11_ElectroFish_Milestone2.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# Welcome to your 11th assignment!\n",
    "\n",
    "Please submit your solution of this notebook in the Whiteboard at the corresponding Assignment entry. We need you to upload the .ipynb-file and the exported .pdf of this notebook.\n",
    "\n",
    "If you have any questions, ask them in either in the tutorials or in the \"Mattermost\" channel: https://mattermost.imp.fu-berlin.de/biorobotics/channels/ssl_ws_2324\n",
    "\n",
    "\n",
    "For the last assignment of the course, we want you to solve a small project working on data from the ***ElectroFish*** project.\n",
    "\n",
    "Useful links: <br>\n",
    "[Blogpost (1)](https://shchur.github.io/blog/2020/tpp1-conditional-intensity/)<br>\n",
    "[Blogpost (2)](https://shchur.github.io/blog/2021/tpp2-neural-tpps/)<br>"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "# ElectroFish\n",
    "\n",
    "## Short Description\n",
    "The ElectroFish project is based around a fish species with an electric organ. The fish can emit electric impulses that are used for communication and navigation. One purpose of the experiments is to gain insight into the fish's social interactions.\n",
    "\n",
    "\n",
    "## ElectroFish Assignment Goal\n",
    "Using SSL, we want to learn a joint embedding of the fish's locomotory behaviour and the fish's EOD (Electric Organ Discharge) activity. <br> \n",
    "We want to use this embedding to create counterfactuals, i.e. we want to find neural activity that would lead to a certain behaviour and vice versa. <br>\n",
    "\n",
    "\n",
    "## Data\n",
    "You will find the data here: <br>\n",
    "[gDrive](https://drive.google.com/drive/folders/11kz1BGG4qCm7uGdHDINltBo53fD_YvPD?usp=drive_link)\n",
    "\n",
    "The data is structured as follows:\n",
    "\n",
    "-> Mormyrus_Pair_xx <br>\n",
    "&nbsp;&nbsp;&nbsp;&nbsp;-> 20230316_Mormyrus_Pair_xx.json.gz <br>\n",
    "&nbsp;&nbsp;&nbsp;&nbsp;-> fish_ids <br>\n",
    "&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;-> labels.png <br>\n",
    "&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;-> labels.xlsx <br>\n",
    "&nbsp;&nbsp;&nbsp;&nbsp;-> metadata <br>\n",
    "&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;-> tank_annotations.json <br>\n",
    "&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;-> video_meta.json <br>\n",
    "&nbsp;&nbsp;&nbsp;&nbsp;-> poses <br>\n",
    "&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;-> poses.h5 <br>\n",
    "&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;-> poses.slp <br>\n",
    "&nbsp;&nbsp;&nbsp;&nbsp;-> raw <br>\n",
    "&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;-> 20230316_Mormyrus_Pair_xx.mp4 <br>\n",
    "&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;-> 20230316_Mormyrus_Pair_xx.smr <br>\n",
    "\n",
    "The **.json.gz** file contains the processed data used for training a location prediction model. We recommend checking this out first to get a nice overview. <br>\n",
    "The **fish_ids** folder contains the manually labeled EOD signals. <br>\n",
    "The **metadata** folder contains the metadata for the tank and video of the experiment. <br>\n",
    "The **poses** folder contains the tracking data of the fish obtained with a [SLEAP](https://sleap.ai/) model. <br>\n",
    "The **raw** folder contains the raw data of the experiment: .mp4 video recording and .smr continuous electrode data recodring. <br>\n",
    "\n",
    "To get a nice overview of the data, we recommend checking out the **.json.gz** file and the **metadata**-folder first.<br>\n",
    "Like mentioned before, the **.json.gz** file contains the processed data used for training a location prediction model. It contains data points for every eod event happening in the experiment. For our tasks, we recommend processing the full data yourself (.h5 or .slp file for full tracking data; .smr file for full electrode data), to not loose any information. <br>\n",
