_OBJ_IS_TYPE(o_in, &ulab_ndarray_type)) {
-        mp_raise_TypeError("only ndarrays can be inverted");
+        mp_raise_TypeError(translate("only ndarrays can be inverted"));
    }
    ndarray_obj_t *o = MP_OBJ_TO_PTR(o_in);
    if(!MP_OBJ_IS_TYPE(o_in, &ulab_ndarray_type)) {
-        mp_raise_TypeError("only ndarray objects can be inverted");
+        mp_raise_TypeError(translate("only ndarray objects can be inverted"));
    }
    if(o->m != o->n) {
-        mp_raise_ValueError("only square matrices can be inverted");
+        mp_raise_ValueError(translate("only square matrices can be inverted"));
    }
    ndarray_obj_t *inverted = create_new_ndarray(o->m, o->n, NDARRAY_FLOAT);
    mp_float_t *data = (mp_float_t *)inverted->array->items;
@ -178,17 +129,22 @@ mp_obj_t linalg_inv(mp_obj_t o_in) {
        // TODO: I am not sure this is needed here. Otherwise, 
        // how should we free up the unused RAM of inverted?
        m_del(mp_float_t, inverted->array->items, o->n*o->n);
-        mp_raise_ValueError("input matrix is singular");
+        mp_raise_ValueError(translate("input matrix is singular"));
    }
    return MP_OBJ_FROM_PTR(inverted);
 }

+MP_DEFINE_CONST_FUN_OBJ_1(linalg_inv_obj, linalg_inv);
+
 mp_obj_t linalg_dot(mp_obj_t _m1, mp_obj_t _m2) {
    // TODO: should the results be upcast?
+    if(!MP_OBJ_IS_TYPE(_m1, &ulab_ndarray_type) || !MP_OBJ_IS_TYPE(_m2, &ulab_ndarray_type)) {
+        mp_raise_TypeError(translate("arguments must be ndarrays"));
+    }
    ndarray_obj_t *m1 = MP_OBJ_TO_PTR(_m1);
    ndarray_obj_t *m2 = MP_OBJ_TO_PTR(_m2);    
    if(m1->n != m2->m) {
-        mp_raise_ValueError("matrix dimensions do not match");
+        mp_raise_ValueError(translate("matrix dimensions do not match"));
    }
    // TODO: numpy uses upcasting here
    ndarray_obj_t *out = create_new_ndarray(m1->m, m2->n, NDARRAY_FLOAT);
@ -203,102 +159,21 @@ mp_obj_t linalg_dot(mp_obj_t _m1, mp_obj_t _m2) {
                v2 = ndarray_get_float_value(m2->array->items, m2->array->typecode, k*m2->n+j);
                sum += v1 * v2;
            }
-            outdata[i*m1->m+j] = sum;
+            outdata[j*m1->m+i] = sum;
        }
    }
    return MP_OBJ_FROM_PTR(out);
 }

-mp_obj_t linalg_zeros_ones(size_t n_args, const mp_obj_t *pos_args, mp_map_t *kw_args, uint8_t kind) {
-    static const mp_arg_t allowed_args[] = {
-        { MP_QSTR_, MP_ARG_REQUIRED | MP_ARG_OBJ, {.u_obj = MP_OBJ_NULL} } ,
-        { MP_QSTR_dtype, MP_ARG_KW_ONLY | MP_ARG_INT, {.u_int = NDARRAY_FLOAT} },
-    };
-
-    mp_arg_val_t args[MP_ARRAY_SIZE(allowed_args)];
-    mp_arg_parse_all(n_args, pos_args, kw_args, MP_ARRAY_SIZE(allowed_args), allowed_args, args);
-    
-    uint8_t dtype = args[1].u_int;
-    if(!mp_obj_is_int(args[0].u_obj) && !mp_obj_is_type(args[0].u_obj, &mp_type_tuple)) {
-        mp_raise_TypeError("input argument must be an integer or a 2-tuple");
-    }
-    ndarray_obj_t *ndarray = NULL;
-    if(mp_obj_is_int(args[0].u_obj)) {
-        size_t n = mp_obj_get_int(args[0].u_obj);
-        ndarray = create_new_ndarray(1, n, dtype);
-    } else if(mp_obj_is_type(args[0].u_obj, &mp_type_tuple)) {
-        mp_obj_tuple_t *tuple = MP_OBJ_TO_PTR(args[0].u_obj);
-        if(tuple->len != 2) {
-            mp_raise_TypeError("input argument must be an integer or a 2-tuple");            
-        }
-        ndarray = create_new_ndarray(mp_obj_get_int(tuple->items[0]), 
-                                                  mp_obj_get_int(tuple->items[1]), dtype);
-    }
-    if(kind == 1) {
-        mp_obj_t one = mp_obj_new_int(1);
-        for(size_t i=0; i < ndarray->array->len; i++) {
-            mp_binary_set_val_array(dtype, ndarray->array->items, i, one);
-        }
-    }
-    return MP_OBJ_FROM_PTR(ndarray);
-}
-
-mp_obj_t linalg_zeros(size_t n_args, const mp_obj_t *pos_args, mp_map_t *kw_args) {
-    return linalg_zeros_ones(n_args, pos_args, kw_args, 0);
-}
-
-mp_obj_t linalg_ones(size_t n_args, const mp_obj_t *pos_args, mp_map_t *kw_args) {
-    return linalg_zeros_ones(n_args, pos_args, kw_args, 1);
-}
-
-mp_obj_t linalg_eye(size_t n_args, const mp_obj_t *pos_args, mp_map_t *kw_args) {
-    static const mp_arg_t allowed_args[] = {
-        { MP_QSTR_, MP_ARG_REQUIRED | MP_ARG_INT, {.u_int = 0} },
-        { MP_QSTR_M, MP_ARG_KW_ONLY | MP_ARG_OBJ, {.u_rom_obj = MP_ROM_PTR(&mp_const_none_obj) } },
-        { MP_QSTR_k, MP_ARG_KW_ONLY | MP_ARG_INT, {.u_int = 0} },        
-        { MP_QSTR_dtype, MP_ARG_KW_ONLY | MP_ARG_INT, {.u_int = NDARRAY_FLOAT} },
-    };
-
-    mp_arg_val_t args[MP_ARRAY_SIZE(allowed_args)];
-    mp_arg_parse_all(n_args, pos_args, kw_args, MP_ARRAY_SIZE(allowed_args), allowed_args, args);
-
-    size_t n = args[0].u_int, m;
-    int16_t k = args[2].u_int;
-    uint8_t dtype = args[3].u_int;
-    if(args[1].u_rom_obj == mp_const_none) {
-        m = n;
-    } else {
-        m = mp_obj_get_int(args[1].u_rom_obj);
-    }
-    
-    ndarray_obj_t *ndarray = create_new_ndarray(m, n, dtype);
-    mp_obj_t one = mp_obj_new_int(1);
-    size_t i = 0;
-    if((k >= 0) && (k < n)) {
-        while(k < n) {
-            mp_binary_set_val_array(dtype, ndarray->array->items, i*n+k, one);
-            k++;
-            i++;
-        }
-    } else if((k < 0) && (-k < m)) {
-        k = -k;
-        i = 0;
-        while(k < m) {
-            mp_binary_set_val_array(dtype, ndarray->array->items, k*n+i, one);
-            k++;
-            i++;
-        }
-    }
-    return MP_OBJ_FROM_PTR(ndarray);
-}
+MP_DEFINE_CONST_FUN_OBJ_2(linalg_dot_obj, linalg_dot);

 mp_obj_t linalg_det(mp_obj_t oin) {
-    if(!mp_obj_is_type(oin, &ulab_ndarray_type)) {
-        mp_raise_TypeError("function defined for ndarrays only");
+    if(!MP_OBJ_IS_TYPE(oin, &ulab_ndarray_type)) {
+        mp_raise_TypeError(translate("function defined for ndarrays only"));
    }
    ndarray_obj_t *in = MP_OBJ_TO_PTR(oin);
    if(in->m != in->n) {
-        mp_raise_ValueError("input must be square matrix");
+        mp_raise_ValueError(translate("input must be square matrix"));
    }
    
    mp_float_t *tmp = m_new(mp_float_t, in->n*in->n);
@ -329,13 +204,15 @@ mp_obj_t linalg_det(mp_obj_t oin) {
    return mp_obj_new_float(det);
 }

+MP_DEFINE_CONST_FUN_OBJ_1(linalg_det_obj, linalg_det);
+
 mp_obj_t linalg_eig(mp_obj_t oin) {
-    if(!mp_obj_is_type(oin, &ulab_ndarray_type)) {
-        mp_raise_TypeError("function defined for ndarrays only");
+    if(!MP_OBJ_IS_TYPE(oin, &ulab_ndarray_type)) {
+        mp_raise_TypeError(translate("function defined for ndarrays only"));
    }
    ndarray_obj_t *in = MP_OBJ_TO_PTR(oin);
    if(in->m != in->n) {
-        mp_raise_ValueError("input must be square matrix");
+        mp_raise_ValueError(translate("input must be square matrix"));
    }
    mp_float_t *array = m_new(mp_float_t, in->array->len);
    for(size_t i=0; i < in->array->len; i++) {
@ -347,7 +224,7 @@ mp_obj_t linalg_eig(mp_obj_t oin) {
            // compare entry (m, n) to (n, m)
            // TODO: this must probably be scaled!
            if(epsilon < MICROPY_FLOAT_C_FUN(fabs)(array[m*in->n + n] - array[n*in->n + m])) {
-                mp_raise_ValueError("input matrix is asymmetric");
+                mp_raise_ValueError(translate("input matrix is asymmetric"));
            }
        }
    }
@ -440,7 +317,7 @@ mp_obj_t linalg_eig(mp_obj_t oin) {
    if(iterations == 0) { 
        // the computation did not converge; numpy raises LinAlgError
        m_del(mp_float_t, array, in->array->len);
-        mp_raise_ValueError("iterations did not converge");
+        mp_raise_ValueError(translate("iterations did not converge"));
    }
    ndarray_obj_t *eigenvalues = create_new_ndarray(1, in->n, NDARRAY_FLOAT);
    mp_float_t *eigvalues = (mp_float_t *)eigenvalues->array->items;
@ -455,3 +332,23 @@ mp_obj_t linalg_eig(mp_obj_t oin) {
    return tuple;
    return MP_OBJ_FROM_PTR(eigenvalues);
 }
+
+MP_DEFINE_CONST_FUN_OBJ_1(linalg_eig_obj, linalg_eig);
+
+STATIC const mp_rom_map_elem_t ulab_linalg_globals_table[] = {
+    { MP_OBJ_NEW_QSTR(MP_QSTR___name__), MP_OBJ_NEW_QSTR(MP_QSTR_linalg) },
+    { MP_ROM_QSTR(MP_QSTR_size), (mp_obj_t)&linalg_size_obj },
+    { MP_ROM_QSTR(MP_QSTR_inv), (mp_obj_t)&linalg_inv_obj },
+    { MP_ROM_QSTR(MP_QSTR_dot), (mp_obj_t)&linalg_dot_obj },
+    { MP_ROM_QSTR(MP_QSTR_det), (mp_obj_t)&linalg_det_obj },
+    { MP_ROM_QSTR(MP_QSTR_eig), (mp_obj_t)&linalg_eig_obj },    
+};
+
+STATIC MP_DEFINE_CONST_DICT(mp_module_ulab_linalg_globals, ulab_linalg_globals_table);
+
+mp_obj_module_t ulab_linalg_module = {
+    .base = { &mp_type_module },
+    .globals = (mp_obj_dict_t*)&mp_module_ulab_linalg_globals,
+};
+
+#endif
--- a/code/linalg.h
+++ b/code/linalg.h
@ -1,3 +1,4 @@
+
 /*
 * This file is part of the micropython-ulab project, 
 *
@ -5,16 +6,15 @@
 *
 * The MIT License (MIT)
 *
- * Copyright (c) 2019 Zoltán Vörös
+ * Copyright (c) 2019-2020 Zoltán Vörös
 */
-    
+
 #ifndef _LINALG_
 #define _LINALG_

+#include "ulab.h"
 #include "ndarray.h"

-#define SWAP(t, a, b) { t tmp = a; a = b; b = tmp; }
-
 #if MICROPY_FLOAT_IMPL == MICROPY_FLOAT_IMPL_FLOAT
 #define epsilon        1.2e-7
 #elif MICROPY_FLOAT_IMPL == MICROPY_FLOAT_IMPL_DOUBLE
@ -23,17 +23,13 @@

 #define JACOBI_MAX     20

-mp_obj_t linalg_transpose(mp_obj_t );
-mp_obj_t linalg_reshape(mp_obj_t , mp_obj_t );
-mp_obj_t linalg_size(size_t , const mp_obj_t *, mp_map_t *);
+#if ULAB_LINALG_MODULE || ULAB_POLY_MODULE
 bool linalg_invert_matrix(mp_float_t *, size_t );
-mp_obj_t linalg_inv(mp_obj_t );
-mp_obj_t linalg_dot(mp_obj_t , mp_obj_t );
-mp_obj_t linalg_zeros(size_t , const mp_obj_t *, mp_map_t *);
-mp_obj_t linalg_ones(size_t , const mp_obj_t *, mp_map_t *);
-mp_obj_t linalg_eye(size_t , const mp_obj_t *, mp_map_t *);
+#endif

-mp_obj_t linalg_det(mp_obj_t );
-mp_obj_t linalg_eig(mp_obj_t );
+#if ULAB_LINALG_MODULE
+
+extern mp_obj_module_t ulab_linalg_module;

 #endif
+#endif
--- a/code/micropython.mk
+++ b/code/micropython.mk
@ -3,13 +3,18 @@ USERMODULES_DIR := $(USERMOD_DIR)

 # Add all C files to SRC_USERMOD.
 SRC_USERMOD += $(USERMODULES_DIR)/ndarray.c
+SRC_USERMOD += $(USERMODULES_DIR)/create.c
 SRC_USERMOD += $(USERMODULES_DIR)/linalg.c
 SRC_USERMOD += $(USERMODULES_DIR)/vectorise.c
 SRC_USERMOD += $(USERMODULES_DIR)/poly.c
 SRC_USERMOD += $(USERMODULES_DIR)/fft.c
 SRC_USERMOD += $(USERMODULES_DIR)/numerical.c
+SRC_USERMOD += $(USERMODULES_DIR)/filter.c
+SRC_USERMOD += $(USERMODULES_DIR)/extras.c
 SRC_USERMOD += $(USERMODULES_DIR)/ulab.c

 # We can add our module folder to include paths if needed
 # This is not actually needed in this example.
 CFLAGS_USERMOD += -I$(USERMODULES_DIR)
+
+CFLAGS_EXTRA = -DMODULE_ULAB_ENABLED=1
--- a/code/ndarray.c
+++ b/code/ndarray.c
@ -1,3 +1,4 @@
+
 /*
 * This file is part of the micropython-ulab project, 
 *
@ -5,9 +6,9 @@
 *
 * The MIT License (MIT)
 *
- * Copyright (c) 2019 Zoltán Vörös
+ * Copyright (c) 2019-2020 Zoltán Vörös
 */
-    
+
 #include <math.h>
 #include <stdio.h>
 #include <stdlib.h>
@ -153,7 +154,7 @@ mp_obj_t ndarray_copy(mp_obj_t self_in) {

 STATIC uint8_t ndarray_init_helper(size_t n_args, const mp_obj_t *pos_args, mp_map_t *kw_args) {
    static const mp_arg_t allowed_args[] = {
-        { MP_QSTR_, MP_ARG_REQUIRED | MP_ARG_OBJ, {.u_rom_obj = MP_ROM_PTR(&mp_const_none_obj)} },
+        { MP_QSTR_, MP_ARG_REQUIRED | MP_ARG_OBJ, {.u_rom_obj = mp_const_none } },
        { MP_QSTR_dtype, MP_ARG_KW_ONLY | MP_ARG_INT, {.u_int = NDARRAY_FLOAT } },
    };
    
@ -164,16 +165,13 @@ STATIC uint8_t ndarray_init_helper(size_t n_args, const mp_obj_t *pos_args, mp_m
    return dtype;
 }

-mp_obj_t ndarray_make_new(const mp_obj_type_t *type, size_t n_args, size_t n_kw, const mp_obj_t *args) {
-    mp_arg_check_num(n_args, n_kw, 1, 2, true);
-    mp_map_t kw_args;
-    mp_map_init_fixed_table(&kw_args, n_kw, args + n_args);
-    uint8_t dtype = ndarray_init_helper(n_args, args, &kw_args);
+STATIC mp_obj_t ndarray_make_new_core(const mp_obj_type_t *type, size_t n_args, size_t n_kw, const mp_obj_t *args, mp_map_t *kw_args) {
+    uint8_t dtype = ndarray_init_helper(n_args, args, kw_args);

    size_t len1, len2=0, i=0;
    mp_obj_t len_in = mp_obj_len_maybe(args[0]);
    if (len_in == MP_OBJ_NULL) {
-        mp_raise_ValueError("first argument must be an iterable");
+        mp_raise_ValueError(translate("first argument must be an iterable"));
    } else {
        // len1 is either the number of rows (for matrices), or the number of elements (row vectors)
        len1 = MP_OBJ_SMALL_INT_VALUE(len_in);
@ -189,7 +187,7 @@ mp_obj_t ndarray_make_new(const mp_obj_type_t *type, size_t n_args, size_t n_kw,
            // Next, we have to check, whether all elements in the outer loop have the same length
            if(i > 0) {
                if(len2 != MP_OBJ_SMALL_INT_VALUE(len_in)) {
-                    mp_raise_ValueError("iterables are not of the same length");
+                    mp_raise_ValueError(translate("iterables are not of the same length"));
                }
            }
            len2 = MP_OBJ_SMALL_INT_VALUE(len_in);
@ -214,14 +212,31 @@ mp_obj_t ndarray_make_new(const mp_obj_type_t *type, size_t n_args, size_t n_kw,
    return MP_OBJ_FROM_PTR(self);
 }

-size_t slice_length(mp_bound_slice_t slice) {
-    // TODO: check, whether this is true!
-    if(slice.step < 0) {
-        slice.step = -slice.step;
-        return (slice.start - slice.stop) / slice.step;
-    } else {
-        return (slice.stop - slice.start) / slice.step;        
+#ifdef CIRCUITPY
+mp_obj_t ndarray_make_new(const mp_obj_type_t *type, size_t n_args, const mp_obj_t *args, mp_map_t *kw_args) {
+    mp_arg_check_num(n_args, kw_args, 1, 2, true);
+    size_t n_kw = 0;
+    if (kw_args != 0) {
+        n_kw = kw_args->used;
    }
+    mp_map_init_fixed_table(kw_args, n_kw, args + n_args);
+    return ndarray_make_new_core(type, n_args, n_kw, args, kw_args);
+}
+#else
+mp_obj_t ndarray_make_new(const mp_obj_type_t *type, size_t n_args, size_t n_kw, const mp_obj_t *args) {
+    mp_arg_check_num(n_args, n_kw, 1, 2, true);
+    mp_map_t kw_args;
+    mp_map_init_fixed_table(&kw_args, n_kw, args + n_args);
+    return ndarray_make_new_core(type, n_args, n_kw, args, &kw_args);
+}
+#endif
+
+size_t slice_length(mp_bound_slice_t slice) {
+    int32_t len, correction = 1;
+    if(slice.step > 0) correction = -1;
+    len = (slice.stop - slice.start + (slice.step + correction)) / slice.step;
+    if(len < 0) return 0;
+    return (size_t)len;
 }

 size_t true_length(mp_obj_t bool_list) {
@ -231,7 +246,7 @@ size_t true_length(mp_obj_t bool_list) {
    mp_obj_t item, iterable = mp_getiter(bool_list, &iter_buf);
    size_t trues = 0;
    while((item = mp_iternext(iterable)) != MP_OBJ_STOP_ITERATION) {
-        if(!mp_obj_is_type(item, &mp_type_bool)) {
+        if(!MP_OBJ_IS_TYPE(item, &mp_type_bool)) {
            // numpy seems to be a little bit inconsistent in when an index is considered
            // to be True/False. Bail out immediately, if the items are not True/False
            return 0;
@ -248,19 +263,19 @@ mp_bound_slice_t generate_slice(mp_uint_t n, mp_obj_t index) {
    mp_bound_slice_t slice;
    if(MP_OBJ_IS_TYPE(index, &mp_type_slice)) {
        mp_seq_get_fast_slice_indexes(n, index, &slice);
-    } else if(mp_obj_is_int(index)) {
+    } else if(MP_OBJ_IS_INT(index)) {
        int32_t _index = mp_obj_get_int(index);
        if(_index < 0) {
            _index += n;
        } 
        if((_index >= n) || (_index < 0)) {
-            mp_raise_msg(&mp_type_IndexError, "index is out of bounds");
+            mp_raise_msg(&mp_type_IndexError, translate("index is out of bounds"));
        }
        slice.start = _index;
        slice.stop = _index + 1;
        slice.step = 1;
    } else {
-        mp_raise_msg(&mp_type_IndexError, "indices must be integers, slices, or Boolean lists");
+        mp_raise_msg(&mp_type_IndexError, translate("indices must be integers, slices, or Boolean lists"));
    }
    return slice;
 }
@ -289,8 +304,8 @@ mp_obj_t insert_slice_list(ndarray_obj_t *ndarray, size_t m, size_t n,
                            mp_obj_t row_list, mp_obj_t column_list, 
                            ndarray_obj_t *values) {
    if((m != values->m) && (n != values->n)) {
-        if((values->array->len != 1)) { // not a single item
-            mp_raise_ValueError("could not broadast input array from shape");
+        if(values->array->len != 1) { // not a single item
+            mp_raise_ValueError(translate("could not broadast input array from shape"));
        }
    }
    size_t cindex, rindex;
@ -377,7 +392,7 @@ mp_obj_t iterate_slice_list(ndarray_obj_t *ndarray, size_t m, size_t n,
                            mp_obj_t row_list, mp_obj_t column_list, 
                            ndarray_obj_t *values) {
    if((m == 0) || (n == 0)) {
-        mp_raise_msg(&mp_type_IndexError, "empty index range");
+        mp_raise_msg(&mp_type_IndexError, translate("empty index range"));
    }

    if(values != NULL) {
@ -462,7 +477,7 @@ mp_obj_t ndarray_get_slice(ndarray_obj_t *ndarray, mp_obj_t index, ndarray_obj_t
    mp_bound_slice_t row_slice = simple_slice(0, 0, 1), column_slice = simple_slice(0, 0, 1);

    size_t m = 0, n = 0;
-    if(mp_obj_is_int(index) && (ndarray->m == 1) && (values == NULL)) { 
+    if(MP_OBJ_IS_INT(index) && (ndarray->m == 1) && (values == NULL)) { 
        // we have a row vector, and don't want to assign
        column_slice = generate_slice(ndarray->n, index);
        if(slice_length(column_slice) == 1) { // we were asked for a single item
@ -471,7 +486,7 @@ mp_obj_t ndarray_get_slice(ndarray_obj_t *ndarray, mp_obj_t index, ndarray_obj_t
        }
    }
    
-    if(mp_obj_is_int(index) || MP_OBJ_IS_TYPE(index, &mp_type_slice)) {
+    if(MP_OBJ_IS_INT(index) || MP_OBJ_IS_TYPE(index, &mp_type_slice)) {
        if(ndarray->m == 1) { // we have a row vector
            column_slice = generate_slice(ndarray->n, index);
            row_slice = simple_slice(0, 1, 1);
@ -495,15 +510,15 @@ mp_obj_t ndarray_get_slice(ndarray_obj_t *ndarray, mp_obj_t index, ndarray_obj_t
    else { // we certainly have a tuple, so let us deal with it
        mp_obj_tuple_t *tuple = MP_OBJ_TO_PTR(index);
        if(tuple->len != 2) {
-            mp_raise_msg(&mp_type_IndexError, "too many indices");
+            mp_raise_msg(&mp_type_IndexError, translate("too many indices"));
        }
        if(!(MP_OBJ_IS_TYPE(tuple->items[0], &mp_type_list) || 
            MP_OBJ_IS_TYPE(tuple->items[0], &mp_type_slice) || 
-            mp_obj_is_int(tuple->items[0])) || 
+            MP_OBJ_IS_INT(tuple->items[0])) || 
           !(MP_OBJ_IS_TYPE(tuple->items[1], &mp_type_list) || 
            MP_OBJ_IS_TYPE(tuple->items[1], &mp_type_slice) || 
-            mp_obj_is_int(tuple->items[1]))) {
-                mp_raise_msg(&mp_type_IndexError, "indices must be integers, slices, or Boolean lists");
+            MP_OBJ_IS_INT(tuple->items[1]))) {
+                mp_raise_msg(&mp_type_IndexError, translate("indices must be integers, slices, or Boolean lists"));
        }
        if(MP_OBJ_IS_TYPE(tuple->items[0], &mp_type_list)) { // rows are indexed by Boolean list
            m = true_length(tuple->items[0]);
@ -543,11 +558,11 @@ mp_obj_t ndarray_subscr(mp_obj_t self_in, mp_obj_t index, mp_obj_t value) {
        return ndarray_get_slice(self, index, NULL);    
    } else { // assignment to slices; the value must be an ndarray, or a scalar
        if(!MP_OBJ_IS_TYPE(value, &ulab_ndarray_type) && 
-          !mp_obj_is_int(value) && !mp_obj_is_float(value)) {
-            mp_raise_ValueError("right hand side must be an ndarray, or a scalar");
+          !MP_OBJ_IS_INT(value) && !mp_obj_is_float(value)) {
+            mp_raise_ValueError(translate("right hand side must be an ndarray, or a scalar"));
        } else {
            ndarray_obj_t *values = NULL;
-            if(mp_obj_is_int(value)) {
+            if(MP_OBJ_IS_INT(value)) {
                values = create_new_ndarray(1, 1, self->array->typecode);
                mp_binary_set_val_array(values->array->typecode, values->array->items, 0, value);   
            } else if(mp_obj_is_float(value)) {
@ -625,22 +640,14 @@ mp_obj_t ndarray_shape(mp_obj_t self_in) {
    return mp_obj_new_tuple(2, tuple);
 }

