{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Numba 0.54.0 Release Demo\n",
    "\n",
    "This notebook contains a demonstration of new features present in the 0.54.0 release of Numba. Whilst release notes are produced as part of the [CHANGE_LOG](https://github.com/numba/numba/blob/release0.54/CHANGE_LOG), there's nothing like seeing code in action!\n",
    "\n",
    "This release contains a lot of things including new features, updated library support and large scale internal changes/enhancements to permit extending Numba to new targets. In this notebook, the new CPU target features are demonstrated. The [CUDA target](https://numba.pydata.org/numba-doc/latest/cuda/index.html) also gained a lot of new features in 0.54.0 and [@gmarkall](https://github.com/gmarkall) has created a [demo notebook](https://mybinder.org/v2/gh/numba/numba-examples/master?filepath=notebooks%2FNumba_054_CUDA_Release_Demo.ipynb) especially for these!\n",
    "\n",
    "Key internal/support changes in this release are:\n",
    "\n",
    "- NumPy support is restricted to (1.17 <= x < 1.21).\n",
    "- Python 3.7 is the new minimum version of Python accepted.\n",
    "- Intel TBB 2021 is now supported and the new minimum version of TBB.\n",
    "- The LLVM backend has been upgraded to version 11.\n",
    "- Numba has directly vendored cloudpickle into its code base for use with caching and object mode.\n",
    "\n",
    "Intel also kindly sponsored research and development that lead to three new features:\n",
    "\n",
    "- Target extension support.\n",
    "- Custom NumPy array subclass support.\n",
    "- Context managed retargeting of the dispatcher."
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Python support enhancements"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "from numba import njit\n",
    "import numba\n",
    "assert numba.version_info.short >= (0, 54)\n",
    "from numba.core import types\n",
    "import numpy as np"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Numba 0.54 gains some basic `f-string` support ([@ehsantn](https://github.com/ehsantn)) and support for dictionary comprehensions ([@stuartarchibald](https://github.com/stuartarchibald))."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "@njit\n",
    "def demo_fstring_dict_comp():\n",
    "    return {x: f'string_{x * 10}' for x in range(5)}\n",
    "\n",
    "demo_fstring_dict_comp()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "It also gains support for the ``sum`` builtin ([@stuartarchibald](https://github.com/stuartarchibald))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "@njit\n",
    "def demo_sum():\n",
    "    return sum([1, 2, 3]), sum((4, 5, 6)), sum((1j, 3, np.float64(7)))\n",
    "    \n",
    "demo_sum()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "and the ability to `hash()` `IntEnum`s ([@HPLegion](https://github.com/HPLegion))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "from enum import IntEnum\n",
    "\n",
    "class Scale(IntEnum):\n",
    "    C4 = 261\n",
    "    D4 = 293\n",
    "    E4 = 329\n",
    "    F4 = 349\n",
    "    G4 = 391\n",
    "    A4 = 440\n",
    "\n",
    "@njit\n",
    "def demo_intenum_hash():\n",
    "    return hash(Scale.D4), {Scale.C4: 'Middle C'}[Scale.C4]\n",
    "\n",
    "demo_intenum_hash()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## NumPy support enhancements\n",
    "\n",
    "Numba 0.54 contains support for some more NumPy functions:\n",
    "\n",
    "* ``np.clip`` ([@arvoelke](https://github.com/arvoelke), [@gmarkall](https://github.com/gmarkall))\n",
    "* ``np.rot90`` ([@braniii](https://github.com/braniii))\n",
    "* ``np.swapaxes`` ([@braniii](https://github.com/braniii))\n",
    "* ``np.random.dirichlet`` ([@rishi-kulkarni](https://github.com/rishi-kulkarni))\n",
    "* ``np.isreal`` ([@guilhermeleobas](https://github.com/guilhermeleobas))\n",
    "* ``np.iscomplex`` ([@guilhermeleobas](https://github.com/guilhermeleobas))\n",
    "* ``np.isrealobj`` ([@guilhermeleobas](https://github.com/guilhermeleobas))\n",
    "* ``np.iscomplexobj`` ([@guilhermeleobas](https://github.com/guilhermeleobas))\n",
    "* ``np.isscalar`` ([@guilhermeleobas](https://github.com/guilhermeleobas))\n",
    "* ``np.isneginf`` ([@guilhermeleobas](https://github.com/guilhermeleobas))\n",
    "* ``np.isposinf`` ([@guilhermeleobas](https://github.com/guilhermeleobas))\n",
    "\n",
    "and also enhances the support for:\n",
    "\n",
    "* ``np.argmax()`` now supports the `axis` kwarg ([@itamarst](https://github.com/itamarst))\n",
    "* ``__setitem__`` using ``0d`` arrays ([@guilhermeleobas](https://github.com/guilhermeleobas))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "from numba import literal_unroll\n",
    "\n",
    "@njit\n",
    "def demo_numpy_new_features():\n",
    "    x = np.linspace(0, 10, 20)\n",
    "    print(\"np.clip\\n\", np.clip(x, 2., 5.))\n",
    "    print(\"np.rot90\\n\", np.rot90(np.arange(4).reshape((2, 2))))\n",
    "    print(\"np.swapaxes\\n\", np.swapaxes(np.arange(8).reshape((2, 2, 2)), 0, 2))\n",
    "\n",
    "    np.random.seed(0)\n",
    "    print(\"np.random.dirichlet\\n\", np.random.dirichlet((2, 3, 4), 2))\n",
    "\n",
    "    for x in literal_unroll((1, 1j, np.array([1, 2j, 3]))):\n",
    "        print(\"x is \", x)\n",
    "        print(\"  np.iscomplexobj(x)\", np.iscomplexobj(x))\n",
    "        print(\"  np.isreal(x)\", np.isreal(x))\n",
    "        print(\"  np.isrealobj(x)\", np.isrealobj(x))\n",
    "        print(\"  np.isscalar(x)\", np.isscalar(x))\n",
    "    for x in literal_unroll((1, np.inf, -np.inf, np.nan)):\n",
    "        print(\"x is \", x)\n",
    "        print(\"  np.isneginf(x)\", np.isneginf(x))\n",
    "        print(\"  np.isposinf(x)\", np.isposinf(x))\n",
    "        print(\"  np.iscomplex(x)\", np.iscomplex(x))\n",
    "        \n",
    "demo_numpy_new_features()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "Enhancements:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "@njit\n",
    "def demo_numpy_enhancements():\n",
    "    # Axis support in np.argmax\n",
    "    x = np.arange(12).reshape((3, 2, 2))\n",
    "    x[:, 1, :] = 100\n",
    "    print(\"np.argmax(x, axis=2)\\n\", np.argmax(x, axis=2))\n",
    "    \n",
    "    # Setitem using 0d arrays\n",
    "    idx = np.array(1)\n",
    "    x = np.arange(10)\n",
    "    x[idx] = 100\n",
    "    print(\"Setitem of x[idx]\\n\", x, \", idx =\", idx, \", idx.ndim =\", idx.ndim)\n",
    "    \n",
    "demo_numpy_enhancements()"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "----\n",
    "\n",
    "## See Also\n",
    "\n",
    "- [0.54 CUDA Demo](https://mybinder.org/v2/gh/numba/numba-examples/master?filepath=notebooks%2FNumba_054_CUDA_Release_Demo.ipynb)\n",
    "- [0.54 `overload_classmethod` and Array Subclass Demo](https://mybinder.org/v2/gh/numba/numba-examples/master?filepath=notebooks%2FNumba_054_overload_classmethod_array_subclass.ipynb)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.9.5"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}