    "\n",
    "The columns of the DataFrame provide you with a nice overview of the data:\n",
    "- **time_stamp**: The time stamp of the EOD event in seconds\n",
    "- **video_frame**: The video frame in which the EOD event occured\n",
    "- **fish_id**: The ID of the fish that emitted the EOD event\n",
    "    - manually labeled in Pair experiments with 1 being the bigger fish\n",
    "- **pixel_tracks**: all SLEAP track in pixel coordinates of the given video_frame\n",
    "    - The shape of the data [2,6,2] corresponds to [fish_id, bodypart_id, x/y coordinates]\n",
    "    - The bodyparts: \n",
    "        - 0: head\n",
    "        - 1: left_fin\n",
    "        - 2: right_fin\n",
    "        - 3: body_end\n",
    "        - 4: back_fin\n",
    "        - 5: middle\n",
    "- **world_tracks**: all SLEAP tracks in world coordinates\n",
    "- **tracking_scores**: 'confidence' score of SLEAP for all fish \n",
    "- **ced_electrode_data**: of the peak detected from the smr-data\n",
    "    - EOD event sensor data\n",
    "    - The shape of the data [4,70] corresponds to [elecrode_id, value window of size 70]\n",
    "- **efi_electrode_data**: of the peak detected from the board-data\n",
    "    - Different sensor for measurung EOD events\n",
    "    - Shape [4,50]\n",
    "\n",
    "\n",
    "## Milestones\n",
    "To provide you with more detailed subtasks, we split the assignment into 3 milestones. Each milestone represents the workload of 1 week and should be submitted like an assignment sheet. Therefore, please use and submit this notebook in the Whiteboard at the corresponding Assignment entry. We need you to upload the .ipynb-file and the exported .pdf of this notebook. \n",
    "\n",
    "**Milestone 1** <br>\n",
    "Done. <br>\n",
    "\n",
    "**Milestone 2** <br>\n",
    "*Goal*: Train an imiation model of the fish electrical behaviour. <br>\n",
    "*Steps*: <br>\n",
    "&nbsp;&nbsp;&nbsp;&nbsp;1) Preprossessing input and output data <br>\n",
    "&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;Inputs: [EOD signal] <br>\n",
    "&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;Output: [PDF parameters] <br>\n",
    "&nbsp;&nbsp;&nbsp;&nbsp;2) Prepare DataLoader<br>\n",
    "&nbsp;&nbsp;&nbsp;&nbsp;3) Train Imitation model for the EOD signal<br>\n",
    "&nbsp;&nbsp;&nbsp;&nbsp;4) Visualization / Evaluation <br>\n",
    "\n",
    "**Milestone 3** <br>\n",
    "*TBC in 2 weeks, 06.02.24*\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 1,
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "import seaborn as sns\n",
    "import pandas as pd\n",
    "import torch\n",
    "import sys\n",
    "from tqdm.notebook import tqdm\n",
    "\n",
    "# Load processed EOD data\n",
    "from fish_helper import DatasetContainer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 2,
   "metadata": {},
   "outputs": [
    {
     "name": "stderr",
     "output_type": "stream",
     "text": [
      "/home/mi/manueh51/ssl_2324_assignments/fish_helper.py:59: FutureWarning: Passing literal json to 'read_json' is deprecated and will be removed in a future version. To read from a literal string, wrap it in a 'StringIO' object.\n",
      "  self.data = pd.read_json(container_json['data'], dtype=dtype)\n"
     ]
    },