-mp_obj_t ndarray_rawsize(mp_obj_t self_in) {
-    // returns a 5-tuple with the 
-    // 
-    // 0. number of rows
-    // 1. number of columns
-    // 2. length of the storage (should be equal to the product of 1. and 2.)
-    // 3. length of the data storage in bytes
-    // 4. datum size in bytes
+mp_obj_t ndarray_size(mp_obj_t self_in) {
    ndarray_obj_t *self = MP_OBJ_TO_PTR(self_in);
-    mp_obj_tuple_t *tuple = MP_OBJ_TO_PTR(mp_obj_new_tuple(5, NULL));
-    tuple->items[0] = MP_OBJ_NEW_SMALL_INT(self->m);
-    tuple->items[1] = MP_OBJ_NEW_SMALL_INT(self->n);
-    tuple->items[2] = MP_OBJ_NEW_SMALL_INT(self->array->len);
-    tuple->items[3] = MP_OBJ_NEW_SMALL_INT(self->bytes);
-    tuple->items[4] = MP_OBJ_NEW_SMALL_INT(mp_binary_get_size('@', self->array->typecode, NULL));
-    return tuple;
+    return mp_obj_new_int(self->array->len);
+}
+
+mp_obj_t ndarray_itemsize(mp_obj_t self_in) {
+    ndarray_obj_t *self = MP_OBJ_TO_PTR(self_in);
+    return MP_OBJ_NEW_SMALL_INT(mp_binary_get_size('@', self->array->typecode, NULL));
 }

 mp_obj_t ndarray_flatten(size_t n_args, const mp_obj_t *pos_args, mp_map_t *kw_args) {
@ -655,7 +662,7 @@ mp_obj_t ndarray_flatten(size_t n_args, const mp_obj_t *pos_args, mp_map_t *kw_a
    
    GET_STR_DATA_LEN(args[0].u_obj, order, len);    
    if((len != 1) || ((memcmp(order, "C", 1) != 0) && (memcmp(order, "F", 1) != 0))) {
-        mp_raise_ValueError("flattening order must be either 'C', or 'F'");        
+        mp_raise_ValueError(translate("flattening order must be either 'C', or 'F'"));
    }

    // if order == 'C', we simply have to set m, and n, there is nothing else to do
@ -678,11 +685,6 @@ mp_obj_t ndarray_flatten(size_t n_args, const mp_obj_t *pos_args, mp_map_t *kw_a
    return self_copy;
 }

-mp_obj_t ndarray_asbytearray(mp_obj_t self_in) {
-    ndarray_obj_t *self = MP_OBJ_TO_PTR(self_in);
-    return MP_OBJ_FROM_PTR(self->array);
-}
-
 // Binary operations

 mp_obj_t ndarray_binary_op(mp_binary_op_t op, mp_obj_t lhs, mp_obj_t rhs) {
@ -694,7 +696,7 @@ mp_obj_t ndarray_binary_op(mp_binary_op_t op, mp_obj_t lhs, mp_obj_t rhs) {
    // TODO: implement in-place operators
    mp_obj_t RHS = MP_OBJ_NULL;
    bool rhs_is_scalar = true;
-    if(mp_obj_is_int(rhs)) {
+    if(MP_OBJ_IS_INT(rhs)) {
        int32_t ivalue = mp_obj_get_int(rhs);
        if((ivalue > 0) && (ivalue < 256)) {
            CREATE_SINGLE_ITEM(RHS, uint8_t, NDARRAY_UINT8, ivalue);
@ -715,12 +717,12 @@ mp_obj_t ndarray_binary_op(mp_binary_op_t op, mp_obj_t lhs, mp_obj_t rhs) {
        rhs_is_scalar = false;
    }
    //else 
-    if(mp_obj_is_type(lhs, &ulab_ndarray_type) && mp_obj_is_type(RHS, &ulab_ndarray_type)) { 
+    if(MP_OBJ_IS_TYPE(lhs, &ulab_ndarray_type) && MP_OBJ_IS_TYPE(RHS, &ulab_ndarray_type)) { 
        // next, the ndarray stuff
        ndarray_obj_t *ol = MP_OBJ_TO_PTR(lhs);
        ndarray_obj_t *or = MP_OBJ_TO_PTR(RHS);
        if(!rhs_is_scalar && ((ol->m != or->m) || (ol->n != or->n))) {
-            mp_raise_ValueError("operands could not be broadcast together");
+            mp_raise_ValueError(translate("operands could not be broadcast together"));
        }
        // At this point, the operands should have the same shape
        switch(op) {
@ -823,7 +825,7 @@ mp_obj_t ndarray_binary_op(mp_binary_op_t op, mp_obj_t lhs, mp_obj_t rhs) {
                        RUN_BINARY_LOOP(NDARRAY_FLOAT, mp_float_t, mp_float_t, mp_float_t, ol, or, op);
                    }
                } else { // this should never happen
-                    mp_raise_TypeError("wrong input type");
+                    mp_raise_TypeError(translate("wrong input type"));
                }
                // this instruction should never be reached, but we have to make the compiler happy
                return MP_OBJ_NULL; 
@ -831,7 +833,7 @@ mp_obj_t ndarray_binary_op(mp_binary_op_t op, mp_obj_t lhs, mp_obj_t rhs) {
                return MP_OBJ_NULL; // op not supported                                                        
        }
    } else {
-        mp_raise_TypeError("wrong operand type on the right hand side");
+        mp_raise_TypeError(translate("wrong operand type on the right hand side"));
    }
 }

@ -849,7 +851,7 @@ mp_obj_t ndarray_unary_op(mp_unary_op_t op, mp_obj_t self_in) {
        
        case MP_UNARY_OP_INVERT:
            if(self->array->typecode == NDARRAY_FLOAT) {
-                mp_raise_ValueError("operation is not supported for given type");
+                mp_raise_ValueError(translate("operation is not supported for given type"));
            }
            // we can invert the content byte by byte, there is no need to distinguish 
            // between different typecodes
@ -909,3 +911,67 @@ mp_obj_t ndarray_unary_op(mp_unary_op_t op, mp_obj_t self_in) {
        default: return MP_OBJ_NULL; // operator not supported
    }
 }
+
+mp_obj_t ndarray_transpose(mp_obj_t self_in) {
+    ndarray_obj_t *self = MP_OBJ_TO_PTR(self_in);
+    // the size of a single item in the array
+    uint8_t _sizeof = mp_binary_get_size('@', self->array->typecode, NULL);
+    
+    // NOTE: 
+    // if the matrices are square, we can simply swap items, but 
+    // generic matrices can't be transposed in place, so we have to 
+    // declare a temporary variable
+    
+    // NOTE: 
+    //  In the old matrix, the coordinate (m, n) is m*self->n + n
+    //  We have to assign this to the coordinate (n, m) in the new 
+    //  matrix, i.e., to n*self->m + m (since the new matrix has self->m columns)
+    
+    // one-dimensional arrays can be transposed by simply swapping the dimensions
+    if((self->m != 1) && (self->n != 1)) {
+        uint8_t *c = (uint8_t *)self->array->items;
+        // self->bytes is the size of the bytearray, irrespective of the typecode
+        uint8_t *tmp = m_new(uint8_t, self->bytes);
+        for(size_t m=0; m < self->m; m++) {
+            for(size_t n=0; n < self->n; n++) {
+                memcpy(tmp+_sizeof*(n*self->m + m), c+_sizeof*(m*self->n + n), _sizeof);
+            }
+        }
+        memcpy(self->array->items, tmp, self->bytes);
+        m_del(uint8_t, tmp, self->bytes);
+    } 
+    SWAP(size_t, self->m, self->n);
+    return mp_const_none;
+}
+
+MP_DEFINE_CONST_FUN_OBJ_1(ndarray_transpose_obj, ndarray_transpose);
+
+mp_obj_t ndarray_reshape(mp_obj_t self_in, mp_obj_t shape) {
+    ndarray_obj_t *self = MP_OBJ_TO_PTR(self_in);
+    if(!MP_OBJ_IS_TYPE(shape, &mp_type_tuple) || (MP_OBJ_SMALL_INT_VALUE(mp_obj_len_maybe(shape)) != 2)) {
+        mp_raise_ValueError(translate("shape must be a 2-tuple"));
+    }
+
+    mp_obj_iter_buf_t iter_buf;
+    mp_obj_t item, iterable = mp_getiter(shape, &iter_buf);
+    uint16_t m, n;
+    item = mp_iternext(iterable);
+    m = mp_obj_get_int(item);
+    item = mp_iternext(iterable);
+    n = mp_obj_get_int(item);
+    if(m*n != self->m*self->n) {
+        // TODO: the proper error message would be "cannot reshape array of size %d into shape (%d, %d)"
+        mp_raise_ValueError(translate("cannot reshape array (incompatible input/output shape)"));
+    }
+    self->m = m;
+    self->n = n;
+    return MP_OBJ_FROM_PTR(self);
+}
+
+MP_DEFINE_CONST_FUN_OBJ_2(ndarray_reshape_obj, ndarray_reshape);
+
+mp_int_t ndarray_get_buffer(mp_obj_t self_in, mp_buffer_info_t *bufinfo, mp_uint_t flags) {
+    ndarray_obj_t *self = MP_OBJ_TO_PTR(self_in);
+    // buffer_p.get_buffer() returns zero for success, while mp_get_buffer returns true for success
+    return !mp_get_buffer(self->array, bufinfo, flags);
+}
--- a/code/ndarray.h
+++ b/code/ndarray.h
@ -1,3 +1,4 @@
+
 /*
 * This file is part of the micropython-ulab project, 
 *
@ -5,9 +6,9 @@
 *
 * The MIT License (MIT)
 *
- * Copyright (c) 2019 Zoltán Vörös
+ * Copyright (c) 2019-2020 Zoltán Vörös
 */
-    
+
 #ifndef _NDARRAY_
 #define _NDARRAY_

@ -24,6 +25,12 @@
 #define FLOAT_TYPECODE 'd'
 #endif

+#if !CIRCUITPY
+#define translate(x) x
+#endif
+
+#define SWAP(t, a, b) { t tmp = a; a = b; b = tmp; }
+
 extern const mp_obj_type_t ulab_ndarray_type;

 enum NDARRAY_TYPE {
@ -53,16 +60,30 @@ void ndarray_assign_elements(mp_obj_array_t *, mp_obj_t , uint8_t , size_t *);
 ndarray_obj_t *create_new_ndarray(size_t , size_t , uint8_t );

 mp_obj_t ndarray_copy(mp_obj_t );
+#ifdef CIRCUITPY
+mp_obj_t ndarray_make_new(const mp_obj_type_t *type, size_t n_args, const mp_obj_t *args, mp_map_t *kw_args);
+#else
 mp_obj_t ndarray_make_new(const mp_obj_type_t *, size_t , size_t , const mp_obj_t *);
+#endif
 mp_obj_t ndarray_subscr(mp_obj_t , mp_obj_t , mp_obj_t );
 mp_obj_t ndarray_getiter(mp_obj_t , mp_obj_iter_buf_t *);
 mp_obj_t ndarray_binary_op(mp_binary_op_t , mp_obj_t , mp_obj_t );
 mp_obj_t ndarray_unary_op(mp_unary_op_t , mp_obj_t );

 mp_obj_t ndarray_shape(mp_obj_t );
-mp_obj_t ndarray_rawsize(mp_obj_t );
+mp_obj_t ndarray_size(mp_obj_t );
+mp_obj_t ndarray_itemsize(mp_obj_t );
 mp_obj_t ndarray_flatten(size_t , const mp_obj_t *, mp_map_t *);
-mp_obj_t ndarray_asbytearray(mp_obj_t );
+
+mp_obj_t ndarray_reshape(mp_obj_t , mp_obj_t );
+MP_DECLARE_CONST_FUN_OBJ_2(ndarray_reshape_obj);
+
+mp_obj_t ndarray_transpose(mp_obj_t );
+MP_DECLARE_CONST_FUN_OBJ_1(ndarray_transpose_obj);
+
+mp_int_t ndarray_get_buffer(mp_obj_t obj, mp_buffer_info_t *bufinfo, mp_uint_t flags);
+//void ndarray_attributes(mp_obj_t , qstr , mp_obj_t *);
+

 #define CREATE_SINGLE_ITEM(outarray, type, typecode, value) do {\
    ndarray_obj_t *tmp = create_new_ndarray(1, 1, (typecode));\
--- a/code/ndarray_properties.h
+++ b/code/ndarray_properties.h
@ -0,0 +1,61 @@
+
+/*
+ * This file is part of the micropython-ulab project, 
+ *
+ * https://github.com/v923z/micropython-ulab
+ *
+ * The MIT License (MIT)
+ *
+ * Copyright (c) 2020 Jeff Epler for Adafruit Industries
+ *               2020 Zoltán Vörös   
+*/
+
+#ifndef _NDARRAY_PROPERTIES_
+#define _NDARRAY_PROPERTIES_
+
+#include "py/runtime.h"
+#include "py/binary.h"
+#include "py/obj.h"
+#include "py/objarray.h"
+
+#include "ndarray.h"
+
+#if CIRCUITPY
+typedef struct _mp_obj_property_t {
+    mp_obj_base_t base;
+    mp_obj_t proxy[3]; // getter, setter, deleter
+} mp_obj_property_t;
+
+MP_DEFINE_CONST_FUN_OBJ_1(ndarray_get_shape_obj, ndarray_shape);
+MP_DEFINE_CONST_FUN_OBJ_1(ndarray_get_size_obj, ndarray_size);
+MP_DEFINE_CONST_FUN_OBJ_1(ndarray_get_itemsize_obj, ndarray_itemsize);
+
+STATIC const mp_obj_property_t ndarray_shape_obj = {
+    .base.type = &mp_type_property,
+    .proxy = {(mp_obj_t)&ndarray_get_shape_obj,
+              mp_const_none,
+              mp_const_none },
+};
+
+STATIC const mp_obj_property_t ndarray_size_obj = {
+    .base.type = &mp_type_property,
+    .proxy = {(mp_obj_t)&ndarray_get_size_obj,
+              mp_const_none,
+              mp_const_none },
+};
+
+STATIC const mp_obj_property_t ndarray_itemsize_obj = {
+    .base.type = &mp_type_property,
+    .proxy = {(mp_obj_t)&ndarray_get_itemsize_obj,
+              mp_const_none,
+              mp_const_none },
+};
+#else
+
+MP_DEFINE_CONST_FUN_OBJ_1(ndarray_size_obj, ndarray_size);
+MP_DEFINE_CONST_FUN_OBJ_1(ndarray_itemsize_obj, ndarray_itemsize);
+MP_DEFINE_CONST_FUN_OBJ_1(ndarray_shape_obj, ndarray_shape);
+
+#endif
+
+#endif
--- a/code/numerical.c
+++ b/code/numerical.c
@ -1,3 +1,4 @@
+
 /*
 * This file is part of the micropython-ulab project, 
 *
@ -5,9 +6,9 @@
 *
 * The MIT License (MIT)
 *
- * Copyright (c) 2019 Zoltán Vörös
+ * Copyright (c) 2019-2020 Zoltán Vörös
 */
-    
+
 #include <math.h>
 #include <stdlib.h>
 #include <string.h>
@ -18,6 +19,8 @@
 #include "py/misc.h"
 #include "numerical.h"

+#if ULAB_NUMERICAL_MODULE
+
 enum NUMERICAL_FUNCTION_TYPE {
    NUMERICAL_MIN,
    NUMERICAL_MAX,
@ -28,56 +31,33 @@ enum NUMERICAL_FUNCTION_TYPE {
    NUMERICAL_STD,
 };

-mp_obj_t numerical_linspace(size_t n_args, const mp_obj_t *pos_args, mp_map_t *kw_args) {
-    static const mp_arg_t allowed_args[] = {
-        { MP_QSTR_, MP_ARG_REQUIRED | MP_ARG_OBJ, {.u_rom_obj = MP_ROM_PTR(&mp_const_none_obj) } },
-        { MP_QSTR_, MP_ARG_REQUIRED | MP_ARG_OBJ, {.u_rom_obj = MP_ROM_PTR(&mp_const_none_obj) } },
-        { MP_QSTR_num, MP_ARG_INT, {.u_int = 50} },
-        { MP_QSTR_endpoint, MP_ARG_KW_ONLY | MP_ARG_OBJ, {.u_rom_obj = MP_ROM_PTR(&mp_const_true_obj)} },
-        { MP_QSTR_retstep, MP_ARG_KW_ONLY | MP_ARG_OBJ, {.u_rom_obj = MP_ROM_PTR(&mp_const_false_obj)} },
-        { MP_QSTR_dtype, MP_ARG_KW_ONLY | MP_ARG_INT, {.u_int = NDARRAY_FLOAT} },
-    };
-
-    mp_arg_val_t args[MP_ARRAY_SIZE(allowed_args)];
-    mp_arg_parse_all(2, pos_args, kw_args, MP_ARRAY_SIZE(allowed_args), allowed_args, args);
-
-    uint16_t len = args[2].u_int;
-    if(len < 2) {
-        mp_raise_ValueError("number of points must be at least 2");
-    }
-    mp_float_t value, step;
-    value = mp_obj_get_float(args[0].u_obj);
-    uint8_t typecode = args[5].u_int;
-    if(args[3].u_obj == mp_const_true) step = (mp_obj_get_float(args[1].u_obj)-value)/(len-1);
-    else step = (mp_obj_get_float(args[1].u_obj)-value)/len;
-    ndarray_obj_t *ndarray = create_new_ndarray(1, len, typecode);
-    if(typecode == NDARRAY_UINT8) {
-        uint8_t *array = (uint8_t *)ndarray->array->items;
-        for(size_t i=0; i < len; i++, value += step) array[i] = (uint8_t)value;
-    } else if(typecode == NDARRAY_INT8) {
-        int8_t *array = (int8_t *)ndarray->array->items;
-        for(size_t i=0; i < len; i++, value += step) array[i] = (int8_t)value;
-    } else if(typecode == NDARRAY_UINT16) {
-        uint16_t *array = (uint16_t *)ndarray->array->items;
-        for(size_t i=0; i < len; i++, value += step) array[i] = (uint16_t)value;
-    } else if(typecode == NDARRAY_INT16) {
-        int16_t *array = (int16_t *)ndarray->array->items;
-        for(size_t i=0; i < len; i++, value += step) array[i] = (int16_t)value;
-    } else {
-        mp_float_t *array = (mp_float_t *)ndarray->array->items;
-        for(size_t i=0; i < len; i++, value += step) array[i] = value;
-    }
-    if(args[4].u_obj == mp_const_false) {
-        return MP_OBJ_FROM_PTR(ndarray);
-    } else {
-        mp_obj_t tuple[2];
-        tuple[0] = ndarray;
-        tuple[1] = mp_obj_new_float(step);
-        return mp_obj_new_tuple(2, tuple);
-    }
+void axis_sorter(ndarray_obj_t *ndarray, mp_obj_t axis, size_t *m, size_t *n, size_t *N, 
+                 size_t *increment, size_t *len, size_t *start_inc) {
+    if(axis == mp_const_none) { // flatten the array
+        *m = 1;
+        *n = 1;
+        *len = ndarray->array->len;
+        *N = 1;
+        *increment = 1;
+        *start_inc = ndarray->array->len;
+    } else if((mp_obj_get_int(axis) == 1)) { // along the horizontal axis
+        *m = ndarray->m;
+        *n = 1;
+        *len = ndarray->n;
+        *N = ndarray->m;
+        *increment = 1;
+        *start_inc = ndarray->n;
+    } else { // along vertical axis
+        *m = 1;
+        *n = ndarray->n;
+        *len = ndarray->m;
+        *N = ndarray->n;
+        *increment = ndarray->n;
+        *start_inc = 1;
+    }    
 }

-mp_obj_t numerical_sum_mean_std_array(mp_obj_t oin, uint8_t optype) {
+mp_obj_t numerical_sum_mean_std_iterable(mp_obj_t oin, uint8_t optype, size_t ddof) {
    mp_float_t value, sum = 0.0, sq_sum = 0.0;
    mp_obj_iter_buf_t iter_buf;
    mp_obj_t item, iterable = mp_getiter(oin, &iter_buf);
@ -85,233 +65,191 @@ mp_obj_t numerical_sum_mean_std_array(mp_obj_t oin, uint8_t optype) {
    while ((item = mp_iternext(iterable)) != MP_OBJ_STOP_ITERATION) {
        value = mp_obj_get_float(item);
        sum += value;
-        if(optype == NUMERICAL_STD) {
-            sq_sum += value*value;
-        }
    }
    if(optype ==  NUMERICAL_SUM) {
        return mp_obj_new_float(sum);
    } else if(optype == NUMERICAL_MEAN) {
        return mp_obj_new_float(sum/len);
-    } else {
+    } else { // this should be the case of the standard deviation
+        // TODO: note that we could get away with a single pass, if we used the Weldorf algorithm
+        // That should save a fair amount of time, because we would have to extract the values only once
+        iterable = mp_getiter(oin, &iter_buf);
        sum /= len; // this is now the mean!
-        return mp_obj_new_float(MICROPY_FLOAT_C_FUN(sqrt)(sq_sum/len-sum*sum));
-    }
-}
-
-STATIC mp_float_t numerical_sum_mean_std_single_line(void *data, size_t start, size_t stop, 
-                                                  size_t stride, uint8_t typecode, uint8_t optype) {
-    
-    mp_float_t sum = 0.0, sq_sum = 0.0, value;
-    size_t len = 0;
-    for(size_t i=start; i < stop; i+=stride, len++) {
-        value = ndarray_get_float_value(data, typecode, i);        
-        sum += value;
-        if(optype == NUMERICAL_STD) {
-            sq_sum += value*value;
-        }
-    }
-    if(len == 0) {
-        mp_raise_ValueError("data length is 0!");
-    }
-    if(optype ==  NUMERICAL_SUM) {
-        return sum;
-    } else if(optype == NUMERICAL_MEAN) {
-        return sum/len;
-    } else {
-        sum /= len; // this is now the mean!
-        return MICROPY_FLOAT_C_FUN(sqrt)(sq_sum/len-sum*sum);
-    }
-}
-
-STATIC mp_obj_t numerical_sum_mean_std_matrix(mp_obj_t oin, mp_obj_t axis, uint8_t optype) {
-    ndarray_obj_t *in = MP_OBJ_TO_PTR(oin);
-    if((axis == mp_const_none) || (in->m == 1) || (in->n == 1)) { 
-        // return the value for the flattened array
-        return mp_obj_new_float(numerical_sum_mean_std_single_line(in->array->items, 0, 
-                                                      in->array->len, 1, in->array->typecode, optype));
-    } else {
-        uint8_t _axis = mp_obj_get_int(axis);
-        size_t m = (_axis == 0) ? 1 : in->m;
-        size_t n = (_axis == 0) ? in->n : 1;
-        size_t len = in->array->len;
-        mp_float_t sms;
-        // TODO: pass in->array->typcode to create_new_ndarray
-        ndarray_obj_t *out = create_new_ndarray(m, n, NDARRAY_FLOAT);
-
-        // TODO: these two cases could probably be combined in a more elegant fashion...
-        if(_axis == 0) { // vertical
-            for(size_t i=0; i < n; i++) {
-                sms = numerical_sum_mean_std_single_line(in->array->items, i, len, 
-                                                               n, in->array->typecode, optype);
-                ((float_t *)out->array->items)[i] = sms;
-            }
-        } else { // horizontal
-            for(size_t i=0; i < m; i++) {
-                sms = numerical_sum_mean_std_single_line(in->array->items, i*in->n, 
-                                                               (i+1)*in->n, 1, in->array->typecode, optype);
-                ((float_t *)out->array->items)[i] = sms;
-            }
-        }
-    return MP_OBJ_FROM_PTR(out);
-    }
-}
-
-size_t numerical_argmin_argmax_array(ndarray_obj_t *in, size_t start, 
-                                       size_t stop, size_t stride, uint8_t op) {
-    size_t best_idx = start;
-    if(in->array->typecode == NDARRAY_UINT8) {
-        ARG_MIN_LOOP(in, uint8_t, start, stop, stride, op);
-    } else if(in->array->typecode == NDARRAY_INT8) {
-        ARG_MIN_LOOP(in, int8_t, start, stop, stride, op);
-    } else if(in->array->typecode == NDARRAY_UINT16) {
-        ARG_MIN_LOOP(in, uint16_t, start, stop, stride, op);
-    } else if(in->array->typecode == NDARRAY_INT16) {
-        ARG_MIN_LOOP(in, uint16_t, start, stop, stride, op);
-    } else if(in->array->typecode == NDARRAY_FLOAT) {
-        ARG_MIN_LOOP(in, mp_float_t, start, stop, stride, op);
-    }
-    return best_idx;
-}
-
-void copy_value_into_ndarray(ndarray_obj_t *target, ndarray_obj_t *source, size_t target_idx, size_t source_idx) {
-    // since we are simply copying, it doesn't matter, whether the arrays are signed or unsigned, 
-    // we can cast them in any way we like
-    // This could also be done with byte copies. I don't know, whether that would have any benefits
-    if((target->array->typecode == NDARRAY_UINT8) || (target->array->typecode == NDARRAY_INT8)) {
-        ((uint8_t *)target->array->items)[target_idx] = ((uint8_t *)source->array->items)[source_idx];
-    } else if((target->array->typecode == NDARRAY_UINT16) || (target->array->typecode == NDARRAY_INT16)) {
-        ((uint16_t *)target->array->items)[target_idx] = ((uint16_t *)source->array->items)[source_idx];
-    } else { 
-        ((float *)target->array->items)[target_idx] = ((float *)source->array->items)[source_idx];
-    }
-}
- 
-STATIC mp_obj_t numerical_argmin_argmax(mp_obj_t oin, mp_obj_t axis, uint8_t optype) {
-    if(MP_OBJ_IS_TYPE(oin, &mp_type_tuple) || MP_OBJ_IS_TYPE(oin, &mp_type_list) || 
-        MP_OBJ_IS_TYPE(oin, &mp_type_range)) {
-        // This case will work for single iterables only 
-        size_t idx = 0, best_idx = 0;
-        mp_obj_iter_buf_t iter_buf;
-        mp_obj_t iterable = mp_getiter(oin, &iter_buf);
-        mp_obj_t best_obj = MP_OBJ_NULL;
-        mp_obj_t item;
-        mp_uint_t op = MP_BINARY_OP_LESS;
-        if((optype == NUMERICAL_ARGMAX) || (optype == NUMERICAL_MAX)) op = MP_BINARY_OP_MORE;
        while ((item = mp_iternext(iterable)) != MP_OBJ_STOP_ITERATION) {
-            if ((best_obj == MP_OBJ_NULL) || (mp_binary_op(op, item, best_obj) == mp_const_true)) {
-                best_obj = item;
-                best_idx = idx;
-            }
-            idx++;
+            value = mp_obj_get_float(item) - sum;
+            sq_sum += value * value;
        }
-        if((optype == NUMERICAL_ARGMIN) || (optype == NUMERICAL_ARGMAX)) {
-            return MP_OBJ_NEW_SMALL_INT(best_idx);
-        } else {
-            return best_obj;
-        }
-    } else if(mp_obj_is_type(oin, &ulab_ndarray_type)) {
-            ndarray_obj_t *in = MP_OBJ_TO_PTR(oin);
-            size_t best_idx;
-            if((axis == mp_const_none) || (in->m == 1) || (in->n == 1)) {
-                // return the value for the flattened array                
-                best_idx = numerical_argmin_argmax_array(in, 0, in->array->len, 1, optype);
-                if((optype == NUMERICAL_ARGMIN) || (optype == NUMERICAL_ARGMAX)) {
-                    return MP_OBJ_NEW_SMALL_INT(best_idx);
-                } else {
-                    if(in->array->typecode == NDARRAY_FLOAT) {
-                        return mp_obj_new_float(ndarray_get_float_value(in->array->items, in->array->typecode, best_idx));
-                    } else {
-                        return mp_binary_get_val_array(in->array->typecode, in->array->items, best_idx);
-                    }
-                }
-            } else { // we have to work with a full matrix here
-                uint8_t _axis = mp_obj_get_int(axis);
-                size_t m = (_axis == 0) ? 1 : in->m;
-                size_t n = (_axis == 0) ? in->n : 1;
-                size_t len = in->array->len;
-                ndarray_obj_t *ndarray = NULL;
-                if((optype == NUMERICAL_MAX) || (optype == NUMERICAL_MIN)) {
-                    ndarray = create_new_ndarray(m, n, in->array->typecode);
-                } else { // argmin/argmax
-                    // TODO: one might get away with uint8_t, if both m, and n < 255
-                    ndarray = create_new_ndarray(m, n, NDARRAY_UINT16);
-                }
-
-                // TODO: these two cases could probably be combined in a more elegant fashion...
-                if(_axis == 0) { // vertical
-                    for(size_t i=0; i < n; i++) {
-                        best_idx = numerical_argmin_argmax_array(in, i, len, n, optype);
-                        if((optype == NUMERICAL_MIN) || (optype == NUMERICAL_MAX)) {
-                            copy_value_into_ndarray(ndarray, in, i, best_idx);
-                        } else {
-                            ((uint16_t *)ndarray->array->items)[i] = (uint16_t)(best_idx / n);
-                        }
-                    }
-                } else { // horizontal
-                    for(size_t i=0; i < m; i++) {
-                        best_idx = numerical_argmin_argmax_array(in, i*in->n, (i+1)*in->n, 1, optype);
-                        if((optype == NUMERICAL_MIN) || (optype == NUMERICAL_MAX)) {
-                             copy_value_into_ndarray(ndarray, in, i, best_idx);
-                        } else {
-                            ((uint16_t *)ndarray->array->items)[i] = (uint16_t)(best_idx - i*in->n);
-                        }
-                    }
-                }
-                return MP_OBJ_FROM_PTR(ndarray);
-            }
-            return mp_const_none;
-        }
-    mp_raise_TypeError("input type is not supported");
+        return mp_obj_new_float(MICROPY_FLOAT_C_FUN(sqrt)(sq_sum/(len-ddof)));
+    }
 }

