test_modules.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "    comm  nodetype  connect\n",
      "1      1         0        0\n",
      "2      1         0        0\n",
      "3      1         1        2\n",
      "4      2         1        1\n",
      "5      2         0        0\n",
      "6      2         0        0\n",
      "7      2         0        0\n",
      "8      2         0        0\n",
      "9      2         0        0\n",
      "10     2         1        3\n",
      "11     3         1        2\n",
      "12     3         0        0\n",
      "13     3         0        0\n",
      "14     3         0        0\n",
      "15     3         0        0\n",
      "16     3         0        0\n",
      "17     3         1        1\n",
      "18     1         1        3\n",
      "19     1         0        0\n",
      "20     1         0        0\n",
      "21     1         0        0\n"
     ]
    }
   ],
   "source": [
    "%matplotlib inline\n",
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "from tesser import network\n",
    "\n",
    "# data path (edit for local setup)\n",
    "data_dir = '/Users/morton/data/tesser'\n",
    "\n",
    "# load node information\n",
    "node_info = network.node_info()\n",
    "print(node_info)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "<matplotlib.image.AxesImage at 0x7fb8e8200518>"
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 432x288 with 1 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# get transition probability and plot it\n",
    "T = network.adjacency(node_info) / 6\n",
    "fig, ax = plt.subplots()\n",
    "ax.matshow(T)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "   SubjNum  run  trial  objnum           file orientnam  orientnum resp  \\\n",
      "0      101    1      1       9  object_17.jpg       cor        1.0  NaN   \n",
      "1      101    1      2       4  object_18.jpg       cor        1.0    c   \n",
      "2      101    1      3       3   object_4.jpg       cor        1.0    c   \n",
      "3      101    1      4       4  object_18.jpg       rot        0.0    n   \n",
      "4      101    1      5       8   object_9.jpg       cor        1.0    n   \n",
      "\n",
      "   respnum  acc        rt  part  seqtype  \n",
      "0      0.0  0.0       NaN     1      NaN  \n",
      "1      1.0  1.0  1.479518     1      NaN  \n",
      "2      1.0  1.0  1.334561     1      NaN  \n",
      "3      2.0  1.0  1.065799     1      NaN  \n",
      "4      2.0  0.0  1.234106     1      NaN  \n"
     ]
    }
   ],
   "source": [
    "# load all structure learning runs in a dataframe\n",
    "from tesser import util\n",
    "df = util.load_struct(data_dir, 101)\n",
    "print(df.head())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "   SubjNum  run  trial  objnum           file orientnam  orientnum resp  \\\n",
      "2      101    3      3      21  object_24.jpg       cor        1.0    c   \n",
      "3      101    3      4      18  object_23.jpg       cor        1.0    c   \n",
      "4      101    3      5       2  object_28.jpg       cor        1.0    c   \n",
      "5      101    3      6      18  object_23.jpg       rot        0.0    n   \n",
      "6      101    3      7      19   object_7.jpg       cor        1.0    c   \n",
      "\n",
      "   respnum  acc        rt  part  seqtype  \n",
      "2      1.0  1.0  0.694019     2      1.0  \n",
      "3      1.0  1.0  0.590571     2      1.0  \n",
      "4      1.0  1.0  0.659165     2      1.0  \n",
      "5      2.0  1.0  0.696351     2      1.0  \n",
      "6      1.0  1.0  0.595409     2      1.0  \n"
     ]
    }
   ],
   "source": [
    "# example of getting a specific run\n",
    "run = 3\n",
    "part = 2\n",
    "df_run = df.loc[(df.part == part) & (df.run == run), :]\n",
    "print(df_run.head())"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [],
   "source": [
    "# train SR agent on all runs\n",
    "from tesser import learn\n",
    "n_state = len(np.unique(df_run.objnum))\n",
    "M = np.zeros((n_state, n_state))\n",
    "\n",
    "M_all = {}\n",
    "for part in (1, 2):\n",
    "    runs = np.unique(df.loc[df.part == part, 'run'])\n",
    "    for run in runs:\n",
    "        # get data for this run\n",
    "        df_run = df.loc[(df.part == part) & (df.run == run), :]\n",
    "        M_new = learn.run_experiment(df_run.objnum.values, .5, .1, M)        \n",
    "        M = M_new.copy()\n",
    "        M_all[(part, run)] = M_new.copy()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "image/png": "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\n",
      "text/plain": [
       "<Figure size 1008x432 with 11 Axes>"
      ]
     },
     "metadata": {
      "needs_background": "light"
     },
     "output_type": "display_data"
    }
   ],
   "source": [
    "# plot SR after each run\n",
    "fig, ax = plt.subplots(2, 6, figsize=(14, 6))\n",
    "for i, part in enumerate((1, 2)):\n",
    "    for j, run in enumerate(range(1, 7)):\n",
    "        if (part, run) not in M_all:\n",
    "            fig.delaxes(ax[i, j])\n",
    "            continue\n",
    "        ax[i,j].matshow(M_all[(part, run)], vmin=0, vmax=.5)"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "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": "3.7.3"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 4
}