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "{'tank_corner_points': [[128.95050048828125, 1697.019775390625], [113.4000015258789, 280.3999938964844], [1831.026123046875, 279.8823547363281], [1831.59716796875, 1672.93994140625]], 'pixel_to_world_matrix': [[0.06644436292251645, -0.0007293721365335534, -7.330274914167141], [-0.0009595544796270599, -0.06784856731417248, 115.26409549446636], [-9.74359066112888e-06, -6.028094366956945e-06, 1.0]], 'world_to_pixel_matrix': [[14.893747708751922, -0.1715645723739672, 128.95050048828125], [0.03626900649996236, -14.720576310935625, 1697.019775390625], [0.0001453372084619642, -9.04086815969854e-05, 1.0]], 'tracks_minimum_score': 3.0, 'first_frame_of_sync': 136, 'time_of_video': 308.6, 'fps': 25, 'peak_left_offset': 35, 'peak_right_offset': 35, 'peak_prominence': 0.03, 'filter_plateau': True, 'min_plateau_size': 3, 'num_real_fish': 2, 'dtypes': {'time_stamp': 'float64', 'video_frame': 'int64', 'fish_id': 'Int64', 'pixel_tracks': 'object', 'world_tracks': 'object', 'tracking_scores': 'object', 'ced_electrode_data': 'object', 'efi_electrode_data': 'object'}, 'board_data_lag_to_smr': 207376, 'board_data_lag_to_smr_in_seconds': 4.14752, 'board_data_lag_to_smr_manuel_adjustment_to_automatic_in_seconds': 0.0}\n"
     ]
    },
    {
     "data": {
      "text/html": [
       "<div>\n",
       "<style scoped>\n",
       "    .dataframe tbody tr th:only-of-type {\n",
       "        vertical-align: middle;\n",
       "    }\n",
       "\n",
       "    .dataframe tbody tr th {\n",
       "        vertical-align: top;\n",
       "    }\n",
       "\n",
       "    .dataframe thead th {\n",
       "        text-align: right;\n",
       "    }\n",
       "</style>\n",
       "<table border=\"1\" class=\"dataframe\">\n",
       "  <thead>\n",
       "    <tr style=\"text-align: right;\">\n",
       "      <th></th>\n",
       "      <th>time_stamp</th>\n",
       "      <th>video_frame</th>\n",
       "      <th>fish_id</th>\n",
       "      <th>pixel_tracks</th>\n",
       "      <th>world_tracks</th>\n",
       "      <th>tracking_scores</th>\n",
       "      <th>ced_electrode_data</th>\n",
       "      <th>efi_electrode_data</th>\n",
       "    </tr>\n",
       "  </thead>\n",
       "  <tbody>\n",
       "    <tr>\n",
       "      <th>0</th>\n",
       "      <td>0.01120</td>\n",
       "      <td>136</td>\n",
       "      <td>1</td>\n",
       "      <td>[[[1637.6125488281, 1552.1987304688], [1686.54...</td>\n",
       "      <td>[[[102.9537891214, 8.5958438126], [106.3467447...</td>\n",
       "      <td>[4.6689133644, 4.9788632393]</td>\n",
       "      <td>[[1.8263244629000002, 1.8272399902, 1.82739257...</td>\n",
       "      <td>[[285, 299, 280, 333, 283, 261, 184, 252, 93, ...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>1</th>\n",
       "      <td>0.02490</td>\n",
       "      <td>137</td>\n",
       "      <td>2</td>\n",
       "      <td>[[[1627.2991943359, 1553.25], [1681.615234375,...</td>\n",
       "      <td>[[[102.2400654784, 8.5319933544], [106.0039933...</td>\n",
       "      <td>[4.7536907196, 4.8814210892]</td>\n",
       "      <td>[[1.8272399902, 1.8278503418, 1.8267822266, 1....</td>\n",
       "      <td>[[327, 354, 319, 323, 357, 424, 294, 409, 419,...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>2</th>\n",
       "      <td>0.06948</td>\n",
       "      <td>138</td>\n",
       "      <td>1</td>\n",
       "      <td>[[[1616.6354980469, 1557.095703125], [1676.027...</td>\n",
       "      <td>[[[101.5019092768, 8.2741293738], [105.6156216...</td>\n",
       "      <td>[4.6453266144, 4.9770188332]</td>\n",
       "      <td>[[1.8240356444999999, 1.8254089355, 1.82205200...</td>\n",
       "      <td>[[359, 343, 399, 331, 350, 376, 346, 378, 260,...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>3</th>\n",
       "      <td>0.08270</td>\n",
       "      <td>138</td>\n",
       "      <td>2</td>\n",
       "      <td>[[[1616.6354980469, 1557.095703125], [1676.027...</td>\n",
       "      <td>[[[101.5019092768, 8.2741293738], [105.6156216...</td>\n",
       "      <td>[4.6453266144, 4.9770188332]</td>\n",
       "      <td>[[1.8228149413999999, 1.8264770508, 1.82662963...</td>\n",
       "      <td>[[357, 423, 353, 366, 413, 469, 437, 478, 537,...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>4</th>\n",
       "      <td>0.13838</td>\n",
       "      <td>139</td>\n",