-STATIC mp_obj_t numerical_function(size_t n_args, const mp_obj_t *pos_args, mp_map_t *kw_args, uint8_t type) {
+STATIC mp_obj_t numerical_sum_mean_ndarray(ndarray_obj_t *ndarray, mp_obj_t axis, uint8_t optype) {
+    size_t m, n, increment, start, start_inc, N, len; 
+    axis_sorter(ndarray, axis, &m, &n, &N, &increment, &len, &start_inc);
+    ndarray_obj_t *results = create_new_ndarray(m, n, NDARRAY_FLOAT);
+    mp_float_t sum, sq_sum;
+    mp_float_t *farray = (mp_float_t *)results->array->items;
+    for(size_t j=0; j < N; j++) { // result index
+        start = j * start_inc;
+        sum = sq_sum = 0.0;
+        if(ndarray->array->typecode == NDARRAY_UINT8) {
+            RUN_SUM(ndarray, uint8_t, optype, len, start, increment);
+        } else if(ndarray->array->typecode == NDARRAY_INT8) {
+            RUN_SUM(ndarray, int8_t, optype, len, start, increment);
+        } else if(ndarray->array->typecode == NDARRAY_UINT16) {
+            RUN_SUM(ndarray, uint16_t, optype, len, start, increment);
+        } else if(ndarray->array->typecode == NDARRAY_INT16) {
+            RUN_SUM(ndarray, int16_t, optype, len, start, increment);
+        } else { // this will be mp_float_t, no need to check
+            RUN_SUM(ndarray, mp_float_t, optype, len, start, increment);
+        }
+        if(optype == NUMERICAL_SUM) {
+            farray[j] = sum;
+        } else { // this is the case of the mean
+            farray[j] = sum / len;
+        }
+    }
+    if(results->array->len == 1) {
+        return mp_obj_new_float(farray[0]);
+    }
+    return MP_OBJ_FROM_PTR(results);
+}
+
+mp_obj_t numerical_std_ndarray(ndarray_obj_t *ndarray, mp_obj_t axis, size_t ddof) {
+    size_t m, n, increment, start, start_inc, N, len; 
+    mp_float_t sum, sum_sq;
+    
+    axis_sorter(ndarray, axis, &m, &n, &N, &increment, &len, &start_inc);
+    if(ddof > len) {
+        mp_raise_ValueError(translate("ddof must be smaller than length of data set"));
+    }
+    ndarray_obj_t *results = create_new_ndarray(m, n, NDARRAY_FLOAT);
+    mp_float_t *farray = (mp_float_t *)results->array->items;
+    for(size_t j=0; j < N; j++) { // result index
+        start = j * start_inc;
+        sum = 0.0;
+        sum_sq = 0.0;
+        if(ndarray->array->typecode == NDARRAY_UINT8) {
+            RUN_STD(ndarray, uint8_t, len, start, increment);
+        } else if(ndarray->array->typecode == NDARRAY_INT8) {
+            RUN_STD(ndarray, int8_t, len, start, increment);
+        } else if(ndarray->array->typecode == NDARRAY_UINT16) {
+            RUN_STD(ndarray, uint16_t, len, start, increment);
+        } else if(ndarray->array->typecode == NDARRAY_INT16) {
+            RUN_STD(ndarray, int16_t, len, start, increment);
+        } else { // this will be mp_float_t, no need to check
+            RUN_STD(ndarray, mp_float_t, len, start, increment);
+        }
+        farray[j] = MICROPY_FLOAT_C_FUN(sqrt)(sum_sq/(len - ddof));
+    }
+    if(results->array->len == 1) {
+        return mp_obj_new_float(farray[0]);
+    }
+    return MP_OBJ_FROM_PTR(results);
+}
+
+mp_obj_t numerical_argmin_argmax_iterable(mp_obj_t oin, mp_obj_t axis, uint8_t optype) {
+    size_t idx = 0, best_idx = 0;
+    mp_obj_iter_buf_t iter_buf;
+    mp_obj_t iterable = mp_getiter(oin, &iter_buf);
+    mp_obj_t best_obj = MP_OBJ_NULL;
+    mp_obj_t item;
+    mp_uint_t op = MP_BINARY_OP_LESS;
+    if((optype == NUMERICAL_ARGMAX) || (optype == NUMERICAL_MAX)) op = MP_BINARY_OP_MORE;
+    while ((item = mp_iternext(iterable)) != MP_OBJ_STOP_ITERATION) {
+        if ((best_obj == MP_OBJ_NULL) || (mp_binary_op(op, item, best_obj) == mp_const_true)) {
+            best_obj = item;
+            best_idx = idx;
+        }
+        idx++;
+    }
+    if((optype == NUMERICAL_ARGMIN) || (optype == NUMERICAL_ARGMAX)) {
+        return MP_OBJ_NEW_SMALL_INT(best_idx);
+    } else {
+        return best_obj;
+    }    
+}
+
+mp_obj_t numerical_argmin_argmax_ndarray(ndarray_obj_t *ndarray, mp_obj_t axis, uint8_t optype) {
+    size_t m, n, increment, start, start_inc, N, len;
+    axis_sorter(ndarray, axis, &m, &n, &N, &increment, &len, &start_inc);
+    ndarray_obj_t *results;
+    if((optype == NUMERICAL_ARGMIN) || (optype == NUMERICAL_ARGMAX)) {
+        // we could save some RAM by taking NDARRAY_UINT8, if the dimensions 
+        // are smaller than 256, but the code would become more involving 
+        // (we would also need extra flash space)
+        results = create_new_ndarray(m, n, NDARRAY_UINT16);
+    } else {
+        results = create_new_ndarray(m, n, ndarray->array->typecode);
+    }
+    
+    for(size_t j=0; j < N; j++) { // result index
+        start = j * start_inc;
+        if((ndarray->array->typecode == NDARRAY_UINT8) || (ndarray->array->typecode == NDARRAY_INT8)) {
+            if((optype == NUMERICAL_MAX) || (optype == NUMERICAL_MIN)) {
+                RUN_ARGMIN(ndarray, results, uint8_t, uint8_t, len, start, increment, optype, j);
+            } else {
+                RUN_ARGMIN(ndarray, results, uint8_t, uint16_t, len, start, increment, optype, j);                
+            }
+        } else if((ndarray->array->typecode == NDARRAY_UINT16) || (ndarray->array->typecode == NDARRAY_INT16)) {
+            RUN_ARGMIN(ndarray, results, uint16_t, uint16_t, len, start, increment, optype, j);
+        } else {
+            if((optype == NUMERICAL_MAX) || (optype == NUMERICAL_MIN)) {
+                RUN_ARGMIN(ndarray, results, mp_float_t, mp_float_t, len, start, increment, optype, j);
+            } else {
+                RUN_ARGMIN(ndarray, results, mp_float_t, uint16_t, len, start, increment, optype, j);                
+            }
+        }
+    }
+    return MP_OBJ_FROM_PTR(results);
+}
+
+STATIC mp_obj_t numerical_function(size_t n_args, const mp_obj_t *pos_args, mp_map_t *kw_args, uint8_t optype) {
    static const mp_arg_t allowed_args[] = {
-        { MP_QSTR_, MP_ARG_REQUIRED | MP_ARG_OBJ, {.u_rom_obj = MP_ROM_PTR(&mp_const_none_obj)} } ,
-        { MP_QSTR_axis, MP_ARG_KW_ONLY | MP_ARG_OBJ, {.u_rom_obj = MP_ROM_PTR(&mp_const_none_obj)} },
+        { MP_QSTR_, MP_ARG_REQUIRED | MP_ARG_OBJ, {.u_rom_obj = mp_const_none} } ,
+        { MP_QSTR_axis, MP_ARG_OBJ, {.u_rom_obj = mp_const_none } },
    };

    mp_arg_val_t args[MP_ARRAY_SIZE(allowed_args)];
-    mp_arg_parse_all(1, pos_args, kw_args, MP_ARRAY_SIZE(allowed_args), allowed_args, args);
+    mp_arg_parse_all(n_args, pos_args, kw_args, MP_ARRAY_SIZE(allowed_args), allowed_args, args);
    
    mp_obj_t oin = args[0].u_obj;
    mp_obj_t axis = args[1].u_obj;
    if((axis != mp_const_none) && (mp_obj_get_int(axis) != 0) && (mp_obj_get_int(axis) != 1)) {
        // this seems to pass with False, and True...
-        mp_raise_ValueError("axis must be None, 0, or 1");
+        mp_raise_ValueError(translate("axis must be None, 0, or 1"));
    }
    
    if(MP_OBJ_IS_TYPE(oin, &mp_type_tuple) || MP_OBJ_IS_TYPE(oin, &mp_type_list) || 
        MP_OBJ_IS_TYPE(oin, &mp_type_range)) {
-        switch(type) {
+        switch(optype) {
            case NUMERICAL_MIN:
            case NUMERICAL_ARGMIN:
            case NUMERICAL_MAX:
            case NUMERICAL_ARGMAX:
-                return numerical_argmin_argmax(oin, axis, type);
+                return numerical_argmin_argmax_iterable(oin, axis, optype);
            case NUMERICAL_SUM:
            case NUMERICAL_MEAN:
-            case NUMERICAL_STD:
-                return numerical_sum_mean_std_array(oin, type);
+                return numerical_sum_mean_std_iterable(oin, optype, 0);
            default: // we should never reach this point, but whatever
                return mp_const_none;
        }
    } else if(MP_OBJ_IS_TYPE(oin, &ulab_ndarray_type)) {
-        switch(type) {
+        ndarray_obj_t *ndarray = MP_OBJ_TO_PTR(oin);
+        switch(optype) {
            case NUMERICAL_MIN:
            case NUMERICAL_MAX:
            case NUMERICAL_ARGMIN:
            case NUMERICAL_ARGMAX:
-                return numerical_argmin_argmax(oin, axis, type);
+                return numerical_argmin_argmax_ndarray(ndarray, axis, optype);
            case NUMERICAL_SUM:
            case NUMERICAL_MEAN:
-            case NUMERICAL_STD:
-                return numerical_sum_mean_std_matrix(oin, axis, type);            
+                return numerical_sum_mean_ndarray(ndarray, axis, optype);
            default:
-                mp_raise_NotImplementedError("operation is not implemented on ndarrays");
+                mp_raise_NotImplementedError(translate("operation is not implemented on ndarrays"));
        }
    } else {
-        mp_raise_TypeError("input must be tuple, list, range, or ndarray");
+        mp_raise_TypeError(translate("input must be tuple, list, range, or ndarray"));
    }
    return mp_const_none;
 }
@ -320,35 +258,73 @@ mp_obj_t numerical_min(size_t n_args, const mp_obj_t *pos_args, mp_map_t *kw_arg
    return numerical_function(n_args, pos_args, kw_args, NUMERICAL_MIN);
 }

+MP_DEFINE_CONST_FUN_OBJ_KW(numerical_min_obj, 1, numerical_min);
+
 mp_obj_t numerical_max(size_t n_args, const mp_obj_t *pos_args, mp_map_t *kw_args) {
    return numerical_function(n_args, pos_args, kw_args, NUMERICAL_MAX);
 }

+MP_DEFINE_CONST_FUN_OBJ_KW(numerical_max_obj, 1, numerical_max);
+
 mp_obj_t numerical_argmin(size_t n_args, const mp_obj_t *pos_args, mp_map_t *kw_args) {
    return numerical_function(n_args, pos_args, kw_args, NUMERICAL_ARGMIN);
 }

+MP_DEFINE_CONST_FUN_OBJ_KW(numerical_argmin_obj, 1, numerical_argmin);
+
 mp_obj_t numerical_argmax(size_t n_args, const mp_obj_t *pos_args, mp_map_t *kw_args) {
    return numerical_function(n_args, pos_args, kw_args, NUMERICAL_ARGMAX);
 }

+MP_DEFINE_CONST_FUN_OBJ_KW(numerical_argmax_obj, 1, numerical_argmax);
+
 mp_obj_t numerical_sum(size_t n_args, const mp_obj_t *pos_args, mp_map_t *kw_args) {
    return numerical_function(n_args, pos_args, kw_args, NUMERICAL_SUM);
 }

+MP_DEFINE_CONST_FUN_OBJ_KW(numerical_sum_obj, 1, numerical_sum);
+
 mp_obj_t numerical_mean(size_t n_args, const mp_obj_t *pos_args, mp_map_t *kw_args) {
    return numerical_function(n_args, pos_args, kw_args, NUMERICAL_MEAN);
 }

+MP_DEFINE_CONST_FUN_OBJ_KW(numerical_mean_obj, 1, numerical_mean);
+
 mp_obj_t numerical_std(size_t n_args, const mp_obj_t *pos_args, mp_map_t *kw_args) {
-    return numerical_function(n_args, pos_args, kw_args, NUMERICAL_STD);
+    static const mp_arg_t allowed_args[] = {
+        { MP_QSTR_, MP_ARG_REQUIRED | MP_ARG_OBJ, {.u_rom_obj = mp_const_none } } ,
+        { MP_QSTR_axis, MP_ARG_OBJ, {.u_rom_obj = mp_const_none } },
+        { MP_QSTR_ddof, MP_ARG_KW_ONLY | MP_ARG_INT, {.u_int = 0} },
+    };
+
+    mp_arg_val_t args[MP_ARRAY_SIZE(allowed_args)];
+    mp_arg_parse_all(n_args, pos_args, kw_args, MP_ARRAY_SIZE(allowed_args), allowed_args, args);
+    
+    mp_obj_t oin = args[0].u_obj;
+    mp_obj_t axis = args[1].u_obj;
+    size_t ddof = args[2].u_int;
+    if((axis != mp_const_none) && (mp_obj_get_int(axis) != 0) && (mp_obj_get_int(axis) != 1)) {
+        // this seems to pass with False, and True...
+        mp_raise_ValueError(translate("axis must be None, 0, or 1"));
+    }
+    if(MP_OBJ_IS_TYPE(oin, &mp_type_tuple) || MP_OBJ_IS_TYPE(oin, &mp_type_list) || MP_OBJ_IS_TYPE(oin, &mp_type_range)) {
+        return numerical_sum_mean_std_iterable(oin, NUMERICAL_STD, ddof);
+    } else if(MP_OBJ_IS_TYPE(oin, &ulab_ndarray_type)) {
+        ndarray_obj_t *ndarray = MP_OBJ_TO_PTR(oin);
+        return numerical_std_ndarray(ndarray, axis, ddof);
+    } else {
+        mp_raise_TypeError(translate("input must be tuple, list, range, or ndarray"));
+    }
+    return mp_const_none;
 }

+MP_DEFINE_CONST_FUN_OBJ_KW(numerical_std_obj, 1, numerical_std);
+
 mp_obj_t numerical_roll(size_t n_args, const mp_obj_t *pos_args, mp_map_t *kw_args) {
    static const mp_arg_t allowed_args[] = {
-        { MP_QSTR_, MP_ARG_REQUIRED | MP_ARG_OBJ, {.u_rom_obj = MP_ROM_PTR(&mp_const_none_obj) } },
-        { MP_QSTR_, MP_ARG_REQUIRED | MP_ARG_OBJ, {.u_rom_obj = MP_ROM_PTR(&mp_const_none_obj) } },
-        { MP_QSTR_axis, MP_ARG_KW_ONLY | MP_ARG_OBJ, {.u_rom_obj = MP_ROM_PTR(&mp_const_none_obj)} },
+        { MP_QSTR_, MP_ARG_REQUIRED | MP_ARG_OBJ, {.u_rom_obj = mp_const_none  } },
+        { MP_QSTR_, MP_ARG_REQUIRED | MP_ARG_OBJ, {.u_rom_obj = mp_const_none } },
+        { MP_QSTR_axis, MP_ARG_KW_ONLY | MP_ARG_OBJ, {.u_rom_obj = mp_const_none } },
    };

    mp_arg_val_t args[MP_ARRAY_SIZE(allowed_args)];
@ -359,7 +335,7 @@ mp_obj_t numerical_roll(size_t n_args, const mp_obj_t *pos_args, mp_map_t *kw_ar
    if((args[2].u_obj != mp_const_none) && 
           (mp_obj_get_int(args[2].u_obj) != 0) && 
           (mp_obj_get_int(args[2].u_obj) != 1)) {
-        mp_raise_ValueError("axis must be None, 0, or 1");
+        mp_raise_ValueError(translate("axis must be None, 0, or 1"));
    }

    ndarray_obj_t *in = MP_OBJ_TO_PTR(oin);
@ -424,22 +400,24 @@ mp_obj_t numerical_roll(size_t n_args, const mp_obj_t *pos_args, mp_map_t *kw_ar
    }
 }

+MP_DEFINE_CONST_FUN_OBJ_KW(numerical_roll_obj, 2, numerical_roll);
+
 mp_obj_t numerical_flip(size_t n_args, const mp_obj_t *pos_args, mp_map_t *kw_args) {
    static const mp_arg_t allowed_args[] = {
-        { MP_QSTR_, MP_ARG_REQUIRED | MP_ARG_OBJ, {.u_rom_obj = MP_ROM_PTR(&mp_const_none_obj) } },
-        { MP_QSTR_axis, MP_ARG_KW_ONLY | MP_ARG_OBJ, {.u_rom_obj = MP_ROM_PTR(&mp_const_none_obj)} },
+        { MP_QSTR_, MP_ARG_REQUIRED | MP_ARG_OBJ, {.u_rom_obj = mp_const_none } },
+        { MP_QSTR_axis, MP_ARG_KW_ONLY | MP_ARG_OBJ, {.u_rom_obj = mp_const_none } },
    };

    mp_arg_val_t args[MP_ARRAY_SIZE(allowed_args)];
    mp_arg_parse_all(1, pos_args, kw_args, MP_ARRAY_SIZE(allowed_args), allowed_args, args);
    
-    if(!mp_obj_is_type(args[0].u_obj, &ulab_ndarray_type)) {
-        mp_raise_TypeError("flip argument must be an ndarray");
+    if(!MP_OBJ_IS_TYPE(args[0].u_obj, &ulab_ndarray_type)) {
+        mp_raise_TypeError(translate("flip argument must be an ndarray"));
    }
    if((args[1].u_obj != mp_const_none) && 
           (mp_obj_get_int(args[1].u_obj) != 0) && 
           (mp_obj_get_int(args[1].u_obj) != 1)) {
-        mp_raise_ValueError("axis must be None, 0, or 1");
+        mp_raise_ValueError(translate("axis must be None, 0, or 1"));
    }

    ndarray_obj_t *in = MP_OBJ_TO_PTR(args[0].u_obj);
@ -471,9 +449,11 @@ mp_obj_t numerical_flip(size_t n_args, const mp_obj_t *pos_args, mp_map_t *kw_ar
    return out;
 }

+MP_DEFINE_CONST_FUN_OBJ_KW(numerical_flip_obj, 1, numerical_flip);
+
 mp_obj_t numerical_diff(size_t n_args, const mp_obj_t *pos_args, mp_map_t *kw_args) {
    static const mp_arg_t allowed_args[] = {
-        { MP_QSTR_, MP_ARG_REQUIRED | MP_ARG_OBJ, {.u_rom_obj = MP_ROM_PTR(&mp_const_none_obj) } },
+        { MP_QSTR_, MP_ARG_REQUIRED | MP_ARG_OBJ, {.u_rom_obj = mp_const_none } },
        { MP_QSTR_n, MP_ARG_KW_ONLY | MP_ARG_INT, {.u_int = 1 } },
        { MP_QSTR_axis, MP_ARG_KW_ONLY | MP_ARG_INT, {.u_int = -1 } },
    };
@ -481,8 +461,8 @@ mp_obj_t numerical_diff(size_t n_args, const mp_obj_t *pos_args, mp_map_t *kw_ar
    mp_arg_val_t args[MP_ARRAY_SIZE(allowed_args)];
    mp_arg_parse_all(1, pos_args, kw_args, MP_ARRAY_SIZE(allowed_args), allowed_args, args);
    
-    if(!mp_obj_is_type(args[0].u_obj, &ulab_ndarray_type)) {
-        mp_raise_TypeError("diff argument must be an ndarray");
+    if(!MP_OBJ_IS_TYPE(args[0].u_obj, &ulab_ndarray_type)) {
+        mp_raise_TypeError(translate("diff argument must be an ndarray"));
    }
    
    ndarray_obj_t *in = MP_OBJ_TO_PTR(args[0].u_obj);
@ -492,10 +472,10 @@ mp_obj_t numerical_diff(size_t n_args, const mp_obj_t *pos_args, mp_map_t *kw_ar
    } else if(args[2].u_int == 0) { // differtiate along vertical axis
        increment = in->n;
    } else {
-        mp_raise_ValueError("axis must be -1, 0, or 1");        
+        mp_raise_ValueError(translate("axis must be -1, 0, or 1"));
    }
    if((args[1].u_int < 0) || (args[1].u_int > 9)) {
-        mp_raise_ValueError("n must be between 0, and 9");
+        mp_raise_ValueError(translate("n must be between 0, and 9"));
    }
    uint8_t n = args[1].u_int;
    int8_t *stencil = m_new(int8_t, n+1);
@ -539,9 +519,11 @@ mp_obj_t numerical_diff(size_t n_args, const mp_obj_t *pos_args, mp_map_t *kw_ar
    return MP_OBJ_FROM_PTR(out);
 }

+MP_DEFINE_CONST_FUN_OBJ_KW(numerical_diff_obj, 1, numerical_diff);
+
 mp_obj_t numerical_sort_helper(mp_obj_t oin, mp_obj_t axis, uint8_t inplace) {
-    if(!mp_obj_is_type(oin, &ulab_ndarray_type)) {
-        mp_raise_TypeError("sort argument must be an ndarray");
+    if(!MP_OBJ_IS_TYPE(oin, &ulab_ndarray_type)) {
+        mp_raise_TypeError(translate("sort argument must be an ndarray"));
    }

    ndarray_obj_t *ndarray;
@ -572,7 +554,7 @@ mp_obj_t numerical_sort_helper(mp_obj_t oin, mp_obj_t axis, uint8_t inplace) {
        end = ndarray->m;
        N = ndarray->m;
    } else {
-        mp_raise_ValueError("axis must be -1, 0, None, or 1");        
+        mp_raise_ValueError(translate("axis must be -1, 0, None, or 1"));
    }
    
    size_t q, k, p, c;
@ -598,7 +580,7 @@ mp_obj_t numerical_sort_helper(mp_obj_t oin, mp_obj_t axis, uint8_t inplace) {
 // numpy function
 mp_obj_t numerical_sort(size_t n_args, const mp_obj_t *pos_args, mp_map_t *kw_args) {
    static const mp_arg_t allowed_args[] = {
-        { MP_QSTR_, MP_ARG_REQUIRED | MP_ARG_OBJ, {.u_rom_obj = MP_ROM_PTR(&mp_const_none_obj) } },
+        { MP_QSTR_, MP_ARG_REQUIRED | MP_ARG_OBJ, {.u_rom_obj = mp_const_none } },
        { MP_QSTR_axis, MP_ARG_KW_ONLY | MP_ARG_OBJ, {.u_int = -1 } },
    };

@ -607,10 +589,13 @@ mp_obj_t numerical_sort(size_t n_args, const mp_obj_t *pos_args, mp_map_t *kw_ar

    return numerical_sort_helper(args[0].u_obj, args[1].u_obj, 0);
 }
+
+MP_DEFINE_CONST_FUN_OBJ_KW(numerical_sort_obj, 1, numerical_sort);
+
 // method of an ndarray
 mp_obj_t numerical_sort_inplace(size_t n_args, const mp_obj_t *pos_args, mp_map_t *kw_args) {
    static const mp_arg_t allowed_args[] = {
-        { MP_QSTR_, MP_ARG_REQUIRED | MP_ARG_OBJ, {.u_rom_obj = MP_ROM_PTR(&mp_const_none_obj) } },
+        { MP_QSTR_, MP_ARG_REQUIRED | MP_ARG_OBJ, {.u_rom_obj = mp_const_none } },
        { MP_QSTR_axis, MP_ARG_KW_ONLY | MP_ARG_OBJ, {.u_int = -1 } },
    };

@ -620,15 +605,17 @@ mp_obj_t numerical_sort_inplace(size_t n_args, const mp_obj_t *pos_args, mp_map_
    return numerical_sort_helper(args[0].u_obj, args[1].u_obj, 1);
 }

