test_acrcd.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 21,
   "id": "70f13a71",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "The autoreload extension is already loaded. To reload it, use:\n",
      "  %reload_ext autoreload\n"
     ]
    }
   ],
   "source": [
    "%load_ext autoreload\n",
    "%autoreload 2\n",
    "from tqdm import tqdm\n",
    "\n",
    "from transport_problem import DualOracle, OptimParams\n",
    "from acrcd_no_adaptive import ACRCDOracleStacker, ACRCD_star\n",
    "import numpy as np\n",
    "\n",
    "from transport_problem import HyperParams, OptimParams\n",
    "import numpy as np\n",
    "import scipy\n",
    "import matplotlib.pyplot as plt\n",
    "from pathlib import Path\n",
    "from read_data import get_network_df, get_corrs, get_full_df\n",
    "from numba.core import types\n",
    "import numba\n",
    "\n",
    "from numba import njit\n",
    "from graph_tool.topology import shortest_distance\n",
    "import pandas as pd\n",
    "import networkx as nx\n",
    "import graph_tool as gt\n",
    "from graph_tool.topology import shortest_distance\n",
    "import numpy as np\n",
    "import typing\n",
    "\n",
    "import numba\n",
    "from numba.core import types\n",
    "from get_graph import get_graph_by_name"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 10,
   "id": "fcdc4d72",
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "graph type:  SiouxFalls\n"
     ]
    }
   ],
   "source": [
    "net_df, corrs = get_graph_by_name(\"SiouxFalls\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 16,
   "id": "f9d4ad04",
   "metadata": {},
   "outputs": [],
   "source": [
    "graph = gt.Graph(net_df.values, eprops=[('capacity', 'double'), ('fft', 'double')])\n",
    "\n",
    "l = np.maximum(np.sum(corrs, axis=1), 1e-6)\n",
    "w = np.maximum(np.sum(corrs, axis=0), 1e-6)\n",
    "\n",
    "zones_num = len(l)\n",
    "sources = np.arange(zones_num)\n",
    "targets = np.arange(zones_num)\n",
    "\n",
    "params = HyperParams(gamma=10, mu_pow=0.25, rho=0.15)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 17,
   "id": "d4a6853e",
   "metadata": {},
   "outputs": [],
   "source": [
    "def test_transport_problem(graph, net_df, l, w, params, sources, targets):\n",
    "    dual_oracle = DualOracle(\n",
    "        graph, l, w, params\n",
    "    )\n",
    "    oracle_stacker = ACRCDOracleStacker(\n",
    "        dual_oracle,\n",
    "        graph,\n",
    "        sources,\n",
    "        targets\n",
    "    )\n",
    "    \n",
    "    return ACRCD_star(oracle_stacker, *oracle_stacker.get_init_vars_block(), 1000)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 18,
   "id": "e24ed042",
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "100%|██████████| 1000/1000 [00:01<00:00, 648.87it/s]\n"
     ]
    }
   ],
   "source": [
    "results = test_transport_problem(graph,net_df, l, w, params, sources, targets)\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 22,
   "id": "1391b0cb",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x7f521c5baa90>]"
      ]
     },
     "execution_count": 22,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(results.history_la_mu_calls, results.history_dual_gap)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 23,
   "id": "407fd0b4",
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "[<matplotlib.lines.Line2D at 0x7f521c52be50>]"
      ]
     },
     "execution_count": 23,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 640x480 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "plt.plot(results.history_la_mu_grad_norm)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "4ff3e560",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "9c7181cc",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "cb69a744",
   "metadata": {},
   "source": []
  },
  {
   "cell_type": "markdown",
   "id": "cca10c0e",
   "metadata": {},
   "source": [
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "0e553a55",
   "metadata": {},
   "outputs": [],
   "source": [
    "plt.yscale('log')\n",
    "plt.plot(history, label=\"dgap\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "80e52f97",
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "\n",
    "data = []\n",
    "x = t_calls\n",
    "for i, j, k in zip(x, np.abs(history[1:]), np.abs(la_mu_grad_norms)):\n",
    "    data.append((i, j, k))\n",
    "\n",
    "df = pd.DataFrame(data)\n",
    "df.to_csv(\"plot_ustm.csv\", index=False)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "a483cac1",
   "metadata": {},
   "outputs": [],
   "source": [
    "plt.yscale('log')\n",
    "plt.plot(la_mu_grad_norms)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "3d396921",
   "metadata": {},
   "outputs": [],
   "source": [
    "x = t_calls\n",
    "dual_gap = history\n",
    "dual_gap_clipped = np.maximum(0, dual_gap)\n",
    "                                \n",
    "metric = 2 * np.array(results.history_la_mu_grad_norm) * la_mu_end_norm + dual_gap_clipped\n",
    "plt.plot(x, metric, label=\"custom metric\")\n",
    "plt.yscale(\"log\")\n",
    "plt.legend()\n",
    "plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "1668c448",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "7cd279de",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "55c15e3a",
   "metadata": {},
   "outputs": [],
   "source": []
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "ca41e82b",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "testEnv",
   "language": "python",
   "name": "testenv"
  },
  "language_info": {
   "codemirror_mode": {
    "name": "ipython",
    "version": 3
   },
   "file_extension": ".py",
   "mimetype": "text/x-python",
   "name": "python",
   "nbconvert_exporter": "python",
   "pygments_lexer": "ipython3",
   "version": "3.11.4"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}