       "      <td>1</td>\n",
       "      <td>[[[1606.1176757812, 1557.5185546875], [1674.60...</td>\n",
       "      <td>[[[100.7743917906, 8.2542062312], [105.5168695...</td>\n",
       "      <td>[4.479634285, 5.1406679153]</td>\n",
       "      <td>[[1.8197631836000001, 1.8228149413999999, 1.82...</td>\n",
       "      <td>[[301, 290, 318, 288, 314, 279, 258, 255, 154,...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>5</th>\n",
       "      <td>0.15188</td>\n",
       "      <td>140</td>\n",
       "      <td>2</td>\n",
       "      <td>[[[1599.9650878906, 1558.1341552734], [1669.49...</td>\n",
       "      <td>[[[100.3488395774, 8.2169473611], [105.1628198...</td>\n",
       "      <td>[4.4004626274, 4.8763465881]</td>\n",
       "      <td>[[1.8260192871, 1.8212890625, 1.8229675293, 1....</td>\n",
       "      <td>[[374, 324, 459, 351, 407, 371, 387, 388, 480,...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>6</th>\n",
       "      <td>0.19700</td>\n",
       "      <td>141</td>\n",
       "      <td>2</td>\n",
       "      <td>[[[1590.3410644531, 1562.8796386719], [1663.90...</td>\n",
       "      <td>[[[99.682780818, 7.8956671369], [104.775387448...</td>\n",
       "      <td>[4.092713356, 4.5360798836]</td>\n",
       "      <td>[[1.8270874023, 1.8240356444999999, 1.82586669...</td>\n",
       "      <td>[[275, 328, 296, 305, 290, 312, 268, 297, 239,...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>7</th>\n",
       "      <td>0.20324</td>\n",
       "      <td>141</td>\n",
       "      <td>1</td>\n",
       "      <td>[[[1590.3410644531, 1562.8796386719], [1663.90...</td>\n",
       "      <td>[[[99.682780818, 7.8956671369], [104.775387448...</td>\n",
       "      <td>[4.092713356, 4.5360798836]</td>\n",
       "      <td>[[1.831817627, 1.8243408203, 1.8209838867, 1.8...</td>\n",
       "      <td>[[275, 328, 296, 305, 290, 312, 268, 297, 239,...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>8</th>\n",
       "      <td>0.23648</td>\n",
       "      <td>142</td>\n",
       "      <td>2</td>\n",
       "      <td>[[[1584.3190917969, 1567.75], [1658.6987304688...</td>\n",
       "      <td>[[[99.2658024619, 7.5624747078], [104.41480335...</td>\n",
       "      <td>[4.2152342796, 3.9297819138000003]</td>\n",
       "      <td>[[1.8272399902, 1.8301391602, 1.8284606934, 1....</td>\n",
       "      <td>[[360, 336, 360, 390, 433, 362, 407, 401, 412,...</td>\n",
       "    </tr>\n",
       "    <tr>\n",
       "      <th>9</th>\n",
       "      <td>0.27080</td>\n",
       "      <td>143</td>\n",
       "      <td>1</td>\n",
       "      <td>[[[1579.3872070312, 1568.7956542969], [1649.35...</td>\n",
       "      <td>[[[98.9247250831, 7.4942501428], [103.76764361...</td>\n",
       "      <td>[4.2730569839, 3.4880776405000002]</td>\n",
       "      <td>[[1.8276977538999999, 1.8264770508, 1.82846069...</td>\n",
       "      <td>[[286, 387, 323, 305, 433, 358, 422, 429, 467,...</td>\n",
       "    </tr>\n",
       "  </tbody>\n",
       "</table>\n",
       "</div>"
      ],
      "text/plain": [
       "   time_stamp  video_frame  fish_id  \\\n",
       "0     0.01120          136        1   \n",
       "1     0.02490          137        2   \n",
       "2     0.06948          138        1   \n",
       "3     0.08270          138        2   \n",
       "4     0.13838          139        1   \n",
       "5     0.15188          140        2   \n",
       "6     0.19700          141        2   \n",
       "7     0.20324          141        1   \n",
       "8     0.23648          142        2   \n",
       "9     0.27080          143        1   \n",
       "\n",
       "                                        pixel_tracks  \\\n",
       "0  [[[1637.6125488281, 1552.1987304688], [1686.54...   \n",
       "1  [[[1627.2991943359, 1553.25], [1681.615234375,...   \n",
       "2  [[[1616.6354980469, 1557.095703125], [1676.027...   \n",