+MP_DEFINE_CONST_FUN_OBJ_KW(numerical_sort_inplace_obj, 1, numerical_sort_inplace);
+
 mp_obj_t numerical_argsort(size_t n_args, const mp_obj_t *pos_args, mp_map_t *kw_args) {
    static const mp_arg_t allowed_args[] = {
-        { MP_QSTR_, MP_ARG_REQUIRED | MP_ARG_OBJ, {.u_rom_obj = MP_ROM_PTR(&mp_const_none_obj) } },
+        { MP_QSTR_, MP_ARG_REQUIRED | MP_ARG_OBJ, {.u_rom_obj = mp_const_none } },
        { MP_QSTR_axis, MP_ARG_KW_ONLY | MP_ARG_OBJ, {.u_int = -1 } },
    };
    mp_arg_val_t args[MP_ARRAY_SIZE(allowed_args)];
    mp_arg_parse_all(1, pos_args, kw_args, MP_ARRAY_SIZE(allowed_args), allowed_args, args);
-    if(!mp_obj_is_type(args[0].u_obj, &ulab_ndarray_type)) {
-        mp_raise_TypeError("argsort argument must be an ndarray");
+    if(!MP_OBJ_IS_TYPE(args[0].u_obj, &ulab_ndarray_type)) {
+        mp_raise_TypeError(translate("argsort argument must be an ndarray"));
    }

    ndarray_obj_t *ndarray = MP_OBJ_TO_PTR(args[0].u_obj);
@ -658,7 +645,7 @@ mp_obj_t numerical_argsort(size_t n_args, const mp_obj_t *pos_args, mp_map_t *kw
        end = m;
        N = m;
    } else {
-        mp_raise_ValueError("axis must be -1, 0, None, or 1");
+        mp_raise_ValueError(translate("axis must be -1, 0, None, or 1"));
    }

    // at the expense of flash, we could save RAM by creating 
@ -689,3 +676,30 @@ mp_obj_t numerical_argsort(size_t n_args, const mp_obj_t *pos_args, mp_map_t *kw
    }
    return MP_OBJ_FROM_PTR(indices);
 }
+
+MP_DEFINE_CONST_FUN_OBJ_KW(numerical_argsort_obj, 1, numerical_argsort);
+
+STATIC const mp_rom_map_elem_t ulab_numerical_globals_table[] = {
+    { MP_OBJ_NEW_QSTR(MP_QSTR___name__), MP_OBJ_NEW_QSTR(MP_QSTR_numerical) },
+    { MP_OBJ_NEW_QSTR(MP_QSTR_sum), (mp_obj_t)&numerical_sum_obj },
+    { MP_OBJ_NEW_QSTR(MP_QSTR_mean), (mp_obj_t)&numerical_mean_obj },
+    { MP_OBJ_NEW_QSTR(MP_QSTR_std), (mp_obj_t)&numerical_std_obj },
+    { MP_OBJ_NEW_QSTR(MP_QSTR_min), (mp_obj_t)&numerical_min_obj },
+    { MP_OBJ_NEW_QSTR(MP_QSTR_max), (mp_obj_t)&numerical_max_obj },
+    { MP_OBJ_NEW_QSTR(MP_QSTR_argmin), (mp_obj_t)&numerical_argmin_obj },
+    { MP_OBJ_NEW_QSTR(MP_QSTR_argmax), (mp_obj_t)&numerical_argmax_obj },
+    { MP_OBJ_NEW_QSTR(MP_QSTR_roll), (mp_obj_t)&numerical_roll_obj },
+    { MP_OBJ_NEW_QSTR(MP_QSTR_flip), (mp_obj_t)&numerical_flip_obj },
+    { MP_OBJ_NEW_QSTR(MP_QSTR_diff), (mp_obj_t)&numerical_diff_obj },
+    { MP_OBJ_NEW_QSTR(MP_QSTR_sort), (mp_obj_t)&numerical_sort_obj },
+    { MP_OBJ_NEW_QSTR(MP_QSTR_argsort), (mp_obj_t)&numerical_argsort_obj },    
+};
+
+STATIC MP_DEFINE_CONST_DICT(mp_module_ulab_numerical_globals, ulab_numerical_globals_table);
+
+mp_obj_module_t ulab_numerical_module = {
+    .base = { &mp_type_module },
+    .globals = (mp_obj_dict_t*)&mp_module_ulab_numerical_globals,
+};
+
+#endif
--- a/code/numerical.h
+++ b/code/numerical.h
@ -1,3 +1,4 @@
+
 /*
 * This file is part of the micropython-ulab project, 
 *
@ -5,47 +6,62 @@
 *
 * The MIT License (MIT)
 *
- * Copyright (c) 2019 Zoltán Vörös
+ * Copyright (c) 2019-2020 Zoltán Vörös
 */
-    
+
 #ifndef _NUMERICAL_
 #define _NUMERICAL_

+#include "ulab.h"
 #include "ndarray.h"

-mp_obj_t numerical_linspace(size_t , const mp_obj_t *, mp_map_t *);
-mp_obj_t numerical_sum(size_t , const mp_obj_t *, mp_map_t *);
-mp_obj_t numerical_mean(size_t , const mp_obj_t *, mp_map_t *);
-mp_obj_t numerical_std(size_t , const mp_obj_t *, mp_map_t *);
-mp_obj_t numerical_min(size_t , const mp_obj_t *, mp_map_t *);
-mp_obj_t numerical_max(size_t , const mp_obj_t *, mp_map_t *);
-mp_obj_t numerical_argmin(size_t , const mp_obj_t *, mp_map_t *);
-mp_obj_t numerical_argmax(size_t , const mp_obj_t *, mp_map_t *);
-mp_obj_t numerical_roll(size_t , const mp_obj_t *, mp_map_t *);
+#if ULAB_NUMERICAL_MODULE
+
+extern mp_obj_module_t ulab_numerical_module;

 // TODO: implement minimum/maximum, and cumsum
-mp_obj_t numerical_minimum(mp_obj_t , mp_obj_t );
-mp_obj_t numerical_maximum(mp_obj_t , mp_obj_t );
-mp_obj_t numerical_cumsum(size_t , const mp_obj_t *, mp_map_t *);
-mp_obj_t numerical_flip(size_t , const mp_obj_t *, mp_map_t *);
-mp_obj_t numerical_diff(size_t , const mp_obj_t *, mp_map_t *);
-mp_obj_t numerical_sort(size_t , const mp_obj_t *, mp_map_t *);
-mp_obj_t numerical_sort_inplace(size_t , const mp_obj_t *, mp_map_t *);
-mp_obj_t numerical_argsort(size_t , const mp_obj_t *, mp_map_t *);
+//mp_obj_t numerical_minimum(mp_obj_t , mp_obj_t );
+//mp_obj_t numerical_maximum(mp_obj_t , mp_obj_t );
+//mp_obj_t numerical_cumsum(size_t , const mp_obj_t *, mp_map_t *);

-// this macro could be tighter, if we moved the ifs to the argmin function, assigned <, as well as >
-#define ARG_MIN_LOOP(in, type, start, stop, stride, op) do {\
-    type *array = (type *)(in)->array->items;\
+#define RUN_ARGMIN(in, out, typein, typeout, len, start, increment, op, pos) do {\
+    typein *array = (typein *)(in)->array->items;\
+    typeout *outarray = (typeout *)(out)->array->items;\
+    size_t best_index = 0;\
    if(((op) == NUMERICAL_MAX) || ((op) == NUMERICAL_ARGMAX)) {\
-        for(size_t i=(start)+(stride); i < (stop); i+=(stride)) {\
-            if((array)[i] > (array)[best_idx]) {\
-                best_idx = i;\
-            }\
+        for(size_t i=1; i < (len); i++) {\
+            if(array[(start)+i*(increment)] > array[(start)+best_index*(increment)]) best_index = i;\
        }\
+        if((op) == NUMERICAL_MAX) outarray[(pos)] = array[(start)+best_index*(increment)];\
+        else outarray[(pos)] = best_index;\
    } else{\
-        for(size_t i=(start)+(stride); i < (stop); i+=(stride)) {\
-            if((array)[i] < (array)[best_idx]) best_idx = i;\
+        for(size_t i=1; i < (len); i++) {\
+            if(array[(start)+i*(increment)] < array[(start)+best_index*(increment)]) best_index = i;\
        }\
+        if((op) == NUMERICAL_MIN) outarray[(pos)] = array[(start)+best_index*(increment)];\
+        else outarray[(pos)] = best_index;\
+    }\
+} while(0)
+
+#define RUN_SUM(ndarray, type, optype, len, start, increment) do {\
+    type *array = (type *)(ndarray)->array->items;\
+    type value;\
+    for(size_t j=0; j < (len); j++) {\
+        value = array[(start)+j*(increment)];\
+        sum += value;\
+    }\
+} while(0)
+
+#define RUN_STD(ndarray, type, len, start, increment) do {\
+    type *array = (type *)(ndarray)->array->items;\
+    mp_float_t value;\
+    for(size_t j=0; j < (len); j++) {\
+        sum += array[(start)+j*(increment)];\
+    }\
+    sum /= (len);\
+    for(size_t j=0; j < (len); j++) {\
+        value = (array[(start)+j*(increment)] - sum);\
+        sum_sq += value * value;\
    }\
 } while(0)

@ -132,4 +148,19 @@ mp_obj_t numerical_argsort(size_t , const mp_obj_t *, mp_map_t *);
    }\
 } while(0)

+MP_DECLARE_CONST_FUN_OBJ_KW(numerical_min_obj);
+MP_DECLARE_CONST_FUN_OBJ_KW(numerical_max_obj);
+MP_DECLARE_CONST_FUN_OBJ_KW(numerical_argmin_obj);
+MP_DECLARE_CONST_FUN_OBJ_KW(numerical_argmax_obj);
+MP_DECLARE_CONST_FUN_OBJ_KW(numerical_sum_obj);
+MP_DECLARE_CONST_FUN_OBJ_KW(numerical_mean_obj);
+MP_DECLARE_CONST_FUN_OBJ_KW(numerical_std_obj);
+MP_DECLARE_CONST_FUN_OBJ_KW(numerical_roll_obj);
+MP_DECLARE_CONST_FUN_OBJ_KW(numerical_flip_obj);
+MP_DECLARE_CONST_FUN_OBJ_KW(numerical_diff_obj);
+MP_DECLARE_CONST_FUN_OBJ_KW(numerical_sort_obj);
+MP_DECLARE_CONST_FUN_OBJ_KW(numerical_sort_inplace_obj);
+MP_DECLARE_CONST_FUN_OBJ_KW(numerical_argsort_obj);
+
+#endif
 #endif
--- a/code/poly.c
+++ b/code/poly.c
@ -1,3 +1,4 @@
+
 /*
 * This file is part of the micropython-ulab project, 
 *
@ -5,9 +6,9 @@
 *
 * The MIT License (MIT)
 *
- * Copyright (c) 2019 Zoltán Vörös
+ * Copyright (c) 2019-2020 Zoltán Vörös
 */
-    
+
 #include "py/obj.h"
 #include "py/runtime.h"
 #include "py/objarray.h"
@ -15,19 +16,19 @@
 #include "linalg.h"
 #include "poly.h"

-
+#if ULAB_POLY_MODULE
 bool object_is_nditerable(mp_obj_t o_in) {
-    if(mp_obj_is_type(o_in, &ulab_ndarray_type) || 
-      mp_obj_is_type(o_in, &mp_type_tuple) || 
-      mp_obj_is_type(o_in, &mp_type_list) || 
-      mp_obj_is_type(o_in, &mp_type_range)) {
+    if(MP_OBJ_IS_TYPE(o_in, &ulab_ndarray_type) || 
+      MP_OBJ_IS_TYPE(o_in, &mp_type_tuple) || 
+      MP_OBJ_IS_TYPE(o_in, &mp_type_list) || 
+      MP_OBJ_IS_TYPE(o_in, &mp_type_range)) {
        return true;
    }
    return false;
 }

 size_t get_nditerable_len(mp_obj_t o_in) {
-    if(mp_obj_is_type(o_in, &ulab_ndarray_type)) {
+    if(MP_OBJ_IS_TYPE(o_in, &ulab_ndarray_type)) {
        ndarray_obj_t *in = MP_OBJ_TO_PTR(o_in);
        return in->array->len;
    } else {
@ -82,12 +83,14 @@ mp_obj_t poly_polyval(mp_obj_t o_p, mp_obj_t o_x) {
    return MP_OBJ_FROM_PTR(out);
 }

+MP_DEFINE_CONST_FUN_OBJ_2(poly_polyval_obj, poly_polyval);
+
 mp_obj_t poly_polyfit(size_t  n_args, const mp_obj_t *args) {
    if((n_args != 2) && (n_args != 3)) {
-        mp_raise_ValueError("number of arguments must be 2, or 3");
+        mp_raise_ValueError(translate("number of arguments must be 2, or 3"));
    }
    if(!object_is_nditerable(args[0])) {
-        mp_raise_ValueError("input data must be an iterable");
+        mp_raise_ValueError(translate("input data must be an iterable"));
    }
    uint16_t lenx = 0, leny = 0;
    uint8_t deg = 0;
@ -99,7 +102,7 @@ mp_obj_t poly_polyfit(size_t  n_args, const mp_obj_t *args) {
        leny = (uint16_t)mp_obj_get_int(mp_obj_len_maybe(args[0]));
        deg = (uint8_t)mp_obj_get_int(args[1]);
        if(leny < deg) {
-            mp_raise_ValueError("more degrees of freedom than data points");
+            mp_raise_ValueError(translate("more degrees of freedom than data points"));
        }
        lenx = leny;
        x = m_new(mp_float_t, lenx); // assume uniformly spaced data points
@ -112,11 +115,11 @@ mp_obj_t poly_polyfit(size_t  n_args, const mp_obj_t *args) {
        lenx = (uint16_t)mp_obj_get_int(mp_obj_len_maybe(args[0]));
        leny = (uint16_t)mp_obj_get_int(mp_obj_len_maybe(args[0]));
        if(lenx != leny) {
-            mp_raise_ValueError("input vectors must be of equal length");
+            mp_raise_ValueError(translate("input vectors must be of equal length"));
        }
        deg = (uint8_t)mp_obj_get_int(args[2]);
        if(leny < deg) {
-            mp_raise_ValueError("more degrees of freedom than data points");
+            mp_raise_ValueError(translate("more degrees of freedom than data points"));
        }
        x = m_new(mp_float_t, lenx);
        fill_array_iterable(x, args[0]);
@ -156,7 +159,7 @@ mp_obj_t poly_polyfit(size_t  n_args, const mp_obj_t *args) {
        m_del(mp_float_t, x, lenx);
        m_del(mp_float_t, y, lenx);
        m_del(mp_float_t, prod, (deg+1)*(deg+1));
-        mp_raise_ValueError("could not invert Vandermonde matrix");
+        mp_raise_ValueError(translate("could not invert Vandermonde matrix"));
    } 
    // at this point, we have the inverse of X^T * X
    // y is a column vector; x is free now, we can use it for storing intermediate values
@ -191,3 +194,20 @@ mp_obj_t poly_polyfit(size_t  n_args, const mp_obj_t *args) {
    }
    return MP_OBJ_FROM_PTR(beta);
 }
+
+MP_DEFINE_CONST_FUN_OBJ_VAR_BETWEEN(poly_polyfit_obj, 2, 3, poly_polyfit);
+
+STATIC const mp_rom_map_elem_t ulab_poly_globals_table[] = {
+    { MP_OBJ_NEW_QSTR(MP_QSTR___name__), MP_OBJ_NEW_QSTR(MP_QSTR_poly) },
+    { MP_OBJ_NEW_QSTR(MP_QSTR_polyval), (mp_obj_t)&poly_polyval_obj },
+    { MP_OBJ_NEW_QSTR(MP_QSTR_polyfit), (mp_obj_t)&poly_polyfit_obj },
+};
+
+STATIC MP_DEFINE_CONST_DICT(mp_module_ulab_poly_globals, ulab_poly_globals_table);
+
+mp_obj_module_t ulab_poly_module = {
+    .base = { &mp_type_module },
+    .globals = (mp_obj_dict_t*)&mp_module_ulab_poly_globals,
+};
+
+#endif
--- a/code/poly.h
+++ b/code/poly.h
@ -1,3 +1,4 @@
+
 /*
 * This file is part of the micropython-ulab project, 
 *
@ -5,13 +6,20 @@
 *
 * The MIT License (MIT)
 *
- * Copyright (c) 2019 Zoltán Vörös
+ * Copyright (c) 2019-2020 Zoltán Vörös
 */
-    
+
 #ifndef _POLY_
 #define _POLY_

-mp_obj_t poly_polyval(mp_obj_t , mp_obj_t );
-mp_obj_t poly_polyfit(size_t  , const mp_obj_t *);
+#include "ulab.h"
+
+#if ULAB_POLY_MODULE
+
+extern mp_obj_module_t ulab_poly_module;
+
+MP_DECLARE_CONST_FUN_OBJ_2(poly_polyval_obj);
+MP_DECLARE_CONST_FUN_OBJ_VAR_BETWEEN(poly_polyfit_obj);

 #endif
+#endif
--- a/code/ulab.c
+++ b/code/ulab.c
@ -1,13 +1,14 @@
+
 /*
- * This file is part of the micropython-ulab project, 
+ * This file is part of the micropython-ulab project,
 *
 * https://github.com/v923z/micropython-ulab
 *
 * The MIT License (MIT)
 *
- * Copyright (c) 2019 Zoltán Vörös
+ * Copyright (c) 2019-2020 Zoltán Vörös
 */
-    
+
 #include <math.h>
 #include <stdio.h>
 #include <stdlib.h>
@ -17,92 +18,32 @@
 #include "py/obj.h"
 #include "py/objarray.h"

+#include "ulab.h"
 #include "ndarray.h"
+#include "ndarray_properties.h"
+#include "create.h"
 #include "linalg.h"
 #include "vectorise.h"
 #include "poly.h"
 #include "fft.h"
+#include "filter.h"
 #include "numerical.h"
+#include "extras.h"

-#define ULAB_VERSION 0.263
+STATIC MP_DEFINE_STR_OBJ(ulab_version_obj, "0.36.0");

-typedef struct _mp_obj_float_t {
-    mp_obj_base_t base;
-    mp_float_t value;
-} mp_obj_float_t;
-
-mp_obj_float_t ulab_version = {{&mp_type_float}, ULAB_VERSION};
-
-MP_DEFINE_CONST_FUN_OBJ_1(ndarray_shape_obj, ndarray_shape);
-MP_DEFINE_CONST_FUN_OBJ_1(ndarray_rawsize_obj, ndarray_rawsize);
 MP_DEFINE_CONST_FUN_OBJ_KW(ndarray_flatten_obj, 1, ndarray_flatten);
-MP_DEFINE_CONST_FUN_OBJ_1(ndarray_asbytearray_obj, ndarray_asbytearray);
-
-MP_DEFINE_CONST_FUN_OBJ_1(linalg_transpose_obj, linalg_transpose);
-MP_DEFINE_CONST_FUN_OBJ_2(linalg_reshape_obj, linalg_reshape);
-MP_DEFINE_CONST_FUN_OBJ_KW(linalg_size_obj, 1, linalg_size);
-MP_DEFINE_CONST_FUN_OBJ_1(linalg_inv_obj, linalg_inv);
-MP_DEFINE_CONST_FUN_OBJ_2(linalg_dot_obj, linalg_dot);
-MP_DEFINE_CONST_FUN_OBJ_KW(linalg_zeros_obj, 0, linalg_zeros);
-MP_DEFINE_CONST_FUN_OBJ_KW(linalg_ones_obj, 0, linalg_ones);
-MP_DEFINE_CONST_FUN_OBJ_KW(linalg_eye_obj, 0, linalg_eye);
-MP_DEFINE_CONST_FUN_OBJ_1(linalg_det_obj, linalg_det);
-MP_DEFINE_CONST_FUN_OBJ_1(linalg_eig_obj, linalg_eig);
-
-MP_DEFINE_CONST_FUN_OBJ_1(vectorise_acos_obj, vectorise_acos);
-MP_DEFINE_CONST_FUN_OBJ_1(vectorise_acosh_obj, vectorise_acosh);
-MP_DEFINE_CONST_FUN_OBJ_1(vectorise_asin_obj, vectorise_asin);
-MP_DEFINE_CONST_FUN_OBJ_1(vectorise_asinh_obj, vectorise_asinh);
-MP_DEFINE_CONST_FUN_OBJ_1(vectorise_atan_obj, vectorise_atan);
-MP_DEFINE_CONST_FUN_OBJ_1(vectorise_atanh_obj, vectorise_atanh);
-MP_DEFINE_CONST_FUN_OBJ_1(vectorise_ceil_obj, vectorise_ceil);
-MP_DEFINE_CONST_FUN_OBJ_1(vectorise_cos_obj, vectorise_cos);
-MP_DEFINE_CONST_FUN_OBJ_1(vectorise_erf_obj, vectorise_erf);
-MP_DEFINE_CONST_FUN_OBJ_1(vectorise_erfc_obj, vectorise_erfc);
-MP_DEFINE_CONST_FUN_OBJ_1(vectorise_exp_obj, vectorise_exp);
-MP_DEFINE_CONST_FUN_OBJ_1(vectorise_expm1_obj, vectorise_expm1);
-MP_DEFINE_CONST_FUN_OBJ_1(vectorise_floor_obj, vectorise_floor);
-MP_DEFINE_CONST_FUN_OBJ_1(vectorise_gamma_obj, vectorise_gamma);
-MP_DEFINE_CONST_FUN_OBJ_1(vectorise_lgamma_obj, vectorise_lgamma);
-MP_DEFINE_CONST_FUN_OBJ_1(vectorise_log_obj, vectorise_log);
-MP_DEFINE_CONST_FUN_OBJ_1(vectorise_log10_obj, vectorise_log10);
-MP_DEFINE_CONST_FUN_OBJ_1(vectorise_log2_obj, vectorise_log2);
-MP_DEFINE_CONST_FUN_OBJ_1(vectorise_sin_obj, vectorise_sin);
-MP_DEFINE_CONST_FUN_OBJ_1(vectorise_sinh_obj, vectorise_sinh);
-MP_DEFINE_CONST_FUN_OBJ_1(vectorise_sqrt_obj, vectorise_sqrt);
-MP_DEFINE_CONST_FUN_OBJ_1(vectorise_tan_obj, vectorise_tan);
-MP_DEFINE_CONST_FUN_OBJ_1(vectorise_tanh_obj, vectorise_tanh);
-
-STATIC MP_DEFINE_CONST_FUN_OBJ_KW(numerical_linspace_obj, 2, numerical_linspace);
-STATIC MP_DEFINE_CONST_FUN_OBJ_KW(numerical_sum_obj, 1, numerical_sum);
-STATIC MP_DEFINE_CONST_FUN_OBJ_KW(numerical_mean_obj, 1, numerical_mean);
-STATIC MP_DEFINE_CONST_FUN_OBJ_KW(numerical_std_obj, 1, numerical_std);
-STATIC MP_DEFINE_CONST_FUN_OBJ_KW(numerical_min_obj, 1, numerical_min);
-STATIC MP_DEFINE_CONST_FUN_OBJ_KW(numerical_max_obj, 1, numerical_max);
-STATIC MP_DEFINE_CONST_FUN_OBJ_KW(numerical_argmin_obj, 1, numerical_argmin);
-STATIC MP_DEFINE_CONST_FUN_OBJ_KW(numerical_argmax_obj, 1, numerical_argmax);
-STATIC MP_DEFINE_CONST_FUN_OBJ_KW(numerical_roll_obj, 2, numerical_roll);
-STATIC MP_DEFINE_CONST_FUN_OBJ_KW(numerical_flip_obj, 1, numerical_flip);
-STATIC MP_DEFINE_CONST_FUN_OBJ_KW(numerical_diff_obj, 1, numerical_diff);
-STATIC MP_DEFINE_CONST_FUN_OBJ_KW(numerical_sort_obj, 1, numerical_sort);
-STATIC MP_DEFINE_CONST_FUN_OBJ_KW(numerical_sort_inplace_obj, 1, numerical_sort_inplace);
-STATIC MP_DEFINE_CONST_FUN_OBJ_KW(numerical_argsort_obj, 1, numerical_argsort);
-
-STATIC MP_DEFINE_CONST_FUN_OBJ_2(poly_polyval_obj, poly_polyval);
-STATIC MP_DEFINE_CONST_FUN_OBJ_VAR_BETWEEN(poly_polyfit_obj, 2, 3, poly_polyfit);
-
-STATIC MP_DEFINE_CONST_FUN_OBJ_VAR_BETWEEN(fft_fft_obj, 1, 2, fft_fft);
-STATIC MP_DEFINE_CONST_FUN_OBJ_VAR_BETWEEN(fft_ifft_obj, 1, 2, fft_ifft);
-STATIC MP_DEFINE_CONST_FUN_OBJ_VAR_BETWEEN(fft_spectrum_obj, 1, 2, fft_spectrum);

 STATIC const mp_rom_map_elem_t ulab_ndarray_locals_dict_table[] = {
+    { MP_ROM_QSTR(MP_QSTR_reshape), MP_ROM_PTR(&ndarray_reshape_obj) },
+    { MP_ROM_QSTR(MP_QSTR_transpose), MP_ROM_PTR(&ndarray_transpose_obj) },
+    { MP_ROM_QSTR(MP_QSTR_flatten), MP_ROM_PTR(&ndarray_flatten_obj) },
    { MP_ROM_QSTR(MP_QSTR_shape), MP_ROM_PTR(&ndarray_shape_obj) },
-    { MP_ROM_QSTR(MP_QSTR_rawsize), MP_ROM_PTR(&ndarray_rawsize_obj) },
-    { MP_ROM_QSTR(MP_QSTR_flatten), MP_ROM_PTR(&ndarray_flatten_obj) },    
-    { MP_ROM_QSTR(MP_QSTR_asbytearray), MP_ROM_PTR(&ndarray_asbytearray_obj) },
-    { MP_ROM_QSTR(MP_QSTR_transpose), MP_ROM_PTR(&linalg_transpose_obj) },
-    { MP_ROM_QSTR(MP_QSTR_reshape), MP_ROM_PTR(&linalg_reshape_obj) },
+    { MP_ROM_QSTR(MP_QSTR_size), MP_ROM_PTR(&ndarray_size_obj) },
+    { MP_ROM_QSTR(MP_QSTR_itemsize), MP_ROM_PTR(&ndarray_itemsize_obj) },
+#if CIRCUITPY
    { MP_ROM_QSTR(MP_QSTR_sort), MP_ROM_PTR(&numerical_sort_inplace_obj) },
+#endif
 };

 STATIC MP_DEFINE_CONST_DICT(ulab_ndarray_locals_dict, ulab_ndarray_locals_dict_table);
@ -116,62 +57,39 @@ const mp_obj_type_t ulab_ndarray_type = {
    .getiter = ndarray_getiter,
    .unary_op = ndarray_unary_op,
    .binary_op = ndarray_binary_op,
+    .buffer_p = { .get_buffer = ndarray_get_buffer, },
    .locals_dict = (mp_obj_dict_t*)&ulab_ndarray_locals_dict,
 };

