{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "d35ddef5",
   "metadata": {},
   "outputs": [],
   "source": [
    "from fastcore.all import *\n",
    "from execnb.nbio import *"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "891eb357",
   "metadata": {},
   "outputs": [],
   "source": [
    "from bs4 import BeautifulSoup\n",
    "from bs4.formatter import HTMLFormatter"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "8f3176d9",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "'Natural'"
      ]
     },
     "execution_count": null,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "b = BeautifulSoup('hi <img alt=\"Natural\" caption=\"artificial\" id=\"neuron\" src=\"images/chapter7.png\" width=\"500\"/>')\n",
    "im = b.img\n",
    "im['alt']"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "9fa51767",
   "metadata": {},
   "outputs": [],
   "source": [
    "def img2md(im):\n",
    "    r = '!['\n",
    "    a = im.get('caption', None)\n",
    "    if a: r+=a\n",
    "    r+=f']({im[\"src\"]}'\n",
    "    r+=')'\n",
    "    d = {}\n",
    "    a = im.get('width', None)\n",
    "    if a: d['width'] = a\n",
    "    a = im.get('id', None)\n",
    "    if a: d['id'] = 'fig-'+re.sub('_', '-', a)\n",
    "    a = im.get('alt', None)\n",
    "    if a: d['fig-alt'] = a\n",
    "    if d: r+='{' + ' '.join(f'{k}=\"{v}\"' for k,v in d.items()) + '}'\n",
    "    return r"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "ac9430d1",
   "metadata": {},
   "outputs": [],
   "source": [
    "class UnsortedAttributes(HTMLFormatter):\n",
    "    def attributes(self, tag): yield from tag.attrs.items()\n",
    "\n",
    "def enc(t): return t.encode_contents(formatter=UnsortedAttributes()).decode('utf-8')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "232f3673",
   "metadata": {},
   "outputs": [],
   "source": [
    "def nb_images(nb):\n",
    "    for c in nb.cells:\n",
    "        o = c.source\n",
    "        if c.cell_type=='markdown' and '<img ' in o:\n",
    "            b = BeautifulSoup(o)\n",
    "            im = b.img\n",
    "            im.replace_with(img2md(im))\n",
    "            o = b.body\n",
    "            if getattr(o,'p',None): o = o.p\n",
    "            c.source = enc(o)\n",
    "    #         print(enc(o))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "fa5aea9f",
   "metadata": {},
   "outputs": [],
   "source": [
    "def fix_sec(o): return '@sec-' + re.sub('_', '-', o.group(1))\n",
    "def fix_fig(o): return '@fig-' + re.sub('_', '-', o.group(1))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "ce23aa79",
   "metadata": {},
   "outputs": [],
   "source": [
    "def nb_xrefs(nb):\n",
    "    for c in nb.cells:\n",
    "        o = c.source\n",
    "        if c.cell_type=='markdown' and '<<' in o:\n",
    "            o = re.sub(r'<<chapter_(\\w+)>>', fix_sec, o)\n",
    "            o = re.sub(r'<<(\\w+)>>', fix_fig, o)\n",
    "            c.source = o\n",
    "    #         print(o)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "3b24c7a7",
   "metadata": {},
   "outputs": [],
   "source": [
    "def fix_lab(o): return '#| label: fig-' + re.sub('_', '-', o.group(1))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "35df4b8f",
   "metadata": {},
   "outputs": [],
   "source": [
    "def nb_comput(nb):\n",
    "    for c in nb.cells:\n",
    "        o = c.source\n",
    "        if c.cell_type=='code' and '#id' in o:\n",
    "            o = re.sub(r'^#id +(\\w+)', fix_lab, o, flags=re.MULTILINE)\n",
    "            o = re.sub(r'^#caption +(.+)', r'#| fig-cap: \"\\1\"', o, flags=re.MULTILINE)\n",
    "            o = re.sub(r'^#alt +(.+)', r'#| fig-alt: \"\\1\"', o, flags=re.MULTILINE)\n",
    "            c.source = o\n",
    "    #         print(o)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "762e5dee",
   "metadata": {},
   "outputs": [],
   "source": [
    "def fix_nb(path):\n",
    "    nb = read_nb(path)\n",
    "    nb_images(nb)\n",
    "    nb_xrefs(nb)\n",
    "    nb_comput(nb)\n",
    "    write_nb(nb, path)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "5351c94f",
   "metadata": {},
   "outputs": [],
   "source": [
    "# fix_nb('../mnist_basics.ipynb')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "591ffa9f",
   "metadata": {},
   "outputs": [],
   "source": [
    "# for path in Path('..').ls(file_exts=['.ipynb']): fix_nb(path)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "68b7c95a",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "language": "python",
   "name": "python3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}