{
 "cells": [
  {
   "metadata": {},
   "cell_type": "markdown",
   "source": [
    "# Terrible Search Engine\n",
    "\n",
    "Using simulated annealing on websites with links pointing to neighbours is not a good idea, but a good example for async annealing."
   ],
   "id": "88cb648dc111d148"
  },
  {
   "metadata": {
    "collapsed": true,
    "ExecuteTime": {
     "end_time": "2026-07-01T21:32:47.864907206Z",
     "start_time": "2026-07-01T21:32:47.852781369Z"
    }
   },
   "cell_type": "code",
   "outputs": [],
   "execution_count": 45,
   "source": [
    "from urllib.parse import urljoin\n",
    "from random import choice\n",
    "\n",
    "from anyio import sleep\n",
    "from httpx import AsyncClient\n",
    "from bs4 import BeautifulSoup\n",
    "from altbacken.core.async_annealing import AnnealingState\n",
    "from altbacken.external.async_annealing import SimpleAsyncSimulatedAnnealing"
   ],
   "id": "2233d4dcc94800b5"
  },
  {
   "metadata": {
    "ExecuteTime": {
     "end_time": "2026-07-01T21:08:49.862087569Z",
     "start_time": "2026-07-01T21:08:49.843450567Z"
    }
   },
   "cell_type": "code",
   "source": [
    "type URL = str | None\n",
    "type Weight = int"
   ],
   "id": "1e97599a39a5978b",
   "outputs": [],
   "execution_count": 11
  },
  {
   "metadata": {
    "ExecuteTime": {
     "end_time": "2026-07-01T21:11:12.531277355Z",
     "start_time": "2026-07-01T21:11:12.513488364Z"
    }
   },
   "cell_type": "code",
   "source": [
    "SEARCH_KEYS: dict[str, Weight] = {\n",
    "    \"Python\": -1,\n",
    "    \"Moon\": -5,\n",
    "    \"Bass\": -8,\n",
    "    \"Cheesecake\": -10\n",
    "}"
   ],
   "id": "13ad3f6428ff6688",
   "outputs": [],
   "execution_count": 15
  },
  {
   "metadata": {
    "ExecuteTime": {
     "end_time": "2026-07-01T21:38:30.255073134Z",
     "start_time": "2026-07-01T21:38:30.206041569Z"
    }
   },
   "cell_type": "code",
   "source": [
    "class LinkNeighbourhood:\n",
    "    async def __call__(self, state: AnnealingState[URL]) -> URL:\n",
    "        if current_url := state.current_solution:\n",
    "            try:\n",
    "                async with AsyncClient(follow_redirects=True) as client:\n",
    "                    response = await client.get(current_url)\n",
    "                    if not response.is_success:\n",
    "                        return current_url\n",
    "                    await sleep(0.5)\n",
    "                    soup = BeautifulSoup(response.text, \"html.parser\")\n",
    "                    links: list[str] = [\n",
    "                        urljoin(current_url, link[\"href\"])\n",
    "                        for link in soup.find_all(\"a\", {\"href\": True})\n",
    "                    ]\n",
    "                    return choice(links) if links else None\n",
    "            except Exception:\n",
    "                return current_url\n",
    "        else:\n",
    "            return None\n"
   ],
   "id": "b29ed2546adef332",
   "outputs": [],
   "execution_count": 53
  },
  {
   "metadata": {
    "ExecuteTime": {
     "end_time": "2026-07-01T21:38:31.230987970Z",
     "start_time": "2026-07-01T21:38:31.221101945Z"
    }
   },
   "cell_type": "code",
   "source": [
    "async def fitness(url: URL) -> float:\n",
    "    if url is None:\n",
    "        return 100.0\n",
    "    async with AsyncClient(follow_redirects=True) as client:\n",
    "        response = await client.get(url)\n",
    "        await sleep(0.5)\n",
    "        if not response.is_success:\n",
    "            return 50.0\n",
    "        text: str = response.text\n",
    "        return sum(\n",
    "            text.count(key) * weight\n",
    "            for key, weight in SEARCH_KEYS.items()\n",
    "        )\n"
   ],
   "id": "9161e7c27436d5a9",
   "outputs": [],
   "execution_count": 54
  },
  {
   "metadata": {
    "ExecuteTime": {
     "end_time": "2026-07-01T21:38:32.257157919Z",
     "start_time": "2026-07-01T21:38:32.248315960Z"
    }
   },
   "cell_type": "code",
   "source": [
    "annealing: SimpleAsyncSimulatedAnnealing[URL] = SimpleAsyncSimulatedAnnealing(\n",
    "    fitness,\n",
    "    LinkNeighbourhood()\n",
    ")\n",
    "\n",
    "visited_sites: set[URL] = set()\n",
    "\n",
    "async def tracer(state: AnnealingState[URL]) -> None:\n",
    "    visited_sites.add(state.current_solution)\n",
    "\n",
    "annealing.tracer = tracer"
   ],
   "id": "2d13b6caf8515629",
   "outputs": [],
   "execution_count": 55
  },
  {
   "metadata": {
    "ExecuteTime": {
     "end_time": "2026-07-01T21:38:49.594223134Z",
     "start_time": "2026-07-01T21:38:34.002282Z"
    }
   },
   "cell_type": "code",
   "source": "best_url, value = await annealing.simulate(\"https://stackoverflow.com/questions\")",
   "id": "d49eca2668d5a1c",
   "outputs": [],
   "execution_count": 56
  },
  {
   "metadata": {
    "ExecuteTime": {
     "end_time": "2026-07-01T21:38:49.665670293Z",
     "start_time": "2026-07-01T21:38:49.602114497Z"
    }
   },
   "cell_type": "code",
   "source": "best_url, value",
   "id": "782f72c3b91d1b33",
   "outputs": [
    {
     "data": {
      "text/plain": [
       "('https://stackoverflow.com/questions/79972182/scipy-import-from-module-fails-sometimes-path-dependent',\n",
       " -6)"
      ]
     },
     "execution_count": 57,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "execution_count": 57
  },
  {
   "metadata": {
    "ExecuteTime": {
     "end_time": "2026-07-01T21:38:49.731517017Z",
     "start_time": "2026-07-01T21:38:49.675371214Z"
    }
   },
   "cell_type": "code",
   "source": "visited_sites",
   "id": "785b4f7b7c27ada5",
   "outputs": [
    {
     "data": {
      "text/plain": [
       "{None,\n",
       " 'https://biology.stackexchange.com/users/129209/keiv',\n",
       " 'https://stackexchange.com',\n",
       " 'https://stackoverflow.com/questions/79972182/scipy-import-from-module-fails-sometimes-path-dependent',\n",
       " 'https://stackoverflowteams.com',\n",
       " 'https://stackoverflowteams.com/teams/create/free/?utm_medium=referral&utm_source=biology-community&utm_campaign=side-bar&utm_content=explore-teams',\n",
       " 'https://stackoverflowteams.com/teams/create/free?isLogin=false',\n",
       " 'https://stackoverflowteams.com/teams/create/free?isLogin=true',\n",
       " 'https://www.instagram.com/thestackoverflow'}"
      ]
     },
     "execution_count": 58,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "execution_count": 58
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 2
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython2",
   "version": "2.7.6"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}