 STATIC const mp_map_elem_t ulab_globals_table[] = {
    { MP_OBJ_NEW_QSTR(MP_QSTR___name__), MP_OBJ_NEW_QSTR(MP_QSTR_ulab) },
-    { MP_ROM_QSTR(MP_QSTR___version__), MP_ROM_PTR(&ulab_version) },
+    { MP_ROM_QSTR(MP_QSTR___version__), MP_ROM_PTR(&ulab_version_obj) },
    { MP_OBJ_NEW_QSTR(MP_QSTR_array), (mp_obj_t)&ulab_ndarray_type },
-    { MP_OBJ_NEW_QSTR(MP_QSTR_size), (mp_obj_t)&linalg_size_obj },
-    { MP_OBJ_NEW_QSTR(MP_QSTR_inv), (mp_obj_t)&linalg_inv_obj },
-    { MP_ROM_QSTR(MP_QSTR_dot), (mp_obj_t)&linalg_dot_obj },
-    { MP_ROM_QSTR(MP_QSTR_zeros), (mp_obj_t)&linalg_zeros_obj },
-    { MP_ROM_QSTR(MP_QSTR_ones), (mp_obj_t)&linalg_ones_obj },
-    { MP_ROM_QSTR(MP_QSTR_eye), (mp_obj_t)&linalg_eye_obj },
-    { MP_ROM_QSTR(MP_QSTR_det), (mp_obj_t)&linalg_det_obj },
-    { MP_ROM_QSTR(MP_QSTR_eig), (mp_obj_t)&linalg_eig_obj },    
-    { MP_OBJ_NEW_QSTR(MP_QSTR_acos), (mp_obj_t)&vectorise_acos_obj },
-    { MP_OBJ_NEW_QSTR(MP_QSTR_acosh), (mp_obj_t)&vectorise_acosh_obj },
-    { MP_OBJ_NEW_QSTR(MP_QSTR_asin), (mp_obj_t)&vectorise_asin_obj },
-    { MP_OBJ_NEW_QSTR(MP_QSTR_asinh), (mp_obj_t)&vectorise_asinh_obj },
-    { MP_OBJ_NEW_QSTR(MP_QSTR_atan), (mp_obj_t)&vectorise_atan_obj },
-    { MP_OBJ_NEW_QSTR(MP_QSTR_atanh), (mp_obj_t)&vectorise_atanh_obj },
-    { MP_OBJ_NEW_QSTR(MP_QSTR_ceil), (mp_obj_t)&vectorise_ceil_obj },
-    { MP_OBJ_NEW_QSTR(MP_QSTR_cos), (mp_obj_t)&vectorise_cos_obj },
-    { MP_OBJ_NEW_QSTR(MP_QSTR_erf), (mp_obj_t)&vectorise_erf_obj },
-    { MP_OBJ_NEW_QSTR(MP_QSTR_erfc), (mp_obj_t)&vectorise_erfc_obj },
-    { MP_OBJ_NEW_QSTR(MP_QSTR_exp), (mp_obj_t)&vectorise_exp_obj },
-    { MP_OBJ_NEW_QSTR(MP_QSTR_expm1), (mp_obj_t)&vectorise_expm1_obj },
-    { MP_OBJ_NEW_QSTR(MP_QSTR_floor), (mp_obj_t)&vectorise_floor_obj },
-    { MP_OBJ_NEW_QSTR(MP_QSTR_gamma), (mp_obj_t)&vectorise_gamma_obj },
-    { MP_OBJ_NEW_QSTR(MP_QSTR_lgamma), (mp_obj_t)&vectorise_lgamma_obj },
-    { MP_OBJ_NEW_QSTR(MP_QSTR_log), (mp_obj_t)&vectorise_log_obj },
-    { MP_OBJ_NEW_QSTR(MP_QSTR_log10), (mp_obj_t)&vectorise_log10_obj },
-    { MP_OBJ_NEW_QSTR(MP_QSTR_log2), (mp_obj_t)&vectorise_log2_obj },
-    { MP_OBJ_NEW_QSTR(MP_QSTR_sin), (mp_obj_t)&vectorise_sin_obj },
-    { MP_OBJ_NEW_QSTR(MP_QSTR_sinh), (mp_obj_t)&vectorise_sinh_obj },
-    { MP_OBJ_NEW_QSTR(MP_QSTR_sqrt), (mp_obj_t)&vectorise_sqrt_obj },
-    { MP_OBJ_NEW_QSTR(MP_QSTR_tan), (mp_obj_t)&vectorise_tan_obj },
-    { MP_OBJ_NEW_QSTR(MP_QSTR_tanh), (mp_obj_t)&vectorise_tanh_obj },
-    { MP_OBJ_NEW_QSTR(MP_QSTR_linspace), (mp_obj_t)&numerical_linspace_obj },
-    { MP_OBJ_NEW_QSTR(MP_QSTR_sum), (mp_obj_t)&numerical_sum_obj },
-    { MP_OBJ_NEW_QSTR(MP_QSTR_mean), (mp_obj_t)&numerical_mean_obj },
-    { MP_OBJ_NEW_QSTR(MP_QSTR_std), (mp_obj_t)&numerical_std_obj },
-    { MP_OBJ_NEW_QSTR(MP_QSTR_min), (mp_obj_t)&numerical_min_obj },
-    { MP_OBJ_NEW_QSTR(MP_QSTR_max), (mp_obj_t)&numerical_max_obj },
-    { MP_OBJ_NEW_QSTR(MP_QSTR_argmin), (mp_obj_t)&numerical_argmin_obj },
-    { MP_OBJ_NEW_QSTR(MP_QSTR_argmax), (mp_obj_t)&numerical_argmax_obj },
-    { MP_OBJ_NEW_QSTR(MP_QSTR_roll), (mp_obj_t)&numerical_roll_obj },
-    { MP_OBJ_NEW_QSTR(MP_QSTR_flip), (mp_obj_t)&numerical_flip_obj },
-    { MP_OBJ_NEW_QSTR(MP_QSTR_diff), (mp_obj_t)&numerical_diff_obj },
-    { MP_OBJ_NEW_QSTR(MP_QSTR_sort), (mp_obj_t)&numerical_sort_obj },
-    { MP_OBJ_NEW_QSTR(MP_QSTR_argsort), (mp_obj_t)&numerical_argsort_obj },
-    { MP_OBJ_NEW_QSTR(MP_QSTR_polyval), (mp_obj_t)&poly_polyval_obj },
-    { MP_OBJ_NEW_QSTR(MP_QSTR_polyfit), (mp_obj_t)&poly_polyfit_obj },
-    { MP_OBJ_NEW_QSTR(MP_QSTR_fft), (mp_obj_t)&fft_fft_obj },
-    { MP_OBJ_NEW_QSTR(MP_QSTR_ifft), (mp_obj_t)&fft_ifft_obj },
-    { MP_OBJ_NEW_QSTR(MP_QSTR_spectrum), (mp_obj_t)&fft_spectrum_obj },
+    { MP_ROM_QSTR(MP_QSTR_zeros), (mp_obj_t)&create_zeros_obj },
+    { MP_ROM_QSTR(MP_QSTR_ones), (mp_obj_t)&create_ones_obj },
+    { MP_ROM_QSTR(MP_QSTR_eye), (mp_obj_t)&create_eye_obj },
+    { MP_ROM_QSTR(MP_QSTR_linspace), (mp_obj_t)&create_linspace_obj },
+    #if ULAB_LINALG_MODULE
+    { MP_ROM_QSTR(MP_QSTR_linalg), MP_ROM_PTR(&ulab_linalg_module) },
+    #endif
+    #if ULAB_VECTORISE_MODULE
+    { MP_ROM_QSTR(MP_QSTR_vector), MP_ROM_PTR(&ulab_vectorise_module) },
+    #endif
+    #if ULAB_NUMERICAL_MODULE
+    { MP_ROM_QSTR(MP_QSTR_numerical), MP_ROM_PTR(&ulab_numerical_module) },
+    #endif
+    #if ULAB_POLY_MODULE
+    { MP_ROM_QSTR(MP_QSTR_poly), MP_ROM_PTR(&ulab_poly_module) },
+    #endif
+    #if ULAB_FFT_MODULE
+    { MP_ROM_QSTR(MP_QSTR_fft), MP_ROM_PTR(&ulab_fft_module) },
+    #endif
+    #if ULAB_FILTER_MODULE
+    { MP_ROM_QSTR(MP_QSTR_filter), MP_ROM_PTR(&ulab_filter_module) },
+    #endif
+    #if ULAB_EXTRAS_MODULE
+    { MP_ROM_QSTR(MP_QSTR_extras), MP_ROM_PTR(&ulab_extras_module) },
+    #endif
    // class constants
    { MP_ROM_QSTR(MP_QSTR_uint8), MP_ROM_INT(NDARRAY_UINT8) },
    { MP_ROM_QSTR(MP_QSTR_int8), MP_ROM_INT(NDARRAY_INT8) },
@ -185,7 +103,7 @@ STATIC MP_DEFINE_CONST_DICT (
    ulab_globals_table
 );

-const mp_obj_module_t ulab_user_cmodule = {
+mp_obj_module_t ulab_user_cmodule = {
    .base = { &mp_type_module },
    .globals = (mp_obj_dict_t*)&mp_module_ulab_globals,
 };
--- a/code/ulab.h
+++ b/code/ulab.h
@ -0,0 +1,39 @@
+
+/*
+ * This file is part of the micropython-ulab project, 
+ *
+ * https://github.com/v923z/micropython-ulab
+ *
+ * The MIT License (MIT)
+ *
+ * Copyright (c) 2019-2020 Zoltán Vörös
+*/
+
+#ifndef __ULAB__
+#define __ULAB__
+
+// create
+#define ULAB_CREATE_MODULE (1)
+
+// vectorise (all functions) takes approx. 3 kB of flash space
+#define ULAB_VECTORISE_MODULE (1)
+
+// linalg adds around 6 kB
+#define ULAB_LINALG_MODULE (1)
+
+// poly is approx. 2.5 kB
+#define ULAB_POLY_MODULE (1)
+
+// numerical is about 12 kB
+#define ULAB_NUMERICAL_MODULE (1)
+
+// FFT costs about 2 kB of flash space
+#define ULAB_FFT_MODULE (1)
+
+// the filter module takes about 1 kB of flash space
+#define ULAB_FILTER_MODULE (1)
+
+// user-defined modules
+#define ULAB_EXTRAS_MODULE (0)
+
+#endif
--- a/code/vectorise.c
+++ b/code/vectorise.c
@ -1,3 +1,4 @@
+
 /*
 * This file is part of the micropython-ulab project, 
 *
@ -5,9 +6,9 @@
 *
 * The MIT License (MIT)
 *
- * Copyright (c) 2019 Zoltán Vörös
+ * Copyright (c) 2019-2020 Zoltán Vörös
 */
-    
+
 #include <math.h>
 #include <stdio.h>
 #include <stdlib.h>
@ -21,9 +22,10 @@
 #define MP_PI MICROPY_FLOAT_CONST(3.14159265358979323846)
 #endif
    
+#if ULAB_VECTORISE_MODULE
 mp_obj_t vectorise_generic_vector(mp_obj_t o_in, mp_float_t (*f)(mp_float_t)) {
    // Return a single value, if o_in is not iterable
-    if(mp_obj_is_float(o_in) || mp_obj_is_integer(o_in)) {
+    if(mp_obj_is_float(o_in) || MP_OBJ_IS_INT(o_in)) {
            return mp_obj_new_float(f(mp_obj_get_float(o_in)));
    }
    mp_float_t x;
@ -60,26 +62,111 @@ mp_obj_t vectorise_generic_vector(mp_obj_t o_in, mp_float_t (*f)(mp_float_t)) {
    return mp_const_none;
 }

+
 MATH_FUN_1(acos, acos);
+MP_DEFINE_CONST_FUN_OBJ_1(vectorise_acos_obj, vectorise_acos);
+
 MATH_FUN_1(acosh, acosh);
+MP_DEFINE_CONST_FUN_OBJ_1(vectorise_acosh_obj, vectorise_acosh);
+
 MATH_FUN_1(asin, asin);
+MP_DEFINE_CONST_FUN_OBJ_1(vectorise_asin_obj, vectorise_asin);
+
 MATH_FUN_1(asinh, asinh);
+MP_DEFINE_CONST_FUN_OBJ_1(vectorise_asinh_obj, vectorise_asinh);
+
 MATH_FUN_1(atan, atan);
+MP_DEFINE_CONST_FUN_OBJ_1(vectorise_atan_obj, vectorise_atan);
+
 MATH_FUN_1(atanh, atanh);
+MP_DEFINE_CONST_FUN_OBJ_1(vectorise_atanh_obj, vectorise_atanh);
+
 MATH_FUN_1(ceil, ceil);
+MP_DEFINE_CONST_FUN_OBJ_1(vectorise_ceil_obj, vectorise_ceil);
+
 MATH_FUN_1(cos, cos);
+MP_DEFINE_CONST_FUN_OBJ_1(vectorise_cos_obj, vectorise_cos);
+
+MATH_FUN_1(cosh, cosh);
+MP_DEFINE_CONST_FUN_OBJ_1(vectorise_cosh_obj, vectorise_cosh);
+
 MATH_FUN_1(erf, erf);
+MP_DEFINE_CONST_FUN_OBJ_1(vectorise_erf_obj, vectorise_erf);
+
 MATH_FUN_1(erfc, erfc);
+MP_DEFINE_CONST_FUN_OBJ_1(vectorise_erfc_obj, vectorise_erfc);
+
 MATH_FUN_1(exp, exp);
+MP_DEFINE_CONST_FUN_OBJ_1(vectorise_exp_obj, vectorise_exp);
+
 MATH_FUN_1(expm1, expm1);
+MP_DEFINE_CONST_FUN_OBJ_1(vectorise_expm1_obj, vectorise_expm1);
+
 MATH_FUN_1(floor, floor);
+MP_DEFINE_CONST_FUN_OBJ_1(vectorise_floor_obj, vectorise_floor);
+
 MATH_FUN_1(gamma, tgamma);
+MP_DEFINE_CONST_FUN_OBJ_1(vectorise_gamma_obj, vectorise_gamma);
+
 MATH_FUN_1(lgamma, lgamma);
+MP_DEFINE_CONST_FUN_OBJ_1(vectorise_lgamma_obj, vectorise_lgamma);
+
 MATH_FUN_1(log, log);
+MP_DEFINE_CONST_FUN_OBJ_1(vectorise_log_obj, vectorise_log);
+
 MATH_FUN_1(log10, log10);
+MP_DEFINE_CONST_FUN_OBJ_1(vectorise_log10_obj, vectorise_log10);
+
 MATH_FUN_1(log2, log2);
+MP_DEFINE_CONST_FUN_OBJ_1(vectorise_log2_obj, vectorise_log2);
+
 MATH_FUN_1(sin, sin);
+MP_DEFINE_CONST_FUN_OBJ_1(vectorise_sin_obj, vectorise_sin);
+
 MATH_FUN_1(sinh, sinh);
+MP_DEFINE_CONST_FUN_OBJ_1(vectorise_sinh_obj, vectorise_sinh);
+
 MATH_FUN_1(sqrt, sqrt);
+MP_DEFINE_CONST_FUN_OBJ_1(vectorise_sqrt_obj, vectorise_sqrt);
+
 MATH_FUN_1(tan, tan);
+MP_DEFINE_CONST_FUN_OBJ_1(vectorise_tan_obj, vectorise_tan);
+
 MATH_FUN_1(tanh, tanh);
+MP_DEFINE_CONST_FUN_OBJ_1(vectorise_tanh_obj, vectorise_tanh);
+
+STATIC const mp_rom_map_elem_t ulab_vectorise_globals_table[] = {
+    { MP_OBJ_NEW_QSTR(MP_QSTR___name__), MP_OBJ_NEW_QSTR(MP_QSTR_vector) },
+    { MP_OBJ_NEW_QSTR(MP_QSTR_acos), (mp_obj_t)&vectorise_acos_obj },
+    { MP_OBJ_NEW_QSTR(MP_QSTR_acosh), (mp_obj_t)&vectorise_acosh_obj },
+    { MP_OBJ_NEW_QSTR(MP_QSTR_asin), (mp_obj_t)&vectorise_asin_obj },
+    { MP_OBJ_NEW_QSTR(MP_QSTR_asinh), (mp_obj_t)&vectorise_asinh_obj },
+    { MP_OBJ_NEW_QSTR(MP_QSTR_atan), (mp_obj_t)&vectorise_atan_obj },
+    { MP_OBJ_NEW_QSTR(MP_QSTR_atanh), (mp_obj_t)&vectorise_atanh_obj },
+    { MP_OBJ_NEW_QSTR(MP_QSTR_ceil), (mp_obj_t)&vectorise_ceil_obj },
+    { MP_OBJ_NEW_QSTR(MP_QSTR_cos), (mp_obj_t)&vectorise_cos_obj },
+    { MP_OBJ_NEW_QSTR(MP_QSTR_erf), (mp_obj_t)&vectorise_erf_obj },
+    { MP_OBJ_NEW_QSTR(MP_QSTR_erfc), (mp_obj_t)&vectorise_erfc_obj },
+    { MP_OBJ_NEW_QSTR(MP_QSTR_exp), (mp_obj_t)&vectorise_exp_obj },
+    { MP_OBJ_NEW_QSTR(MP_QSTR_expm1), (mp_obj_t)&vectorise_expm1_obj },
+    { MP_OBJ_NEW_QSTR(MP_QSTR_floor), (mp_obj_t)&vectorise_floor_obj },
+    { MP_OBJ_NEW_QSTR(MP_QSTR_gamma), (mp_obj_t)&vectorise_gamma_obj },
+    { MP_OBJ_NEW_QSTR(MP_QSTR_lgamma), (mp_obj_t)&vectorise_lgamma_obj },
+    { MP_OBJ_NEW_QSTR(MP_QSTR_log), (mp_obj_t)&vectorise_log_obj },
+    { MP_OBJ_NEW_QSTR(MP_QSTR_log10), (mp_obj_t)&vectorise_log10_obj },
+    { MP_OBJ_NEW_QSTR(MP_QSTR_log2), (mp_obj_t)&vectorise_log2_obj },
+    { MP_OBJ_NEW_QSTR(MP_QSTR_sin), (mp_obj_t)&vectorise_sin_obj },
+    { MP_OBJ_NEW_QSTR(MP_QSTR_sinh), (mp_obj_t)&vectorise_sinh_obj },
+    { MP_OBJ_NEW_QSTR(MP_QSTR_sqrt), (mp_obj_t)&vectorise_sqrt_obj },
+    { MP_OBJ_NEW_QSTR(MP_QSTR_tan), (mp_obj_t)&vectorise_tan_obj },
+    { MP_OBJ_NEW_QSTR(MP_QSTR_tanh), (mp_obj_t)&vectorise_tanh_obj },
+};
+
+STATIC MP_DEFINE_CONST_DICT(mp_module_ulab_vectorise_globals, ulab_vectorise_globals_table);
+
+mp_obj_module_t ulab_vectorise_module = {
+    .base = { &mp_type_module },
+    .globals = (mp_obj_dict_t*)&mp_module_ulab_vectorise_globals,
+};
+
+#endif
--- a/code/vectorise.h
+++ b/code/vectorise.h
@ -1,3 +1,4 @@
+
 /*
 * This file is part of the micropython-ulab project, 
 *
@ -5,37 +6,18 @@
 *
 * The MIT License (MIT)
 *
- * Copyright (c) 2019 Zoltán Vörös
+ * Copyright (c) 2019-2020 Zoltán Vörös
 */
-    
+
 #ifndef _VECTORISE_
 #define _VECTORISE_

+#include "ulab.h"
 #include "ndarray.h"

-mp_obj_t vectorise_acos(mp_obj_t );
-mp_obj_t vectorise_acosh(mp_obj_t );
-mp_obj_t vectorise_asin(mp_obj_t );
-mp_obj_t vectorise_asinh(mp_obj_t );
-mp_obj_t vectorise_atan(mp_obj_t );
-mp_obj_t vectorise_atanh(mp_obj_t );
-mp_obj_t vectorise_ceil(mp_obj_t );
-mp_obj_t vectorise_cos(mp_obj_t );
-mp_obj_t vectorise_erf(mp_obj_t );
-mp_obj_t vectorise_erfc(mp_obj_t );
-mp_obj_t vectorise_exp(mp_obj_t );
-mp_obj_t vectorise_expm1(mp_obj_t );
-mp_obj_t vectorise_floor(mp_obj_t );
-mp_obj_t vectorise_gamma(mp_obj_t );
-mp_obj_t vectorise_lgamma(mp_obj_t );
-mp_obj_t vectorise_log(mp_obj_t );
-mp_obj_t vectorise_log10(mp_obj_t );
-mp_obj_t vectorise_log2(mp_obj_t );
-mp_obj_t vectorise_sin(mp_obj_t );
-mp_obj_t vectorise_sinh(mp_obj_t );
-mp_obj_t vectorise_sqrt(mp_obj_t );
-mp_obj_t vectorise_tan(mp_obj_t );
-mp_obj_t vectorise_tanh(mp_obj_t );
+#if ULAB_VECTORISE_MODULE
+
+mp_obj_module_t ulab_vectorise_module;

 #define ITERATE_VECTOR(type, source, out) do {\
    type *input = (type *)(source)->array->items;\
@ -47,6 +29,7 @@ mp_obj_t vectorise_tanh(mp_obj_t );
 #define MATH_FUN_1(py_name, c_name) \
    mp_obj_t vectorise_ ## py_name(mp_obj_t x_obj) { \
        return vectorise_generic_vector(x_obj, MICROPY_FLOAT_C_FUN(c_name)); \
-    }
+}
    
 #endif
+#endif
--- a/docs/manual/source/conf.py
+++ b/docs/manual/source/conf.py
@ -22,7 +22,7 @@ copyright = '2019, Zoltán Vörös'
 author = 'Zoltán Vörös'

 # The full version, including alpha/beta/rc tags
-release = '0.26'
+release = '0.32.2'


 # -- General configuration ---------------------------------------------------
--- a/docs/manual/source/ulab.rst
+++ b/docs/manual/source/ulab.rst
@ -1,19 +1,19 @@
 Introduction
 ============

-In
-https://micropython-usermod.readthedocs.io/en/latest/usermods_14.html, I
-mentioned that I have another story, for another day. The day has come,
-so here is my story.
+In the `last
+chapter <https://micropython-usermod.readthedocs.io/en/latest/usermods_15.html>`__
+of the usermod documentation, I mentioned that I have another story, for
+another day. The day has come, so here is my story.

 Enter ulab
 ----------

-``ulab`` is a numpy-like module for micropython, meant to simplify and
-speed up common mathematical operations on arrays. The primary goal was
-to implement a small subset of numpy that might be useful in the context
-of a microcontroller. This means low-level data processing of linear
-(array) and two-dimensional (matrix) data.
+``ulab`` is a numpy-like module for ``micropython``, meant to simplify
+and speed up common mathematical operations on arrays. The primary goal
+was to implement a small subset of numpy that might be useful in the
+context of a microcontroller. This means low-level data processing of
+linear (array) and two-dimensional (matrix) data.

 Purpose
 -------
@ -27,8 +27,9 @@ microcontroller, the data volume is probably small, but it might lead to
 catastrophic system failure, if these data are not processed in time,
 because the microcontroller is supposed to interact with the outside
 world in a timely fashion. In fact, this latter objective was the
-initiator of this project: I needed the Fourier transform of the ADC
-signal, and all the available options were simply too slow.
+initiator of this project: I needed the Fourier transform of a signal
+coming from the ADC of the pyboard, and all available options were
+simply too slow.

 In addition to speed, another issue that one has to keep in mind when
 working with embedded systems is the amount of available RAM: I believe,
@ -42,15 +43,15 @@ matter, whether they are all smaller than 100, or larger than one
 hundred million. This is obviously a waste of resources in an
 environment, where resources are scarce.

-Finally, there is a reason for using micropython in the first place.
+Finally, there is a reason for using ``micropython`` in the first place.
 Namely, that a microcontroller can be programmed in a very elegant, and
 *pythonic* way. But if it is so, why should we not extend this idea to
 other tasks and concepts that might come up in this context? If there
 was no other reason than this *elegance*, I would find that convincing
 enough.

-Based on the above-mentioned considerations, all functions are
-implemented in a way that
+Based on the above-mentioned considerations, all functions in ``ulab``
+are implemented in a way that

 1. conforms to ``numpy`` as much as possible
 2. is so frugal with RAM as possible,
@ -62,15 +63,17 @@ The main points of ``ulab`` are
   2 dimensions (arrays and matrices). These containers support all the
   relevant unary and binary operators (e.g., ``len``, ==, +, \*, etc.)
 -  vectorised computations on micropython iterables and numerical
-   arrays/matrices (in numpy-speak, universal functions)
+   arrays/matrices (in ``numpy``-speak, universal functions)
 -  basic linear algebra routines (matrix inversion, multiplication,
   reshaping, transposition, determinant, and eigenvalues)
 -  polynomial fits to numerical data
 -  fast Fourier transforms

-At the time of writing this manual (for version 0.26), the library adds
-approximately 30 kB of extra compiled code to the micropython
-(pyboard.v.11) firmware.
+At the time of writing this manual (for version 0.33.2), the library
+adds approximately 30 kB of extra compiled code to the micropython
+(pyboard.v.11) firmware. However, if you are tight with flash space, you
+can easily shave off a couple of kB. See the section on `customising
+ulab <#Custom_builds>`__.

 Resources and legal matters
 ---------------------------
@ -95,8 +98,9 @@ Friendly request

 If you use ``ulab``, and bump into a bug, or think that a particular
 function is missing, or its behaviour does not conform to ``numpy``,
-please, raise an issue on github, so that the community can profit from
-your experiences.
+please, raise a `ulab
+issue <#https://github.com/v923z/micropython-ulab/issues>`__ on github,
+so that the community can profit from your experiences.

 Even better, if you find the project useful, and think that it could be
 made better, faster, tighter, and shinier, please, consider
@ -108,6 +112,89 @@ These last comments apply to the documentation, too. If, in your
 opinion, the documentation is obscure, misleading, or not detailed
 enough, please, let me know, so that *we* can fix it.