       "3  [[[1616.6354980469, 1557.095703125], [1676.027...   \n",
       "4  [[[1606.1176757812, 1557.5185546875], [1674.60...   \n",
       "5  [[[1599.9650878906, 1558.1341552734], [1669.49...   \n",
       "6  [[[1590.3410644531, 1562.8796386719], [1663.90...   \n",
       "7  [[[1590.3410644531, 1562.8796386719], [1663.90...   \n",
       "8  [[[1584.3190917969, 1567.75], [1658.6987304688...   \n",
       "9  [[[1579.3872070312, 1568.7956542969], [1649.35...   \n",
       "\n",
       "                                        world_tracks  \\\n",
       "0  [[[102.9537891214, 8.5958438126], [106.3467447...   \n",
       "1  [[[102.2400654784, 8.5319933544], [106.0039933...   \n",
       "2  [[[101.5019092768, 8.2741293738], [105.6156216...   \n",
       "3  [[[101.5019092768, 8.2741293738], [105.6156216...   \n",
       "4  [[[100.7743917906, 8.2542062312], [105.5168695...   \n",
       "5  [[[100.3488395774, 8.2169473611], [105.1628198...   \n",
       "6  [[[99.682780818, 7.8956671369], [104.775387448...   \n",
       "7  [[[99.682780818, 7.8956671369], [104.775387448...   \n",
       "8  [[[99.2658024619, 7.5624747078], [104.41480335...   \n",
       "9  [[[98.9247250831, 7.4942501428], [103.76764361...   \n",
       "\n",
       "                      tracking_scores  \\\n",
       "0        [4.6689133644, 4.9788632393]   \n",
       "1        [4.7536907196, 4.8814210892]   \n",
       "2        [4.6453266144, 4.9770188332]   \n",
       "3        [4.6453266144, 4.9770188332]   \n",
       "4         [4.479634285, 5.1406679153]   \n",
       "5        [4.4004626274, 4.8763465881]   \n",
       "6         [4.092713356, 4.5360798836]   \n",
       "7         [4.092713356, 4.5360798836]   \n",
       "8  [4.2152342796, 3.9297819138000003]   \n",
       "9  [4.2730569839, 3.4880776405000002]   \n",
       "\n",
       "                                  ced_electrode_data  \\\n",
       "0  [[1.8263244629000002, 1.8272399902, 1.82739257...   \n",
       "1  [[1.8272399902, 1.8278503418, 1.8267822266, 1....   \n",
       "2  [[1.8240356444999999, 1.8254089355, 1.82205200...   \n",
       "3  [[1.8228149413999999, 1.8264770508, 1.82662963...   \n",
       "4  [[1.8197631836000001, 1.8228149413999999, 1.82...   \n",
       "5  [[1.8260192871, 1.8212890625, 1.8229675293, 1....   \n",
       "6  [[1.8270874023, 1.8240356444999999, 1.82586669...   \n",
       "7  [[1.831817627, 1.8243408203, 1.8209838867, 1.8...   \n",
       "8  [[1.8272399902, 1.8301391602, 1.8284606934, 1....   \n",
       "9  [[1.8276977538999999, 1.8264770508, 1.82846069...   \n",
       "\n",
       "                                  efi_electrode_data  \n",
       "0  [[285, 299, 280, 333, 283, 261, 184, 252, 93, ...  \n",
       "1  [[327, 354, 319, 323, 357, 424, 294, 409, 419,...  \n",
       "2  [[359, 343, 399, 331, 350, 376, 346, 378, 260,...  \n",
       "3  [[357, 423, 353, 366, 413, 469, 437, 478, 537,...  \n",
       "4  [[301, 290, 318, 288, 314, 279, 258, 255, 154,...  \n",
       "5  [[374, 324, 459, 351, 407, 371, 387, 388, 480,...  \n",
       "6  [[275, 328, 296, 305, 290, 312, 268, 297, 239,...  \n",
       "7  [[275, 328, 296, 305, 290, 312, 268, 297, 239,...  \n",
       "8  [[360, 336, 360, 390, 433, 362, 407, 401, 412,...  \n",
       "9  [[286, 387, 323, 305, 433, 358, 422, 429, 467,...  "
      ]
     },
     "execution_count": 2,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Getting an idea of the data. \n",
    "# (And maybe to determine the starting video frame of the eod data for future milestones)\n",
    "\n",
    "data = DatasetContainer()\n",
    "data.load(path='data/Mormyrus_Pair_01/dataset/20230316_Mormyrus_Pair_01.json.gz')\n",
    "\n",
    "# Show meta data\n",
    "print(data.meta)\n",
    "\n",
    "# Show data\n",
    "# data.data.info()\n",
    "data.data.head(10)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "(2, 6, 2)"
