{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "b54830fc",
   "metadata": {},
   "source": [
    "# NumPy Tutorial 02: Array Creation, Shape, and Memory Model"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e76399d1",
   "metadata": {},
   "source": [
    "## Download Notebook\n",
    "\n",
    "{download}`Download this notebook <02_array_creation_and_attributes.ipynb>`"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "fe7b4a0c",
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "1b912b0a",
   "metadata": {},
   "source": [
    "## 1. Core constructors\n",
    "\n",
    "Use specialized constructors when possible for clarity and performance."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "d7500161",
   "metadata": {},
   "outputs": [],
   "source": [
    "a = np.array([[1, 2, 3], [4, 5, 6]], dtype=np.int32)\n",
    "z = np.zeros((2, 4), dtype=np.float64)\n",
    "o = np.ones((2, 4), dtype=np.float64)\n",
    "r = np.arange(0, 10, 2)\n",
    "l = np.linspace(0, 1, 6)\n",
    "\n",
    "print('a:\\n', a)\n",
    "print('zeros shape:', z.shape)\n",
    "print('ones shape:', o.shape)\n",
    "print('arange:', r)\n",
    "print('linspace:', l)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "56d64418",
   "metadata": {},
   "source": [
    "## 2. Inspecting structure\n",
    "\n",
    "`shape`, `ndim`, `size`, `dtype`, `itemsize`, and `nbytes` are foundational attributes."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "19141687",
   "metadata": {},
   "outputs": [],
   "source": [
    "x = np.arange(24, dtype=np.int16).reshape(2, 3, 4)\n",
    "print('shape:', x.shape)\n",
    "print('ndim:', x.ndim)\n",
    "print('size:', x.size)\n",
    "print('dtype:', x.dtype)\n",
    "print('itemsize:', x.itemsize)\n",
    "print('nbytes:', x.nbytes)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9efa64d7",
   "metadata": {},
   "source": [
    "## 3. Reshape and flatten family\n",
    "\n",
    "`reshape` changes view/interpretation when possible; `ravel` prefers view; `flatten` always returns a copy."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "e74908af",
   "metadata": {},
   "outputs": [],
   "source": [
    "x = np.arange(12)\n",
    "m = x.reshape(3, 4)\n",
    "rv = m.ravel()\n",
    "fl = m.flatten()\n",
    "\n",
    "m[0, 0] = 999\n",
    "print('m:\\n', m)\n",
    "print('ravel reflects change:', rv[0])\n",
    "print('flatten copy value:', fl[0])"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "8366f16f",
   "metadata": {},
   "source": [
    "## 4. Memory order and transposition\n",
    "\n",
    "NumPy supports C-order (row-major) and Fortran-order (column-major). Some operations create views, others create copies."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "4fe069cb",
   "metadata": {},
   "outputs": [],
   "source": [
    "m = np.arange(12).reshape(3, 4)\n",
    "t = m.T\n",
    "\n",
    "print('Original shape:', m.shape, 'strides:', m.strides)\n",
    "print('Transposed shape:', t.shape, 'strides:', t.strides)\n",
    "print('Share memory?', np.shares_memory(m, t))"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "79d135db",
   "metadata": {},
   "source": [
    "## 5. Type casting strategies\n",
    "\n",
    "Use `astype` to cast. Prefer controlled casting (`casting='safe'`) in production pipelines."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "bb875ea4",
   "metadata": {},
   "outputs": [],
   "source": [
    "x = np.array([1.2, 3.7, -2.9])\n",
    "print('Original:', x, x.dtype)\n",
    "xi = x.astype(np.int32)\n",
    "print('Cast to int32:', xi, xi.dtype)\n",
    "\n",
    "y = np.array([1, 2, 3], dtype=np.int64)\n",
    "print('Safe cast int64 -> float64:', y.astype(np.float64, casting='safe'))"
   ]
  }
 ],
 "metadata": {
  "language_info": {
   "name": "python"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}