+Differences between micropython-ulab and circuitpython-ulab
+-----------------------------------------------------------
+
+``ulab`` has originally been developed for ``micropython``, but has
+since been integrated into a number of its flavours. Most of these
+flavours are simply forks of ``micropython`` itself, with some
+additional functionality. One of the notable exceptions is
+``circuitpython``, which has slightly diverged at the core level, and
+this has some minor consequences. Some of these concern the C
+implementation details only, which all have been sorted out with the
+generous and enthusiastic support of Jeff Epler from `Adafruit
+Industries <http://www.adafruit.com>`__.
+
+There are, however, a couple of instances, where the usage in the two
+environments is slightly different at the python level. These are how
+the packges can be imported, and how the class properties can be
+accessed. In both cases, the ``circuitpython`` implementation results in
+``numpy``-conform code. ``numpy``-compatibility in ``micropython`` will
+be implemented as soon as ``micropython`` itself has the required tools.
+Till then we have to live with a workaround, which I will point out at
+the relevant places.
+
+Customising ``ulab``
+====================
+
+``ulab`` implements a great number of functions, which are organised in
+sub-modules. E.g., functions related to Fourier transforms are located
+in the ``ulab.fft`` sub-module, so you would import ``fft`` as
+
+.. code:: python
+
+   import ulab
+   from ulab import fft
+
+by which point you can get the FFT of your data by calling
+``fft.fft(...)``.
+
+The idea of such grouping of functions and methods is to provide a means
+for granularity: It is quite possible that you do not need all functions
+in a particular application. If you want to save some flash space, you
+can easily exclude arbitrary sub-modules from the firmware. The
+`ulab.h <https://github.com/v923z/micropython-ulab/blob/master/code/ulab.h>`__
+header file contains a pre-processor flag for each sub-module. The
+default setting is 1 for each of them. Setting them to 0 removes the
+module from the compiled firmware.
+
+The first couple of lines of the file look like this
+
+.. code:: c
+
+   // vectorise (all functions) takes approx. 3 kB of flash space
+   #define ULAB_VECTORISE_MODULE (1)
+
+   // linalg adds around 6 kB
+   #define ULAB_LINALG_MODULE (1)
+
+   // poly is approx. 2.5 kB
+   #define ULAB_POLY_MODULE (1)
+
+In order to simplify navigation in the header, each flag begins with
+``ULAB_``, and continues with the name of the sub-module. This name is
+also the ``.c`` file, where the sub-module is implemented. So, e.g., the
+linear algebra routines can be found in ``linalg.c``, and the
+corresponding compiler flag is ``ULAB_LINALG_MODULE``. Each section
+displays a hint as to how much space you can save by un-setting the
+flag.
+
+At first, having to import everything in this way might appear to be
+overly complicated, but there is a very good reason behind all this: you
+can find out at the time of importing, whether a function or sub-module
+is part of your ``ulab`` firmware, or not. The alternative, namely, that
+you do not have to import anything beyond ``ulab``, could prove
+catastrophic: you would learn only at run time (at the moment of calling
+the function in your code) that a particular function is not in the
+firmware, and that is most probably too late.
+
+Except for ``fft``, the standard sub-modules, ``vector``, ``linalg``,
+``numerical``, ``and poly``\ all ``numpy``-compatible. User-defined
+functions that accept ``ndarray``\ s as their argument should be
+implemented in the ``extra`` sub-module, or its sub-modules. Hints as to
+how to do that can be found in the section `Extending
+ulab <#Extending-ulab>`__.
+
 Supported functions and methods
 ===============================

@ -162,23 +249,21 @@ calls on general iterables)
 Methods of ndarrays
 -------------------

-`.shape <#.shape>`__
+`.shape\* <#.shape>`__
+
+`size\* <#size>`__
+
+`itemsize\* <#itemsize>`__

 `.reshape <#.reshape>`__

-`.rawsize\*\* <#.rawsize>`__
-
 `.transpose <#.transpose>`__

 `.flatten\*\* <#.flatten>`__

-`.asbytearray <#.asbytearray>`__
-
 Matrix methods
 --------------

-`size <#size>`__
-
 `inv <#inv>`__

 `dot <#dot>`__
@ -239,14 +324,20 @@ FFT routines

 `spectrum\*\* <#spectrum>`__

+Filter functions
+----------------
+
+`convolve <#convolve>`__
+
 ndarray, the basic container
 ============================

 The ``ndarray`` is the underlying container of numerical data. It is
 derived from micropython’s own ``array`` object, but has a great number
-of extra features starting with how it can be initialised, how
+of extra features starting with how it can be initialised, which
 operations can be done on it, and which functions can accept it as an
-argument.
+argument. One important property of an ``ndarray`` is that it is also a
+proper ``micropython`` iterable.

 Since the ``ndarray`` is a binary container, it is also compact, meaning
 that it takes only a couple of bytes of extra RAM in addition to what is
@ -260,7 +351,7 @@ precision/size of the ``float`` type depends on the definition of
 ``mp_float_t``. Some platforms, e.g., the PYBD, implement ``double``\ s,
 but some, e.g., the pyboard.v.11, don’t. You can find out, what type of
 float your particular platform implements by looking at the output of
-the `.rawsize <#.rawsize>`__ class method.
+the `.itemsize <#.itemsize>`__ class property.

 On the following pages, we will see how one can work with
 ``ndarray``\ s. Those familiar with ``numpy`` should find that the
@ -268,7 +359,7 @@ nomenclature and naming conventions of ``numpy`` are adhered to as
 closely as possible. I will point out the few differences, where
 necessary.

-For the sake of comparison, in addition to ``ulab`` code snippets,
+For the sake of comparison, in addition to the ``ulab`` code snippets,
 sometimes the equivalent ``numpy`` code is also presented. You can find
 out, where the snippet is supposed to run by looking at its first line,
 the header.
@ -392,8 +483,11 @@ Methods of ndarrays
 .shape
 ~~~~~~

-The ``.shape`` method returns a 2-tuple with the number of rows, and
-columns.
+The ``.shape`` method (property) returns a 2-tuple with the number of
+rows, and columns.
+
+**WARNING:** In ``circuitpython``, you can call the method as a
+property, i.e.,

 .. code::
        
@ -403,7 +497,38 @@ columns.
    
    a = np.array([1, 2, 3, 4], dtype=np.int8)
    print("a:\n", a)
-    print("shape of a:", a.shape())
+    print("shape of a:", a.shape)
+    
+    b= np.array([[1, 2], [3, 4]], dtype=np.int8)
+    print("\nb:\n", b)
+    print("shape of b:", b.shape)
+
+.. parsed-literal::
+
+    a:
+     array([1, 2, 3, 4], dtype=int8)
+    shape of a: (1, 4)
+    
+    b:
+     array([[1, 2],
+    	 [3, 4]], dtype=int8)
+    shape of b: (2, 2)
+    
+    
+
+
+**WARNING:** On the other hand, since properties are not implemented in
+``micropython``, there you would call the method as a function, i.e.,
+
+.. code::
+        
+    # code to be run in micropython
+    
+    import ulab as np
+    
+    a = np.array([1, 2, 3, 4], dtype=np.int8)
+    print("a:\n", a)
+    print("shape of a:", a.shape)
    
    b= np.array([[1, 2], [3, 4]], dtype=np.int8)
    print("\nb:\n", b)
@ -423,6 +548,139 @@ columns.
    


+.size
+~~~~~
+
+The ``.size`` method (property) returns an integer with the number of
+elements in the array.
+
+**WARNING:** In ``circuitpython``, the ``numpy`` nomenclature applies,
+i.e.,
+
+.. code::
+        
+    # code to be run in micropython
+    
+    import ulab as np
+    
+    a = np.array([1, 2, 3], dtype=np.int8)
+    print("a:\n", a)
+    print("size of a:", a.size)
+    
+    b= np.array([[1, 2], [3, 4]], dtype=np.int8)
+    print("\nb:\n", b)
+    print("size of b:", b.size)
+
+.. parsed-literal::
+
+    a:
+     array([1, 2, 3], dtype=int8)
+    size of a: 3
+    
+    b:
+     array([[1, 2],
+    	 [3, 4]], dtype=int8)
+    size of b: 4
+    
+    
+
+
+**WARNING:** In ``micropython``, ``size`` is a method, i.e.,
+
+.. code::
+        
+    # code to be run in micropython
+    
+    import ulab as np
+    
+    a = np.array([1, 2, 3], dtype=np.int8)
+    print("a:\n", a)
+    print("size of a:", a.size)
+    
+    b= np.array([[1, 2], [3, 4]], dtype=np.int8)
+    print("\nb:\n", b)
+    print("size of b:", b.size())
+
+.. parsed-literal::
+
+    a:
+     array([1, 2, 3], dtype=int8)
+    size of a: 3
+    
+    b:
+     array([[1, 2],
+    	 [3, 4]], dtype=int8)
+    size of b: 4
+    
+    
+
+
+.itemsize
+~~~~~~~~~
+
+The ``.itemsize`` method (property) returns an integer with the siz
+enumber of elements in the array.
+
+**WARNING:** In ``circuitpython``:
+
+.. code::
+        
+    # code to be run in micropython
+    
+    import ulab as np
+    
+    a = np.array([1, 2, 3], dtype=np.int8)
+    print("a:\n", a)
+    print("itemsize of a:", a.itemsize)
+    
+    b= np.array([[1, 2], [3, 4]], dtype=np.float)
+    print("\nb:\n", b)
+    print("itemsize of b:", b.itemsize)
+
+.. parsed-literal::
+
+    a:
+     array([1, 2, 3], dtype=int8)
+    itemsize of a: 1
+    
+    b:
+     array([[1.0, 2.0],
+    	 [3.0, 4.0]], dtype=float)
+    itemsize of b: 8
+    
+    
+
+
+**WARNING:** In ``micropython``:
+
+.. code::
+        
+    # code to be run in micropython
+    
+    import ulab as np
+    
+    a = np.array([1, 2, 3], dtype=np.int8)
+    print("a:\n", a)
+    print("itemsize of a:", a.itemsize)
+    
+    b= np.array([[1, 2], [3, 4]], dtype=np.float)
+    print("\nb:\n", b)
+    print("itemsize of b:", b.itemsize())
+
+.. parsed-literal::
+
+    a:
+     array([1, 2, 3], dtype=int8)
+    itemsize of a: 1
+    
+    b:
+     array([[1.0, 2.0],
+    	 [3.0, 4.0]], dtype=float)
+    itemsize of b: 8
+    
+    
+
+
 .reshape
 ~~~~~~~~

@ -459,41 +717,6 @@ consistent with the old, a ``ValueError`` exception will be raised.
    


-.rawsize
-~~~~~~~~
-
-The ``rawsize`` method of the ``ndarray`` returns a 5-tuple with the
-following data
-
-1. number of rows
-2. number of columns
-3. length of the storage (should be equal to the product of 1. and 2.)
-4. length of the data storage in bytes
-5. datum size in bytes (1 for ``uint8``/``int8``, 2 for
-   ``uint16``/``int16``, and 4, or 8 for ``floats``, see `ndarray, the
-   basic container <#ndarray,-the-basic-container>`__)
-
-**WARNING:** ``rawsize`` is a ``ulab``-only method; it has no equivalent
-in ``numpy``.
-
-.. code::
-        
-    # code to be run in micropython
-    
-    import ulab as np
-    
-    a = np.array([1, 2, 3, 4], dtype=np.float)
-    print("a: \t\t", a)
-    print("rawsize of a: \t", a.rawsize())
-
-.. parsed-literal::
-
-    a: 		 array([1.0, 2.0, 3.0, 4.0], dtype=float)
-    rawsize of a: 	 (1, 4, 4, 16, 4)
-    
-    
-
-
 .flatten
 ~~~~~~~~

@ -535,83 +758,6 @@ verbatim copy of the contents.
    


-.asbytearray
-~~~~~~~~~~~~
-
-The contents of an ``ndarray`` can be accessed directly by calling the
-``.asbytearray`` method. This will simply return a pointer to the
-underlying flat ``array`` object, which can then be manipulated
-directly.
-
-**WARNING:** ``asbytearray`` is a ``ulab``-only method; it has no
-equivalent in ``numpy``.
-
-In the example below, note the difference between ``a``, and ``buffer``:
-while both are designated as an array, you recognise the micropython
-array from the fact that it prints the typecode (``b`` in this
-particular case). The ``ndarray``, on the other hand, prints out the
-``dtype`` (``int8`` here).
-
-.. code::
-        
-    # code to be run in micropython
-    
-    import ulab as np
-    
-    a = np.array([1, 2, 3, 4], dtype=np.int8)
-    print('a: ', a)
-    buffer = a.asbytearray()
-    print("array content:", buffer)
-    buffer[1] = 123
-    print("array content:", buffer)
-
-.. parsed-literal::
-
-    a:  array([1, 2, 3, 4], dtype=int8)
-    array content: array('b', [1, 2, 3, 4])
-    array content: array('b', [1, 123, 3, 4])
-    
-    
-
-
-This in itself wouldn’t be very interesting, but since ``buffer`` is a
-proper micropython ``array``, we can pass it to functions that can
-employ the buffer protocol. E.g., all the ``ndarray`` facilities can be
-applied to the results of timed ADC conversions.
-
-.. code::
-        
-    # code to be run in micropython
-    
-    import pyb
-    import ulab as np
-    
-    n = 100
-    
-    adc = pyb.ADC(pyb.Pin.board.X19)
-    tim = pyb.Timer(6, freq=10)
-    
-    a = np.array([0]*n, dtype=np.uint8)
-    buffer = a.asbytearray()
-    adc.read_timed(buffer, tim)
-    
-    print("ADC results:\t", a)
-    print("mean of results:\t", np.mean(a))
-    print("std of results:\t", np.std(a))
-
-.. parsed-literal::
-
-    ADC results:	 array([48, 2, 2, ..., 0, 0, 0], dtype=uint8)
-    mean of results:	 1.22
-    std of results:	 4.744639
-    
-
-
-Likewise, data can be read directly into ``ndarray``\ s from other
-interfaces, e.g., SPI, I2C etc, and also, by laying bare the
-``ndarray``, we can pass results of ``ulab`` computations to anything
-that can read from a buffer.
-
 .transpose
 ~~~~~~~~~~

@ -2786,28 +2932,32 @@ parts of the transform separately.
    # code to be run in micropython
    
    import ulab as np
+    from ulab import numerical
+    from ulab import vector
+    from ulab import fft
+    from ulab import linalg
    
-    x = np.linspace(0, 10, num=1024)
-    y = np.sin(x)
-    z = np.zeros(len(x))
+    x = numerical.linspace(0, 10, num=1024)
+    y = vector.sin(x)
+    z = linalg.zeros(len(x))
    
-    a, b = np.fft(x)
+    a, b = fft.fft(x)
    print('real part:\t', a)
    print('\nimaginary part:\t', b)
    
-    c, d = np.fft(x, z)
+    c, d = fft.fft(x, z)
    print('\nreal part:\t', c)
    print('\nimaginary part:\t', d)

 .. parsed-literal::

-    real part:	 array([5119.996, -5.004663, -5.004798, ..., -5.005482, -5.005643, -5.006577], dtype=float)
-    
-    imaginary part:	 array([0.0, 1631.333, 815.659, ..., -543.764, -815.6588, -1631.333], dtype=float)
-    
-    real part:	 array([5119.996, -5.004663, -5.004798, ..., -5.005482, -5.005643, -5.006577], dtype=float)
-    
-    imaginary part:	 array([0.0, 1631.333, 815.659, ..., -543.764, -815.6588, -1631.333], dtype=float)
+    real part:	 array([5119.996, -5.004663, -5.004798, ..., -5.005482, -5.005643, -5.006577], dtype=float)
+    
+    imaginary part:	 array([0.0, 1631.333, 815.659, ..., -543.764, -815.6588, -1631.333], dtype=float)
+    
+    real part:	 array([5119.996, -5.004663, -5.004798, ..., -5.005482, -5.005643, -5.006577], dtype=float)
+    
+    imaginary part:	 array([0.0, 1631.333, 815.659, ..., -543.764, -815.6588, -1631.333], dtype=float)
    


@ -2989,10 +3139,48 @@ Y_2(k) &=& -\frac{i}{2}\left(Y(k) - Y^*(N-k)\right)
 :math:`Y_1, Y_2`, and :math:`Y`, respectively, are the Fourier
 transforms of :math:`y_1, y_2`, and :math:`y = y_1 + iy_2`.

+Filter routines
+===============
+
+numpy:
+https://docs.scipy.org/doc/numpy/reference/generated/numpy.convolve.html
+
+convolve
+--------
+
+Returns the discrete, linear convolution of two one-dimensional
+sequences.
+
+Only the ``full`` mode is supported, and the ``mode`` named parameter is
+not accepted. Note that all other modes can be had by slicing a ``full``
+result.
+
+.. code::
+        
+    # code to be run in micropython
+    
+    import ulab as np
+    
+    x = np.array((1,2,3))
+    y = np.array((1,10,100,1000))
+    
+    print(np.convolve(x, y))
+
+.. parsed-literal::
+
+    array([1.0, 12.0, 123.0, 1230.0, 2300.0, 3000.0], dtype=float)
+    
+    
+
+
 Extending ulab
 ==============

-New functions can easily be added to ``ulab`` in a couple of simple
+As mentioned at the beginning, ``ulab`` is a set of sub-modules, out of
+which one, ``extra`` is explicitly reserved for user code. You should
+implement your functions in this sub-module, or sub-modules thereof.
+
+The new functions can easily be added to ``extra`` in a couple of simple
 steps. At the C level, the type definition of an ``ndarray`` is as
 follows:

@ -3042,7 +3230,7 @@ or
 The ambiguity is caused by the fact that not all platforms implement
 ``double``, and there one has to take ``float``\ s. But you haven’t
 actually got to be concerned by this, because at the very beginning of
-``ndarray.h``, this is already taken care of: the preprocessor figures
+``ndarray.h``, this is already taken care of: the pre-processor figures
 out, what the ``float`` implementation of the hardware platform is, and
 defines the ``NDARRAY_FLOAT`` typecode accordingly. All you have to keep
 in mind is that wherever you would use ``float`` or ``double``, you have
@ -3219,47 +3407,15 @@ and return that, you could work along the following lines:
       return MP_PTR_TO_OBJ(ndarray);
   }

-In the boilerplate above, we cast the pointer to ``array->items`` to the
-required type. There are certain operations, however, when you do not
-need the casting. If you do not want to change the array’s values, only
-their position within the array, you can get away with copying the
-memory content, regardless the type. A good example for such a scenario
-is the transpose function in
-https://github.com/v923z/micropython-ulab/blob/master/code/linalg.c.
-
-Compiling your module
---------------------
-
-Once you have implemented the functionality you wanted, you have to
-include the source code in the make file by adding it to
-``micropython.mk``:
-
-.. code:: makefile
-
-   USERMODULES_DIR := $(USERMOD_DIR)
-
-   # Add all C files to SRC_USERMOD.
-   SRC_USERMOD += $(USERMODULES_DIR)/ndarray.c
-   SRC_USERMOD += $(USERMODULES_DIR)/linalg.c
-   SRC_USERMOD += $(USERMODULES_DIR)/vectorise.c
-   SRC_USERMOD += $(USERMODULES_DIR)/poly.c
-   SRC_USERMOD += $(USERMODULES_DIR)/fft.c
-   SRC_USERMOD += $(USERMODULES_DIR)/numerical.c
-   SRC_USERMOD += $(USERMODULES_DIR)/ulab.c
-
-   SRC_USERMOD += $(USERMODULES_DIR)/your_module.c
-
-   CFLAGS_USERMOD += -I$(USERMODULES_DIR)
-
-In addition, you also have to add the function objects to ``ulab.c``,
-and create a ``QString`` for the function name:
+Once the function implementation is done, you should add the function
+object to the globals dictionary of the ``extra`` sub-module as

 .. code:: c

   ...
       MP_DEFINE_CONST_FUN_OBJ_1(useless_function_obj, userless_function);
   ...
-       STATIC const mp_map_elem_t ulab_globals_table[] = {
+       STATIC const mp_map_elem_t extra_globals_table[] = {
   ...
       { MP_OBJ_NEW_QSTR(MP_QSTR_useless), (mp_obj_t)&useless_function_obj },
   ...
@ -3277,24 +3433,18 @@ and so on. For a thorough discussion on how function objects have to be
 defined, see, e.g., the `user module programming
 manual <#https://micropython-usermod.readthedocs.io/en/latest/>`__.

-At this point, you should be able to compile the module with your
-extension by running ``make`` on the command line
+And with that, you are done. You can simply call the compiler as

 .. code:: bash

-   make USER_C_MODULES=../../../ulab all
+   make BOARD=PYBV11 USER_C_MODULES=../../../ulab all

-for the unix port, and
+and the rest is taken care of.

-.. code:: bash
-
-   make BOARD=PYBV11 CROSS_COMPILE=<arm_tools_path>/bin/arm-none-eabi- USER_C_MODULES=../../../ulab all
-
-for the ``pyboard``, provided that the you have defined
-
-.. code:: makefile
-
-   #define MODULE_ULAB_ENABLED (1)
-
-somewhere in ``micropython/port/unix/mpconfigport.h``, or
-``micropython/stm32/mpconfigport.h``, respectively.
+In the boilerplate above, we cast the pointer to ``array->items`` to the
+required type. There are certain operations, however, when you do not
+need the casting. If you do not want to change the array’s values, only
+their position within the array, you can get away with copying the
+memory content, regardless the type. A good example for such a scenario
+is the transpose function in
+https://github.com/v923z/micropython-ulab/blob/master/code/linalg.c.
--- a/docs/ulab-change-log.md
+++ b/docs/ulab-change-log.md
@ -1,3 +1,104 @@
+Thu, 27 Feb 2020
+
+version 0.36.0
+
+	moved zeros, ones, eye and linspace into separate module (they are still bound at the top level)
+	
+Thu, 27 Feb 2020
+
+version 0.35.0
+
+    Move zeros, ones back into top level ulab module
+
+Tue, 18 Feb 2020
+
+version 0.34.0
+
+    split ulab into multiple modules
+
+Sun, 16 Feb 2020
+
+version 0.33.2
+
+    moved properties into ndarray_properties.h, implemented pointer arithmetic in fft.c to save some time
+
+Fri, 14 Feb 2020
+
+version 0.33.1
+
+    added the __name__attribute to all sub-modules
+
+Thu, 13 Feb 2020
+
+version 0.33.0
+
+    sub-modules are now proper sub-modules of ulab
+
+Mon, 17 Feb 2020
+
+version 0.32.1
+
+    temporary fix for issue #40
+    
+Tue, 11 Feb 2020
+
+version 0.32.0
+
+    added itemsize, size and shape attributes to ndarrays, and removed rawsize
+
+Mon, 10 Feb 2020
+
+version 0.31.0
+
+    removed asbytearray, and added buffer protocol to ndarrays, fixed bad error in filter.c
+    
+Sun, 09 Feb 2020
+
+version 0.30.2
+
+    fixed slice_length in ndarray.c
+
+Sat, 08 Feb 2020
+
+version 0.30.1
+
+    fixed typecode error, added variable inspection, and replaced ternary operators in filter.c
+    
+Fri, 07 Feb 2020
+
+version 0.30.0
+
+    ulab functions can arbitrarily be excluded from the firmware via the ulab.h configuration file
+    
+Thu, 06 Feb 2020
+
+version 0.27.0
+
+    add convolve, the start of a 'filter' functionality group
+
+Wed, 29 Jan 2020
+
+version 0.26.7
+
+    fixed indexing error in linalg.dot
+
+Mon, 20 Jan 2020
+
+version 0.26.6
+
+    replaced MP_ROM_PTR(&mp_const_none_obj), so that module can be compiled for the nucleo board
+
+Tue, 7 Jan 2020
+
+version 0.26.5
+
+    fixed glitch in numerical.c, numerical.h
+
+Mon, 6 Jan 2020
+
+version 0.26.4
+
+    switched version constant to string