      ]
     },
     "execution_count": 3,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "# Electrode data sample\n",
    "data.data[\"pixel_tracks\"][0].shape"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(5456, 4, 70)\n"
     ]
    },
    {
     "data": {
      "text/plain": [
       "<Axes: >"
      ]
     },
     "execution_count": 4,
     "metadata": {},
     "output_type": "execute_result"
    },
    {
     "data": {
      "image/png": "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",
      "text/plain": [
       "<Figure size 800x600 with 1 Axes>"
      ]
     },
     "metadata": {},
     "output_type": "display_data"
    }
   ],
   "source": [
    "# Sample plot for electrode spike data\n",
    "to_plot = np.stack(data.data.ced_electrode_data.loc[:].to_numpy())\n",
    "print(to_plot.shape)\n",
    "\n",
    "fig = plt.figure(figsize=(8, 6))\n",
    "ax = fig.add_subplot()\n",
    "\n",
    "sns.lineplot(\n",
    "    pd.DataFrame({\n",
    "        '1': to_plot[0,0],\n",
    "        '2': to_plot[0,1],\n",
    "        '3': to_plot[0,2],\n",
    "        '4': to_plot[0,3],\n",
    "    }))"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "name": "stdout",
     "output_type": "stream",
     "text": [
      "(2, 7715)\n",
      "(7715, 6, 2, 2)\n"
     ]
    }
   ],
   "source": [
    "# Read .h5 or .slp file for full tracking data\n",
    "# https://sleap.ai/tutorials/analysis.html\n",
    "# https://sleap.ai/api/sleap.io.convert.html#module-sleap.io.convert\n",
    "\n",
    "import h5py\n",
    "\n",
    "with h5py.File('data/Mormyrus_Pair_01/poses/20230316_Mormyrus_Pair_01.000_20230316_Mormyrus_Pair_01.analysis.h5', 'r') as f:\n",
    "    occupancy_matrix = f['track_occupancy'][:]\n",
    "    tracks_matrix = f['tracks'][:]\n",
    "\n",
    "print(occupancy_matrix.T.shape)\n",
    "print(tracks_matrix.T.shape)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Package for the raytracing for wall information\n",
    "# https://git.imp.fu-berlin.de/bioroboticslab/robofish/fish_models/\n",
    "\n",
    "# Raycast example usage\n",
    "# https://git.imp.fu-berlin.de/bioroboticslab/robofish/fish_models/-/blob/dev_andi_train/examples/raycast_examples.ipynb"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": []
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "## Milestone 2\n",
    "### Step 1 - Preprocess data\n",
    "\n",
    "As a first step we want you to preprocess the input and output data to feed them into future models. <br>\n",
    "The input data should be preprocessed (see [Blogpost](https://shchur.github.io/blog/2021/tpp2-neural-tpps/)) to feed it into an imitation model.<br>\n",
    "\n",
    "Suggestions for the input data: <br>\n",
    "**Input Vector**: [EOD signal] <br>\n",
    "**Output Vector**: [Parameters for a Probability Density Function] <br>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Code here"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Step 2 - Prepare Dataloader\n",
    "\n",
    "Set up the dataloader to train the autoregressive imitation model on. The labels should future events of the sequence. <br>"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Code here"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Step 3 - Train Imitation model\n",
    "\n",
    "Set up and train your imitation model on your dataloader."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Code here"
   ]
  },
  {
   "cell_type": "markdown",
   "metadata": {},
   "source": [
    "### Step 4 - Visualization / Evaluation\n",
    "\n",
    "Visualize the signal generated by your model. Evaluate the model's performance."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "# Code here"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "uptodate",
   "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": "3.9.18"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}