 Tue, 31 Dec 2019

--- a/docs/ulab-manual.ipynb
+++ b/docs/ulab-manual.ipynb
@ -24,11 +24,11 @@
  },
  {
   "cell_type": "code",
-   "execution_count": 23,
+   "execution_count": 2,
   "metadata": {
    "ExecuteTime": {
-     "end_time": "2019-11-05T16:01:28.817558Z",
-     "start_time": "2019-11-05T16:01:28.810704Z"
+     "end_time": "2020-02-26T17:04:16.562607Z",
+     "start_time": "2020-02-26T17:04:16.502531Z"
    }
   },
   "outputs": [
@ -66,7 +66,7 @@
    "author = 'Zoltán Vörös'\n",
    "\n",
    "# The full version, including alpha/beta/rc tags\n",
-    "release = '0.26'\n",
+    "release = '0.32.2'\n",
    "\n",
    "\n",
    "# -- General configuration ---------------------------------------------------\n",
@ -120,10 +120,11 @@
  },
  {
   "cell_type": "code",
-   "execution_count": null,
+   "execution_count": 3,
   "metadata": {
    "ExecuteTime": {
-     "start_time": "2019-11-06T17:37:29.723Z"
+     "end_time": "2020-02-26T17:04:49.527515Z",
+     "start_time": "2020-02-26T17:04:21.416456Z"
    }
   },
   "outputs": [],
@ -292,11 +293,11 @@
  },
  {
   "cell_type": "code",
-   "execution_count": 3,
+   "execution_count": 17,
   "metadata": {
    "ExecuteTime": {
-     "end_time": "2019-11-04T20:57:36.730820Z",
-     "start_time": "2019-11-04T20:57:36.723446Z"
+     "end_time": "2020-02-16T14:53:49.098172Z",
+     "start_time": "2020-02-16T14:53:49.093201Z"
    }
   },
   "outputs": [],
@ -310,11 +311,11 @@
  },
  {
   "cell_type": "code",
-   "execution_count": 4,
+   "execution_count": 18,
   "metadata": {
    "ExecuteTime": {
-     "end_time": "2019-11-04T20:57:38.576560Z",
-     "start_time": "2019-11-04T20:57:38.521927Z"
+     "end_time": "2020-02-16T14:53:53.396267Z",
+     "start_time": "2020-02-16T14:53:53.375754Z"
    }
   },
   "outputs": [],
@ -383,13 +384,20 @@
    "ip.register_magics(PyboardMagic)"
   ]
  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## pyboard"
+   ]
+  },
  {
   "cell_type": "code",
-   "execution_count": 520,
+   "execution_count": 111,
   "metadata": {
    "ExecuteTime": {
-     "end_time": "2019-10-20T06:48:01.610725Z",
-     "start_time": "2019-10-20T06:48:00.856261Z"
+     "end_time": "2020-02-16T18:36:59.172039Z",
+     "start_time": "2020-02-16T18:36:59.144651Z"
    }
   },
   "outputs": [],
@ -401,11 +409,11 @@
  },
  {
   "cell_type": "code",
-   "execution_count": 501,
+   "execution_count": 115,
   "metadata": {
    "ExecuteTime": {
-     "end_time": "2019-10-19T13:36:42.010602Z",
-     "start_time": "2019-10-19T13:36:42.003900Z"
+     "end_time": "2020-02-16T18:50:42.907664Z",
+     "start_time": "2020-02-16T18:50:42.903709Z"
    }
   },
   "outputs": [],
@ -477,21 +485,21 @@
   "cell_type": "markdown",
   "metadata": {},
   "source": [
-    "In https://micropython-usermod.readthedocs.io/en/latest/usermods_14.html, I mentioned that I have another story, for another day. The day has come, so here is my story.\n",
+    "In the [last chapter](https://micropython-usermod.readthedocs.io/en/latest/usermods_15.html) of the usermod documentation, I mentioned that I have another story, for another day. The day has come, so here is my story.\n",
    "\n",
    "## Enter ulab\n",
    "\n",
-    "`ulab` is a numpy-like module for micropython, meant to simplify and speed up common mathematical operations on arrays. The primary goal was to implement a small subset of numpy that might be useful in the context of a microcontroller. This means low-level data processing of linear (array) and two-dimensional (matrix) data.\n",
+    "`ulab` is a numpy-like module for `micropython`, meant to simplify and speed up common mathematical operations on arrays. The primary goal was to implement a small subset of numpy that might be useful in the context of a microcontroller. This means low-level data processing of linear (array) and two-dimensional (matrix) data.\n",
    "\n",
    "## Purpose\n",
    "\n",
-    "Of course, the first question that one has to answer is, why on Earth one would need a fast math library on a microcontroller. After all, it is not expected that heavy number crunching is going to take place on bare metal. It is not meant to. On a PC, the main reason for writing fast code is the sheer amount of data that one wants to process. On a microcontroller, the data volume is probably small, but it might lead to catastrophic system failure, if these data are not processed in time, because the microcontroller is supposed to interact with the outside world in a timely fashion. In fact, this latter objective was the initiator of this project: I needed the Fourier transform of the ADC signal, and all the available options were simply too slow. \n",
+    "Of course, the first question that one has to answer is, why on Earth one would need a fast math library on a microcontroller. After all, it is not expected that heavy number crunching is going to take place on bare metal. It is not meant to. On a PC, the main reason for writing fast code is the sheer amount of data that one wants to process. On a microcontroller, the data volume is probably small, but it might lead to catastrophic system failure, if these data are not processed in time, because the microcontroller is supposed to interact with the outside world in a timely fashion. In fact, this latter objective was the initiator of this project: I needed the Fourier transform of a signal coming from the ADC of the pyboard, and all available options were simply too slow. \n",
    "\n",
    "In addition to speed, another issue that one has to keep in mind when working with embedded systems is the amount of available RAM: I believe, everything here could be implemented in pure python with relatively little effort, but the price we would have to pay for that is not only speed, but RAM, too. python code, if is not frozen, and compiled into the firmware, has to be compiled at runtime, which is not exactly a cheap process. On top of that, if numbers are stored in a list or tuple, which would be the high-level container, then they occupy 8 bytes, no matter, whether they are all smaller than 100, or larger than one hundred million. This is obviously a waste of resources in an environment, where resources are scarce. \n",
    "\n",
-    "Finally, there is a reason for using micropython in the first place. Namely, that a microcontroller can be programmed in a very elegant, and *pythonic* way. But if it is so, why should we not extend this idea to other tasks and concepts that might come up in this context? If there was no other reason than this *elegance*, I would find that convincing enough.\n",
+    "Finally, there is a reason for using `micropython` in the first place. Namely, that a microcontroller can be programmed in a very elegant, and *pythonic* way. But if it is so, why should we not extend this idea to other tasks and concepts that might come up in this context? If there was no other reason than this *elegance*, I would find that convincing enough.\n",
    "\n",
-    "Based on the above-mentioned considerations, all functions are implemented in a way that \n",
+    "Based on the above-mentioned considerations, all functions in `ulab` are implemented in a way that \n",
    "\n",
    "1. conforms to `numpy` as much as possible\n",
    "2. is so frugal with RAM as possible,\n",
@ -500,12 +508,12 @@
    "The main points of `ulab` are \n",
    "\n",
    "- compact, iterable and slicable containers of numerical data in 1, and 2 dimensions (arrays and matrices). These containers support all the relevant unary and binary operators (e.g., `len`, ==, +, *, etc.)\n",
-    "- vectorised computations on micropython iterables and numerical arrays/matrices (in numpy-speak, universal functions)\n",
+    "- vectorised computations on micropython iterables and numerical arrays/matrices (in `numpy`-speak, universal functions)\n",
    "- basic linear algebra routines (matrix inversion, multiplication, reshaping, transposition, determinant, and eigenvalues)\n",
    "- polynomial fits to numerical data\n",
    "- fast Fourier transforms\n",
    "\n",
-    "At the time of writing this manual (for version 0.26), the library adds approximately 30 kB of extra compiled code to the micropython (pyboard.v.11) firmware. \n",
+    "At the time of writing this manual (for version 0.33.2), the library adds approximately 30 kB of extra compiled code to the micropython (pyboard.v.11) firmware. However, if you are tight with flash space, you can easily shave off a couple of kB. See the section on [customising ulab](#Custom_builds).\n",
    "\n",
    "## Resources and legal matters\n",
    "\n",
@ -515,11 +523,53 @@
    "\n",
    "## Friendly request\n",
    "\n",
-    "If you use `ulab`, and bump into a bug, or think that a particular function is missing, or its behaviour does not conform to `numpy`, please, raise an issue on github, so that the community can profit from your experiences. \n",
+    "If you use `ulab`, and bump into a bug, or think that a particular function is missing, or its behaviour does not conform to `numpy`, please, raise a [ulab issue](#https://github.com/v923z/micropython-ulab/issues) on github, so that the community can profit from your experiences. \n",
    "\n",
    "Even better, if you find the project useful, and think that it could be made better, faster, tighter, and shinier, please, consider contributing, and issue a pull request with the implementation of your improvements and new features. `ulab` can only become successful, if it offers what the community needs.\n",
    "\n",
-    "These last comments apply to the documentation, too. If, in your opinion, the documentation is obscure, misleading, or not detailed enough, please, let me know, so that *we* can fix it."
+    "These last comments apply to the documentation, too. If, in your opinion, the documentation is obscure, misleading, or not detailed enough, please, let me know, so that *we* can fix it.\n",
+    "\n",
+    "## Differences between micropython-ulab and circuitpython-ulab\n",
+    "\n",
+    "`ulab` has originally been developed for `micropython`, but has since been integrated into a number of its flavours. Most of these flavours are simply forks of `micropython` itself, with some additional functionality. One of the notable exceptions is `circuitpython`, which has slightly diverged at the core level, and this has some minor consequences. Some of these concern the C implementation details only, which all have been sorted out with the generous and enthusiastic support of Jeff Epler from [Adafruit Industries](http://www.adafruit.com).\n",
+    "\n",
+    "There are, however, a couple of instances, where the usage in the two environments is slightly different at the python level. These are how the packges can be imported, and how the class properties can be accessed. In both cases, the `circuitpython` implementation results in `numpy`-conform code. `numpy`-compatibility in `micropython` will be implemented as soon as `micropython` itself has the required tools. Till then we have to live with a workaround, which I will point out at the relevant places."
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "# Customising `ulab`\n",
+    "\n",
+    "`ulab` implements a great number of functions, which are organised in sub-modules. E.g., functions related to Fourier transforms are located in the `ulab.fft` sub-module, so you would import `fft` as\n",
+    "\n",
+    "```python\n",
+    "import ulab\n",
+    "from ulab import fft\n",
+    "```\n",
+    "by which point you can get the FFT of your data by calling `fft.fft(...)`. \n",
+    "\n",
+    "The idea of such grouping of functions and methods is to provide a means for granularity: It is quite possible that you do not need all functions in a particular application. If you want to save some flash space, you can easily exclude arbitrary sub-modules from the firmware. The [ulab.h](https://github.com/v923z/micropython-ulab/blob/master/code/ulab.h) header file contains a pre-processor flag for each sub-module. The default setting is 1 for each of them. Setting them to 0 removes the module from the compiled firmware. \n",
+    "\n",
+    "The first couple of lines of the file look like this\n",
+    "\n",
+    "```c\n",
+    "// vectorise (all functions) takes approx. 3 kB of flash space\n",
+    "#define ULAB_VECTORISE_MODULE (1)\n",
+    "\n",
+    "// linalg adds around 6 kB\n",
+    "#define ULAB_LINALG_MODULE (1)\n",
+    "\n",
+    "// poly is approx. 2.5 kB\n",
+    "#define ULAB_POLY_MODULE (1)\n",
+    "```\n",
+    "\n",
+    "In order to simplify navigation in the header, each flag begins with `ULAB_`, and continues with the name of the sub-module. This name is also the `.c` file, where the sub-module is implemented. So, e.g., the linear algebra routines can be found in `linalg.c`, and the corresponding compiler flag is `ULAB_LINALG_MODULE`. Each section displays a hint as to how much space you can save by un-setting the flag.\n",
+    "\n",
+    "At first, having to import everything in this way might appear to be overly complicated, but there is a very good reason behind all this: you can find out at the time of importing, whether a function or sub-module is part of your `ulab` firmware, or not. The alternative, namely, that you do not have to import anything beyond `ulab`, could prove catastrophic: you would learn only at run time (at the moment of calling the function in your code) that a particular function is not in the firmware, and that is most probably too late.\n",
+    "\n",
+    "Except for `fft`, the standard sub-modules, `vector`, `linalg`, `numerical`, `and poly`all `numpy`-compatible. User-defined functions that accept `ndarray`s as their argument should be implemented in the `extra` sub-module, or its sub-modules. Hints as to how to do that can be found in the section [Extending ulab](#Extending-ulab)."
   ]
  },
  {
@ -589,22 +639,20 @@
    "\n",
    "## Methods of ndarrays\n",
    "\n",
-    "[.shape](#.shape)\n",
+    "[.shape<sup>*</sup>](#.shape)\n",
+    "\n",
+    "[size<sup>*</sup>](#size)\n",
+    "\n",
+    "[itemsize<sup>*</sup>](#itemsize)\n",
    "\n",
    "[.reshape](#.reshape)\n",
    "\n",
-    "[.rawsize<sup>**</sup>](#.rawsize)\n",
-    "\n",
    "[.transpose](#.transpose)\n",
    "\n",
    "[.flatten<sup>**</sup>](#.flatten)\n",
    "\n",
-    "[.asbytearray](#.asbytearray)\n",
-    "\n",
    "## Matrix methods\n",
    "\n",
-    "[size](#size)\n",
-    "\n",
    "[inv](#inv)\n",
    "\n",
    "[dot](#dot)\n",
@ -659,7 +707,11 @@
    "\n",
    "[ifft<sup>**</sup>](#ifft)\n",
    "\n",
-    "[spectrum<sup>**</sup>](#spectrum)"
+    "[spectrum<sup>**</sup>](#spectrum)\n",
+    "\n",
+    "## Filter functions\n",
+    "\n",
+    "[convolve](#convolve)"
   ]
  },
  {
@ -668,13 +720,13 @@
   "source": [
    "# ndarray, the basic container\n",
    "\n",
-    "The `ndarray` is the underlying container of numerical data. It is derived from micropython's own `array` object, but has a great number of extra features starting with how it can be initialised, how operations can be done on it, and which functions can accept it as an argument.\n",
+    "The `ndarray` is the underlying container of numerical data. It is derived from micropython's own `array` object, but has a great number of extra features starting with how it can be initialised, which operations can be done on it, and which functions can accept it as an argument. One important property of an `ndarray` is that it is also a proper `micropython` iterable.\n",
    "\n",
-    "Since the `ndarray` is a binary container, it is also compact, meaning that it takes only a couple of bytes of extra RAM in addition to what is required for storing the numbers themselves. `ndarray`s are also type-aware, i.e., one can save RAM by specifying a data type, and using the smallest reasonable one. Five such types are defined, namely `uint8`, `int8`, which occupy a single byte of memory per datum, `uint16`, and `int16`, which occupy two bytes per datum, and `float`, which occupies four or eight bytes per datum. The precision/size of the `float` type depends on the definition of `mp_float_t`. Some platforms, e.g., the PYBD, implement `double`s, but some, e.g., the pyboard.v.11, don't. You can find out, what type of float your particular platform implements by looking at the output of the [.rawsize](#.rawsize) class method.\n",
+    "Since the `ndarray` is a binary container, it is also compact, meaning that it takes only a couple of bytes of extra RAM in addition to what is required for storing the numbers themselves. `ndarray`s are also type-aware, i.e., one can save RAM by specifying a data type, and using the smallest reasonable one. Five such types are defined, namely `uint8`, `int8`, which occupy a single byte of memory per datum, `uint16`, and `int16`, which occupy two bytes per datum, and `float`, which occupies four or eight bytes per datum. The precision/size of the `float` type depends on the definition of `mp_float_t`. Some platforms, e.g., the PYBD, implement `double`s, but some, e.g., the pyboard.v.11, don't. You can find out, what type of float your particular platform implements by looking at the output of the [.itemsize](#.itemsize) class property.\n",
    "\n",
    "On the following pages, we will see how one can work with `ndarray`s. Those familiar with `numpy` should find that the nomenclature and naming conventions of `numpy` are adhered to as closely as possible. I will point out the few differences, where necessary.\n",
    "\n",
-    "For the sake of comparison, in addition to `ulab` code snippets, sometimes the equivalent `numpy` code is also presented. You can find out, where the snippet is supposed to run by looking at its first line, the header.\n",
+    "For the sake of comparison, in addition to the `ulab` code snippets, sometimes the equivalent `numpy` code is also presented. You can find out, where the snippet is supposed to run by looking at its first line, the header.\n",
    "\n",
    "Hint:  you can easily port existing `numpy` code, if you `import ulab as np`."
   ]
@ -842,16 +894,18 @@
   "source": [
    "### .shape\n",
    "\n",
-    "The `.shape` method returns a 2-tuple with the number of rows, and columns."
+    "The `.shape` method (property) returns a 2-tuple with the number of rows, and columns. \n",
+    "\n",
+    "**WARNING:** In `circuitpython`, you can call the method as a property, i.e., "
   ]
  },
  {
   "cell_type": "code",
-   "execution_count": 283,
+   "execution_count": 12,
   "metadata": {
    "ExecuteTime": {
-     "end_time": "2019-10-16T15:30:33.810628Z",
-     "start_time": "2019-10-16T15:30:33.796088Z"
+     "end_time": "2020-02-11T19:01:40.377272Z",
+     "start_time": "2020-02-11T19:01:40.364448Z"
    }
   },
   "outputs": [
@ -879,13 +933,261 @@
    "\n",
    "a = np.array([1, 2, 3, 4], dtype=np.int8)\n",
    "print(\"a:\\n\", a)\n",
-    "print(\"shape of a:\", a.shape())\n",
+    "print(\"shape of a:\", a.shape)\n",
+    "\n",
+    "b= np.array([[1, 2], [3, 4]], dtype=np.int8)\n",
+    "print(\"\\nb:\\n\", b)\n",
+    "print(\"shape of b:\", b.shape)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "**WARNING:** On the other hand, since properties are not implemented in `micropython`, there you would call the method as a function, i.e., "
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 12,
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2020-02-11T19:01:40.377272Z",
+     "start_time": "2020-02-11T19:01:40.364448Z"
+    }
+   },
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "a:\n",
+      " array([1, 2, 3, 4], dtype=int8)\n",
+      "shape of a: (1, 4)\n",
+      "\n",
+      "b:\n",
+      " array([[1, 2],\n",
+      "\t [3, 4]], dtype=int8)\n",
+      "shape of b: (2, 2)\n",
+      "\n",
+      "\n"
+     ]
+    }
+   ],
+   "source": [
+    "%%micropython -unix 1\n",
+    "\n",
+    "import ulab as np\n",
+    "\n",
+    "a = np.array([1, 2, 3, 4], dtype=np.int8)\n",
+    "print(\"a:\\n\", a)\n",
+    "print(\"shape of a:\", a.shape)\n",
    "\n",
    "b= np.array([[1, 2], [3, 4]], dtype=np.int8)\n",
    "print(\"\\nb:\\n\", b)\n",
    "print(\"shape of b:\", b.shape())"
   ]
  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "### .size\n",
+    "\n",
+    "The `.size` method (property) returns an integer with the number of elements in the array. \n",
+    "\n",
+    "**WARNING:** In `circuitpython`, the `numpy` nomenclature applies, i.e., "
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 10,
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2020-02-11T06:32:22.721112Z",
+     "start_time": "2020-02-11T06:32:22.713111Z"
+    }
+   },
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "a:\n",
+      " array([1, 2, 3], dtype=int8)\n",
+      "size of a: 3\n",
+      "\n",
+      "b:\n",
+      " array([[1, 2],\n",
+      "\t [3, 4]], dtype=int8)\n",
+      "size of b: 4\n",
+      "\n",
+      "\n"
+     ]
+    }
+   ],
+   "source": [
+    "%%micropython -unix 1\n",
+    "\n",
+    "import ulab as np\n",
+    "\n",
+    "a = np.array([1, 2, 3], dtype=np.int8)\n",
+    "print(\"a:\\n\", a)\n",
+    "print(\"size of a:\", a.size)\n",
+    "\n",
+    "b= np.array([[1, 2], [3, 4]], dtype=np.int8)\n",
+    "print(\"\\nb:\\n\", b)\n",
+    "print(\"size of b:\", b.size)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "**WARNING:** In `micropython`, `size` is a method, i.e., "
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 10,
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2020-02-11T06:32:22.721112Z",
+     "start_time": "2020-02-11T06:32:22.713111Z"
+    }
+   },
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "a:\n",
+      " array([1, 2, 3], dtype=int8)\n",
+      "size of a: 3\n",
+      "\n",
+      "b:\n",
+      " array([[1, 2],\n",
+      "\t [3, 4]], dtype=int8)\n",
+      "size of b: 4\n",
+      "\n",
+      "\n"
+     ]
+    }
+   ],
+   "source": [
+    "%%micropython -unix 1\n",
+    "\n",
+    "import ulab as np\n",
+    "\n",
+    "a = np.array([1, 2, 3], dtype=np.int8)\n",
+    "print(\"a:\\n\", a)\n",
+    "print(\"size of a:\", a.size)\n",
+    "\n",
+    "b= np.array([[1, 2], [3, 4]], dtype=np.int8)\n",
+    "print(\"\\nb:\\n\", b)\n",
+    "print(\"size of b:\", b.size())"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "### .itemsize\n",
+    "\n",
+    "The `.itemsize` method (property) returns an integer with the siz enumber of elements in the array.\n",
+    "\n",
+    "**WARNING:** In `circuitpython`:"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 13,
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2020-02-11T19:05:04.296601Z",
+     "start_time": "2020-02-11T19:05:04.280669Z"
+    }
+   },
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "a:\n",
+      " array([1, 2, 3], dtype=int8)\n",
+      "itemsize of a: 1\n",
+      "\n",
+      "b:\n",
+      " array([[1.0, 2.0],\n",
+      "\t [3.0, 4.0]], dtype=float)\n",
+      "itemsize of b: 8\n",
+      "\n",
+      "\n"
+     ]
+    }
+   ],
+   "source": [
+    "%%micropython -unix 1\n",
+    "\n",
+    "import ulab as np\n",
+    "\n",
+    "a = np.array([1, 2, 3], dtype=np.int8)\n",
+    "print(\"a:\\n\", a)\n",
+    "print(\"itemsize of a:\", a.itemsize)\n",
+    "\n",
+    "b= np.array([[1, 2], [3, 4]], dtype=np.float)\n",
+    "print(\"\\nb:\\n\", b)\n",
+    "print(\"itemsize of b:\", b.itemsize)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "**WARNING:** In `micropython`:"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 13,
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2020-02-11T19:05:04.296601Z",
+     "start_time": "2020-02-11T19:05:04.280669Z"
+    }
+   },
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "a:\n",
+      " array([1, 2, 3], dtype=int8)\n",
+      "itemsize of a: 1\n",
+      "\n",
+      "b:\n",
+      " array([[1.0, 2.0],\n",
+      "\t [3.0, 4.0]], dtype=float)\n",
+      "itemsize of b: 8\n",
+      "\n",
+      "\n"
+     ]
+    }
+   ],
+   "source": [
+    "%%micropython -unix 1\n",
+    "\n",
+    "import ulab as np\n",
+    "\n",
+    "a = np.array([1, 2, 3], dtype=np.int8)\n",
+    "print(\"a:\\n\", a)\n",
+    "print(\"itemsize of a:\", a.itemsize)\n",
+    "\n",
+    "b= np.array([[1, 2], [3, 4]], dtype=np.float)\n",
+    "print(\"\\nb:\\n\", b)\n",
+    "print(\"itemsize of b:\", b.itemsize())"
+   ]
+  },
  {
   "cell_type": "markdown",
   "metadata": {},
@ -934,54 +1236,6 @@
    "print('a (1 by 16):', a.reshape((1, 16)))"
   ]
  },
-  {
-   "cell_type": "markdown",
-   "metadata": {},
-   "source": [
-    "### .rawsize\n",
-    "\n",
-    "The `rawsize` method of the `ndarray` returns a 5-tuple with the following data\n",
-    "\n",
-    "1. number of rows\n",
-    "2. number of columns\n",
-    "3. length of the storage (should be equal to the product of 1. and 2.)\n",
-    "4. length of the data storage in bytes \n",
-    "5. datum size in bytes (1 for `uint8`/`int8`, 2 for `uint16`/`int16`, and 4, or 8 for `floats`, see [ndarray, the basic container](#ndarray,-the-basic-container))\n",
-    "\n",
-    "**WARNING:** `rawsize` is a `ulab`-only method; it has no equivalent in `numpy`."
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 510,
-   "metadata": {
-    "ExecuteTime": {
-     "end_time": "2019-10-19T17:44:26.983908Z",
-     "start_time": "2019-10-19T17:44:26.764912Z"
-    }
-   },
-   "outputs": [
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "a: \t\t array([1.0, 2.0, 3.0, 4.0], dtype=float)\n",
-      "rawsize of a: \t (1, 4, 4, 16, 4)\n",
-      "\n",
-      "\n"
-     ]
-    }
-   ],
-   "source": [
-    "%%micropython -unix 1\n",
-    "\n",
-    "import ulab as np\n",
-    "\n",
-    "a = np.array([1, 2, 3, 4], dtype=np.float)\n",
-    "print(\"a: \\t\\t\", a)\n",
-    "print(\"rawsize of a: \\t\", a.rawsize())"
-   ]
-  },
  {
   "cell_type": "markdown",
   "metadata": {},
@ -1035,109 +1289,6 @@
    "print(\"b flattened (F): \\t\", b.flatten(order='F'))"
   ]
  },
-  {
-   "cell_type": "markdown",
-   "metadata": {},
-   "source": [
-    "### .asbytearray\n",
-    "\n",
-    "The contents of an `ndarray` can be accessed directly by calling the `.asbytearray` method. This will simply return a pointer to the underlying flat `array` object, which can then be manipulated directly.\n",
-    "\n",
-    "**WARNING:** `asbytearray` is a `ulab`-only method; it has no equivalent in `numpy`.\n",
-    "\n",
-    "In the example below, note the difference between `a`, and `buffer`: while both are designated as an array, you recognise the micropython array from the fact that it prints the typecode (`b` in this particular case). The `ndarray`, on the other hand, prints out the `dtype` (`int8` here)."
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 211,
-   "metadata": {
-    "ExecuteTime": {
-     "end_time": "2019-11-01T14:25:39.008074Z",
-     "start_time": "2019-11-01T14:25:38.876546Z"
-    }
-   },
-   "outputs": [
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "a:  array([1, 2, 3, 4], dtype=int8)\n",
-      "array content: array('b', [1, 2, 3, 4])\n",
-      "array content: array('b', [1, 123, 3, 4])\n",
-      "\n",
-      "\n"
-     ]
-    }
-   ],
-   "source": [
-    "%%micropython -unix 1\n",
-    "\n",
-    "import ulab as np\n",
-    "\n",
-    "a = np.array([1, 2, 3, 4], dtype=np.int8)\n",
-    "print('a: ', a)\n",
-    "buffer = a.asbytearray()\n",
-    "print(\"array content:\", buffer)\n",
-    "buffer[1] = 123\n",
-    "print(\"array content:\", buffer)"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "metadata": {},
-   "source": [
-    "This in itself wouldn't be very interesting, but since `buffer` is a proper micropython `array`, we can pass it to functions that can employ the buffer protocol. E.g., all the `ndarray` facilities can be applied to the results of timed ADC conversions."
-   ]
-  },
-  {
-   "cell_type": "code",
-   "execution_count": 563,
-   "metadata": {
-    "ExecuteTime": {
-     "end_time": "2019-10-20T07:29:00.153589Z",
-     "start_time": "2019-10-20T07:28:50.210383Z"
-    }
-   },
-   "outputs": [
-    {
-     "name": "stdout",
-     "output_type": "stream",
-     "text": [
-      "ADC results:\t array([48, 2, 2, ..., 0, 0, 0], dtype=uint8)\n",
-      "mean of results:\t 1.22\n",
-      "std of results:\t 4.744639\n",
-      "\n"
-     ]
-    }
-   ],
-   "source": [
-    "%%micropython -pyboard 1\n",
-    "\n",
-    "import pyb\n",
-    "import ulab as np\n",
-    "\n",
-    "n = 100\n",
-    "\n",
-    "adc = pyb.ADC(pyb.Pin.board.X19)\n",
-    "tim = pyb.Timer(6, freq=10)\n",
-    "\n",
-    "a = np.array([0]*n, dtype=np.uint8)\n",
-    "buffer = a.asbytearray()\n",
-    "adc.read_timed(buffer, tim)\n",
-    "\n",
-    "print(\"ADC results:\\t\", a)\n",
-    "print(\"mean of results:\\t\", np.mean(a))\n",
-    "print(\"std of results:\\t\", np.std(a))"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "metadata": {},
-   "source": [
-    "Likewise, data can be read directly into `ndarray`s from other interfaces, e.g., SPI, I2C etc, and also, by laying bare the `ndarray`, we can pass results of `ulab` computations to anything that can read from a buffer."
-   ]
-  },
  {
   "cell_type": "markdown",
   "metadata": {},
@ -3988,11 +4139,11 @@
  },
  {
   "cell_type": "code",
-   "execution_count": 458,
+   "execution_count": 114,
   "metadata": {
    "ExecuteTime": {
-     "end_time": "2019-10-19T13:07:43.168629Z",
-     "start_time": "2019-10-19T13:07:43.130341Z"
+     "end_time": "2020-02-16T18:38:07.294862Z",
+     "start_time": "2020-02-16T18:38:07.233842Z"
    }
   },
   "outputs": [
@ -4000,13 +4151,13 @@
     "name": "stdout",
     "output_type": "stream",
     "text": [
-      "real part:\t array([5119.996, -5.004663, -5.004798, ..., -5.005482, -5.005643, -5.006577], dtype=float)\n",
-      "\n",
-      "imaginary part:\t array([0.0, 1631.333, 815.659, ..., -543.764, -815.6588, -1631.333], dtype=float)\n",
-      "\n",
-      "real part:\t array([5119.996, -5.004663, -5.004798, ..., -5.005482, -5.005643, -5.006577], dtype=float)\n",
-      "\n",
-      "imaginary part:\t array([0.0, 1631.333, 815.659, ..., -543.764, -815.6588, -1631.333], dtype=float)\n",
+      "real part:\t array([5119.996, -5.004663, -5.004798, ..., -5.005482, -5.005643, -5.006577], dtype=float)\r\n",
+      "\r\n",
+      "imaginary part:\t array([0.0, 1631.333, 815.659, ..., -543.764, -815.6588, -1631.333], dtype=float)\r\n",
+      "\r\n",
+      "real part:\t array([5119.996, -5.004663, -5.004798, ..., -5.005482, -5.005643, -5.006577], dtype=float)\r\n",
+      "\r\n",
+      "imaginary part:\t array([0.0, 1631.333, 815.659, ..., -543.764, -815.6588, -1631.333], dtype=float)\r\n",
      "\n"
     ]
    }
@ -4015,16 +4166,20 @@
    "%%micropython -pyboard 1\n",
    "\n",
    "import ulab as np\n",
+    "from ulab import numerical\n",
+    "from ulab import vector\n",
+    "from ulab import fft\n",
+    "from ulab import linalg\n",
    "\n",
-    "x = np.linspace(0, 10, num=1024)\n",
-    "y = np.sin(x)\n",
-    "z = np.zeros(len(x))\n",
+    "x = numerical.linspace(0, 10, num=1024)\n",
+    "y = vector.sin(x)\n",
+    "z = linalg.zeros(len(x))\n",
    "\n",
-    "a, b = np.fft(x)\n",
+    "a, b = fft.fft(x)\n",
    "print('real part:\\t', a)\n",
    "print('\\nimaginary part:\\t', b)\n",
    "\n",
-    "c, d = np.fft(x, z)\n",
+    "c, d = fft.fft(x, z)\n",
    "print('\\nreal part:\\t', c)\n",
    "print('\\nimaginary part:\\t', d)"
   ]
@ -4268,13 +4423,65 @@
    "where $N$ is the length of $y_1$, and $Y_1, Y_2$, and $Y$, respectively, are the Fourier transforms of $y_1, y_2$, and $y = y_1 + iy_2$."
   ]
  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "# Filter routines"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "numpy: https://docs.scipy.org/doc/numpy/reference/generated/numpy.convolve.html\n",
+    "\n",
+    "## convolve\n",
+    "Returns the discrete, linear convolution of two one-dimensional sequences.\n",
+    "\n",
+    "Only the ``full`` mode is supported, and the ``mode`` named parameter is not accepted. Note that all other modes can be had by slicing a ``full`` result."
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 5,
+   "metadata": {
+    "ExecuteTime": {
+     "end_time": "2020-02-10T18:46:06.538207Z",
+     "start_time": "2020-02-10T18:46:06.525851Z"
+    }
+   },
+   "outputs": [
+    {
+     "name": "stdout",
+     "output_type": "stream",
+     "text": [
+      "array([1.0, 12.0, 123.0, 1230.0, 2300.0, 3000.0], dtype=float)\n",
+      "\n",
+      "\n"
+     ]
+    }
+   ],
+   "source": [
+    "%%micropython -unix 1\n",
+    "\n",
+    "import ulab as np\n",
+    "\n",
+    "x = np.array((1,2,3))\n",
+    "y = np.array((1,10,100,1000))\n",
+    "\n",
+    "print(np.convolve(x, y))"
+   ]
+  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Extending ulab\n",
    "\n",
-    "New functions can easily be added to `ulab` in a couple of simple steps. At the C level, the type definition of an `ndarray` is as follows:\n",
+    "As mentioned at the beginning, `ulab` is a set of sub-modules, out of which one, `extra` is explicitly reserved for user code. You should implement your functions in this sub-module, or sub-modules thereof.\n",
+    "\n",
+    "The new functions can easily be added to `extra` in a couple of simple steps. At the C level, the type definition of an `ndarray` is as follows:\n",
    "\n",
    "```c\n",
    "typedef struct _ndarray_obj_t {\n",
@ -4313,7 +4520,7 @@
    "    NDARRAY_FLOAT = 'd',\n",
    "};\n",
    "```\n",
-    "The ambiguity is caused by the fact that not all platforms implement `double`, and there one has to take `float`s. But you haven't actually got to be concerned by this, because at the very beginning of `ndarray.h`, this is already taken care of: the preprocessor figures out, what the `float` implementation of the hardware platform is, and defines the  `NDARRAY_FLOAT` typecode accordingly. All you have to keep in mind is that wherever you would use `float` or `double`, you have to use `mp_float_t`. That type is defined in `py/mpconfig.h` of the micropython code base.\n",
+    "The ambiguity is caused by the fact that not all platforms implement `double`, and there one has to take `float`s. But you haven't actually got to be concerned by this, because at the very beginning of `ndarray.h`, this is already taken care of: the pre-processor figures out, what the `float` implementation of the hardware platform is, and defines the  `NDARRAY_FLOAT` typecode accordingly. All you have to keep in mind is that wherever you would use `float` or `double`, you have to use `mp_float_t`. That type is defined in `py/mpconfig.h` of the micropython code base.\n",
    "\n",
    "Therefore, a 4-by-5 matrix of type float can be created as\n",
    "\n",
@ -4458,46 +4665,13 @@
   "cell_type": "markdown",
   "metadata": {},
   "source": [
-    "In the boilerplate above, we cast the pointer to `array->items` to the required type. There are certain operations, however, when you do not need the casting. If you do not want to change the array's values, only their position within the array, you can get away with copying the memory content, regardless the type. A good example for such a scenario is the transpose function in https://github.com/v923z/micropython-ulab/blob/master/code/linalg.c."
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "metadata": {},
-   "source": [
-    "## Compiling your module\n",
-    "\n",
-    "Once you have implemented the functionality you wanted, you have to include the source code in the make file by adding it to `micropython.mk`:\n",
-    "\n",
-    "```makefile\n",
-    "USERMODULES_DIR := $(USERMOD_DIR)\n",
-    "\n",
-    "# Add all C files to SRC_USERMOD.\n",
-    "SRC_USERMOD += $(USERMODULES_DIR)/ndarray.c\n",
-    "SRC_USERMOD += $(USERMODULES_DIR)/linalg.c\n",
-    "SRC_USERMOD += $(USERMODULES_DIR)/vectorise.c\n",
-    "SRC_USERMOD += $(USERMODULES_DIR)/poly.c\n",
-    "SRC_USERMOD += $(USERMODULES_DIR)/fft.c\n",
-    "SRC_USERMOD += $(USERMODULES_DIR)/numerical.c\n",
-    "SRC_USERMOD += $(USERMODULES_DIR)/ulab.c\n",
-    "\n",
-    "SRC_USERMOD += $(USERMODULES_DIR)/your_module.c\n",
-    "\n",
-    "CFLAGS_USERMOD += -I$(USERMODULES_DIR)\n",
-    "```"
-   ]
-  },
-  {
-   "cell_type": "markdown",
-   "metadata": {},
-   "source": [
-    "In addition, you also have to add the function objects to `ulab.c`, and create a `QString` for the function name:\n",
+    "Once the function implementation is done, you should add the function object to the globals dictionary of the `extra` sub-module as \n",
    "\n",
    "```c\n",
    "...\n",
    "    MP_DEFINE_CONST_FUN_OBJ_1(useless_function_obj, userless_function);\n",
    "...\n",
-    "    STATIC const mp_map_elem_t ulab_globals_table[] = {\n",
+    "    STATIC const mp_map_elem_t extra_globals_table[] = {\n",
    "...\n",
    "    { MP_OBJ_NEW_QSTR(MP_QSTR_useless), (mp_obj_t)&useless_function_obj },\n",
    "...\n",
@ -4509,29 +4683,21 @@
    "```c\n",
    "    MP_DEFINE_CONST_FUN_OBJ_1(useless_function_obj, userless_function);\n",
    "```\n",
-    "depends naturally on what exactly you implemented, i.e., how many arguments the function takes, whether only positional arguments allowed and so on. For a thorough discussion on how function objects have to be defined, see, e.g., the [user module programming manual](#https://micropython-usermod.readthedocs.io/en/latest/)."
+    "depends naturally on what exactly you implemented, i.e., how many arguments the function takes, whether only positional arguments allowed and so on. For a thorough discussion on how function objects have to be defined, see, e.g., the [user module programming manual](#https://micropython-usermod.readthedocs.io/en/latest/).\n",
+    "\n",
+    "And with that, you are done. You can simply call the compiler as \n",
+    "\n",
+    "```bash\n",
+    "make BOARD=PYBV11 USER_C_MODULES=../../../ulab all\n",
+    "```\n",
+    "and the rest is taken care of."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
-    "At this point, you should be able to compile the module with your extension by running `make` on the command line\n",
-    "\n",
-    "```bash\n",
-    "make USER_C_MODULES=../../../ulab all\n",
-    "```\n",
-    "for the unix port, and \n",
-    "\n",
-    "```bash\n",
-    "make BOARD=PYBV11 CROSS_COMPILE=<arm_tools_path>/bin/arm-none-eabi- USER_C_MODULES=../../../ulab all\n",
-    "```\n",
-    "for the `pyboard`, provided that the you have defined \n",
-    "\n",
-    "```makefile\n",
-    "#define MODULE_ULAB_ENABLED (1)\n",
-    "```\n",
-    "somewhere in `micropython/port/unix/mpconfigport.h`, or `micropython/stm32/mpconfigport.h`, respectively."
+    "In the boilerplate above, we cast the pointer to `array->items` to the required type. There are certain operations, however, when you do not need the casting. If you do not want to change the array's values, only their position within the array, you can get away with copying the memory content, regardless the type. A good example for such a scenario is the transpose function in https://github.com/v923z/micropython-ulab/blob/master/code/linalg.c."
   ]
  }
 ],
--- a/docs/ulab.ipynb
+++ b/docs/ulab.ipynb
--- a/tests/00smoke.py
+++ b/tests/00smoke.py
@ -0,0 +1,2 @@
+import ulab
+print(ulab.eye(3))
--- a/tests/00smoke.py.exp
+++ b/tests/00smoke.py.exp
@ -0,0 +1,3 @@
+array([[1.0, 0.0, 0.0],
+	 [0.0, 1.0, 0.0],
+	 [0.0, 0.0, 1.0]], dtype=float)
--- a/tests/constructors.py
+++ b/tests/constructors.py
@ -0,0 +1,12 @@
+from ulab import linalg
+import ulab
+print(ulab.ones(3))
+print(ulab.ones((2,3)))
+print(ulab.zeros(3))
+print(ulab.zeros((2,3)))
+print(ulab.eye(3))
+print(ulab.ones(1, dtype=ulab.int8))
+print(ulab.ones(2, dtype=ulab.uint8))
+print(ulab.ones(3, dtype=ulab.int16))
+print(ulab.ones(4, dtype=ulab.uint16))
+print(ulab.ones(5, dtype=ulab.float))
--- a/tests/constructors.py.exp
+++ b/tests/constructors.py.exp
@ -0,0 +1,14 @@
+array([1.0, 1.0, 1.0], dtype=float)
+array([[1.0, 1.0, 1.0],
+	 [1.0, 1.0, 1.0]], dtype=float)
+array([0.0, 0.0, 0.0], dtype=float)
+array([[0.0, 0.0, 0.0],
+	 [0.0, 0.0, 0.0]], dtype=float)
+array([[1.0, 0.0, 0.0],
+	 [0.0, 1.0, 0.0],
+	 [0.0, 0.0, 1.0]], dtype=float)
+array([1], dtype=int8)
+array([1, 1], dtype=uint8)
+array([1, 1, 1], dtype=int16)
+array([1, 1, 1, 1], dtype=uint16)
+array([1.0, 1.0, 1.0, 1.0, 1.0], dtype=float)
--- a/tests/operators.py
+++ b/tests/operators.py
@ -0,0 +1,20 @@
+import ulab
+a = ulab.ones(3)
+print(a+a)
+print(a-a)
+print(a*a)
+print(a/a)
+print(a+2)
+print(a-2)
+print(a*2)
+print(a/2)
+print(a<1)
+print(a<2)
+print(a<=0)
+print(a<=1)
+print(a>1)
+print(a>2)
+print(a>=0)
+print(a>=1)
+#print(a==0)  # These print just true or false.  Is it right?  is it a micropython limitation?
+#print(a==1)
--- a/tests/operators.py.exp
+++ b/tests/operators.py.exp
@ -0,0 +1,16 @@
+array([2.0, 2.0, 2.0], dtype=float)
+array([0.0, 0.0, 0.0], dtype=float)
+array([1.0, 1.0, 1.0], dtype=float)
+array([1.0, 1.0, 1.0], dtype=float)
+array([3.0, 3.0, 3.0], dtype=float)
+array([-1.0, -1.0, -1.0], dtype=float)
+array([2.0, 2.0, 2.0], dtype=float)
+array([0.5, 0.5, 0.5], dtype=float)
+[False, False, False]
+[True, True, True]
+[False, False, False]
+[True, True, True]
+[False, False, False]
+[False, False, False]
+[True, True, True]
+[True, True, True]
Author	SHA1	Message	Date
Jeff Epler	fbf98c16db	Merge pull request #11 from jepler/merge-upstream-4 Merge upstream 4	2020-02-27 14:09:18 -06:00
Jeff Epler	42d831e1e6	Merge remote-tracking branch 'upstream/master' into merge-upstream-4	2020-02-27 14:08:15 -06:00
Jeff Epler	66b89de8c7	Always include creation functions	2020-02-27 14:07:04 -06:00
Jeff Epler	844b85018b	Merge pull request #49 from jepler/gitignore ignore files created by ./build.sh	2020-02-27 13:58:25 -06:00
Jeff Epler	1c1a693a2b	Merge pull request #48 from v923z/create created new create sub-module for ndarray initialisation functions	2020-02-27 13:58:14 -06:00
Jeff Epler	f81e950513	ignore files created by ./build.sh	2020-02-27 13:56:09 -06:00
Jeff Epler	ffff7606c8	fix tests after 'eye' was moved	2020-02-27 13:53:29 -06:00
Zoltán Vörös	bee25781b9	added new source file...	2020-02-27 20:46:52 +01:00
Zoltán Vörös	47bf2ec9a7	created new create sub-module for ndarray initialisation functions	2020-02-27 20:39:13 +01:00
Jeff Epler	6860448096	Merge pull request #10 from jepler/merge-upstream-3 Merge from upstream	2020-02-27 10:16:25 -06:00
Jeff Epler	6b9ea24b2e	remove unused and unneeded file	2020-02-27 10:14:53 -06:00
Jeff Epler	a5ac422f33	Merge remote-tracking branch 'upstream/master' into merge-upstream-3	2020-02-27 10:14:40 -06:00
Jeff Epler	badeee48df	Merge pull request #47 from jepler/local-build-script a script to build and run tests locally	2020-02-27 10:12:03 -06:00
Jeff Epler	e370e56a15	a script to build and run tests locally	2020-02-27 10:11:34 -06:00
Jeff Epler	db5f1f85bb	Merge pull request #46 from jepler/ndarray-properties-sort Ndarray properties sort	2020-02-27 10:11:17 -06:00
Jeff Epler	2bddc94df5	Merge pull request #45 from jepler/move-ones-zeros Move zeros(), ones() to base ulab module	2020-02-27 10:08:29 -06:00
Jeff Epler	aa5ef4afb9	Enable sort method in circuitpython	2020-02-27 10:06:33 -06:00
Jeff Epler	d99d834d87	Enable properties in circuitpython I verified that these work for us as coded.	2020-02-27 10:06:27 -06:00
Jeff Epler	83479f115b	Move zeros(), ones() to base ulab module	2020-02-27 10:05:50 -06:00
Jeff Epler	924dc7012a	Merge pull request #44 from jepler/circuitpy-fixes Circuitpy fixes	2020-02-27 09:29:43 -06:00
Jeff Epler	daaacac16f	Remove CIRCUITPY special cases	2020-02-27 08:56:07 -06:00
Jeff Epler	aa4d53e292	Use circuitpy-compat for none	2020-02-27 08:56:04 -06:00
Jeff Epler	3febd79aa0	Merge pull request #41 from v923z/2dim Split ulab into multiple modules	2020-02-26 11:29:17 -06:00
Zoltán Vörös	7e2be88dff	Merge branch '2dim' of github.com:v923z/micropython-ulab into 2dim added circuitpython-related stuff to code and manual	2020-02-26 18:06:19 +01:00
Zoltán Vörös	e0e840f6d5	added circuitpython-related stuff to the manual	2020-02-26 18:05:49 +01:00
Jeff Epler	102ba5032e	add missing expected-file	2020-02-18 21:40:31 -06:00
Jeff Epler	b83ed3e2ca	add more tests	2020-02-18 21:36:49 -06:00
Jeff Epler	1e5ebe739d	numerical: add __name__	2020-02-18 21:27:05 -06:00
Jeff Epler	70af1c8a77	fix test	2020-02-18 21:25:03 -06:00
Jeff Epler	a2962f0fe5	fix printing failure info when tests fail	2020-02-18 21:10:33 -06:00
Jeff Epler	cf057c4df9	roll back constness corrections	2020-02-18 21:00:30 -06:00
Jeff Epler	de6b7772e4	fix ci build	2020-02-18 20:50:01 -06:00
Jeff Epler	43551c9a7a	Merge remote-tracking branch 'upstream/master' into 2dim	2020-02-18 20:46:43 -06:00
Jeff Epler	23adf6e4a7	Mark modules as 'extern const'	2020-02-18 20:32:33 -06:00
Jeff Epler	f6d123beb8	Don't export modules if CIRCUITPY .. We are using the "shared bindings" method of circuitpython	2020-02-18 20:32:28 -06:00
Jeff Epler	27996c9003	Add MP_DELCARE_CONST_FUN declarations to headers	2020-02-18 20:32:22 -06:00
Zoltán Vörös	bec24fe4a4	Merge branch 'master' of github.com:v923z/micropython-ulab	2020-02-17 19:58:27 +01:00
Zoltán Vörös	e71f667114	temporary fix for issue #40	2020-02-17 19:57:55 +01:00
Zoltán Vörös	8300de7f11	backup commit, absolutely nothing essential	2020-02-16 19:53:30 +01:00
Zoltán Vörös	017c1c2c46	added extras.h/extras.c	2020-02-16 19:49:28 +01:00
Zoltán Vörös	6fe015f134	properties are now defined in ndarray_properties.h	2020-02-16 19:49:01 +01:00
Zoltán Vörös	3727c38182	Merge pull request #39 from jepler/github-actions Perform continuous integration using github actions	2020-02-16 16:22:14 +01:00
Jeff Epler	a75903efe5	Add github actions This will build micropython's master branch with ulab support, and then run the tests in tests/ At this time, there's only one test and it's not very useful.	2020-02-15 19:57:50 -06:00
Zoltán Vörös	d8bfe46bd8	added __name__ to all submodules	2020-02-14 19:51:28 +01:00
Zoltán Vörös	2e3a0b4483	separated sub-modules into proper python sub-modules	2020-02-13 21:49:09 +01:00
Jeff Epler	cf0180e05e	Merge pull request #9 from jepler/merge-upstream-2 Merge upstream	2020-02-13 13:31:11 -06:00
Jeff Epler	4e5947afe1	Merge remote-tracking branch 'upstream/master'	2020-02-13 13:27:54 -06:00
Zoltán Vörös	49db707a9f	Merge pull request #37 from jepler/portability Make portable to CircuitPython within same codebase	2020-02-12 18:27:16 +01:00
Zoltán Vörös	f2aaab84cc	Merge pull request #36 from jepler/convolve-optimize filter_convolve: fix build error	2020-02-12 18:24:26 +01:00
Jeff Epler	fc80a25685	Increase CircuitPython compatibility - Adapt to signature of mp_make_new_fun_t for mpy and cpy by ifdef - Use MP_OBJ_IS_TYPE instead of mp_obj_is_type - Ditto MP_OBJ_IS_INT - Use mp_const_none instead of MP_ROM_NONE - Ditto mp_const_true, mp_const_false	2020-02-12 10:15:47 -06:00
Jeff Epler	f47abf90ac	filter_convolve: fix build error My earlier change introduced a build error. I'm not sure why I didn't find this before making the pull request.	2020-02-12 09:50:17 -06:00
Zoltán Vörös	5bf6e89a0a	Merge pull request #35 from jepler/convolve-optimize convolve: Optimize and special-case floats	2020-02-12 08:16:18 +01:00
Jeff Epler	7846b0c469	convolve: Optimize and special-case floats Special casing floats decreases runtime to about 50% (applying a 117-tap filter to 512 points of data goes from 70ms to 32ms) The top_n/bot_n calculations already meant that the a/c indices were never out of range. This decreases runtime further to about 15% of original (11ms) Timings done on an Adafruit Clue (nrf52840 at 64MHz) It does of course increase code size somewhat.	2020-02-11 17:57:54 -06:00
Zoltán Vörös	9153fd8f8a	added a short section to the manual on how to customise ulab	2020-02-11 21:36:12 +01:00
Zoltán Vörös	57cf52838c	Merge pull request #33 from jepler/fix-undef-errors-mpy Fix some define-guards	2020-02-11 20:19:27 +01:00
Zoltán Vörös	c14eee1bd4	trying to fix ulab.h definitions	2020-02-11 20:15:45 +01:00
Zoltán Vörös	2c71467ced	implemented ndarray properties	2020-02-11 20:08:37 +01:00
Jeff Epler	02d74a4d3e	Fix some define-guards These problems were found building in circuitpython: ../../extmod/ulab/code/numerical.c:671:5: error: "ULAB_NUMERICAL_ARGSORT" is not defined, evaluates to 0 [-Werror=undef] 671 \| #if ULAB_NUMERICAL_ARGSORT \| ^~~~~~~~~~~~~~~~~~~~~~ cc1: all warnings being treated as errors ../../extmod/ulab/code/ulab.c:150:9: error: "ULAB_VECTORISE_" is not defined, evaluates to 0 [-Werror=undef] 150 \| #if ULAB_VECTORISE_ \| ^~~~~~~~~~~~~~~ ../../extmod/ulab/code/ulab.c:159:9: error: "ULAB_VECTORISE_TAHN" is not defined, evaluates to 0 [-Werror=undef] 159 \| #if ULAB_VECTORISE_TAHN \| ^~~~~~~~~~~~~~~~~~~ ../../extmod/ulab/code/ulab.c:198:9: error: "ULAB_NUMERICAL_ARGSORT" is not defined, evaluates to 0 [-Werror=undef] 198 \| #if ULAB_NUMERICAL_ARGSORT \| ^~~~~~~~~~~~~~~~~~~~~~	2020-02-11 11:08:25 -06:00
Jeff Epler	72025550d1	Merge pull request #8 from jepler/fix-undef-errors Fix some define-guards	2020-02-11 10:44:05 -06:00
Jeff Epler	1beddec7bb	Fix some define-guards These problems were found building in circuitpython: ../../extmod/ulab/code/numerical.c:671:5: error: "ULAB_NUMERICAL_ARGSORT" is not defined, evaluates to 0 [-Werror=undef] 671 \| #if ULAB_NUMERICAL_ARGSORT \| ^~~~~~~~~~~~~~~~~~~~~~ cc1: all warnings being treated as errors ../../extmod/ulab/code/ulab.c:150:9: error: "ULAB_VECTORISE_" is not defined, evaluates to 0 [-Werror=undef] 150 \| #if ULAB_VECTORISE_ \| ^~~~~~~~~~~~~~~ ../../extmod/ulab/code/ulab.c:159:9: error: "ULAB_VECTORISE_TAHN" is not defined, evaluates to 0 [-Werror=undef] 159 \| #if ULAB_VECTORISE_TAHN \| ^~~~~~~~~~~~~~~~~~~ ../../extmod/ulab/code/ulab.c:198:9: error: "ULAB_NUMERICAL_ARGSORT" is not defined, evaluates to 0 [-Werror=undef] 198 \| #if ULAB_NUMERICAL_ARGSORT \| ^~~~~~~~~~~~~~~~~~~~~~	2020-02-11 10:42:58 -06:00
Jeff Epler	5e81f75661	Trim README to just the basics	2020-02-11 10:08:47 -06:00
Jeff Epler	21b5a130c8	Port to CircuitPython	2020-02-11 10:08:47 -06:00
Zoltán Vörös	4a0677fd14	removed extra ndarray_get_buffer	2020-02-10 19:54:49 +01:00
Zoltán Vörös	800bb3b872	Merge branch 'master' of github.com:v923z/micropython-ulab	2020-02-10 19:51:22 +01:00
Zoltán Vörös	89170a13a6	fixed error in filter.c, removed asbytearray, and added buffer protocol to ndarray.c	2020-02-10 19:50:49 +01:00
Zoltán Vörös	ca23263655	Merge pull request #31 from jepler/memoryview ndarray: let memoryview(arr) work	2020-02-10 18:38:11 +01:00
Zoltán Vörös	936bb3bae5	corrected slicing error in issue #32	2020-02-09 19:56:32 +01:00
Jeff Epler	a90d18caf1	ndarray: let memoryview(arr) work .. this makes ndarray.rawbytes redundant. Circuitpython will remove it.	2020-02-08 10:29:45 -06:00
Zoltán Vörös	918075daaf	fixed a couple of small issues in filter.c	2020-02-08 10:45:13 +01:00
Zoltán Vörös	fac3eb4099	fixed a couple of small issues in filter.c	2020-02-08 10:44:24 +01:00
Zoltán Vörös	76ccd1a118	the master branch is configurable now	2020-02-07 21:23:24 +01:00
Zoltán Vörös	0315408e05	Merge pull request #29 from jepler/convolve convolve: implement and document	2020-02-06 19:53:51 +01:00
Jeff Epler	722401c2e4	Clean up and polish based on review comments no functional change	2020-02-06 12:43:39 -06:00
Jeff Epler	efa4c148c9	convolve: implement and document	2020-02-05 16:33:48 -06:00
Zoltán Vörös	dcd65feb21	fixed indexing error in linalg.dot	2020-01-29 19:27:28 +01:00
Zoltán Vörös	b2c4749962	replaced MP_ROM_PTR macro, wherever it made sense	2020-01-20 19:55:38 +01:00
Zoltán Vörös	941ffceb1b	replaced MP_ROM_PTR macro, wherever it made sense	2020-01-20 19:49:57 +01:00
Zoltán Vörös	ac1111c251	fixed glitch in std/mean functions	2020-01-07 21:53:23 +01:00
Zoltán Vörös	992e48b84e	switched to version string	2020-01-07 21:02:20 +01:00
Zoltán Vörös	68e38bc6b8	backup commit	2020-01-06 07:43:27 +01:00