notebooks/plot_EgoAdapt_user_action_distribution.ipynb

{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "7227d09d",
   "metadata": {},
   "outputs": [],
   "source": [
    "\"\"\"\n",
    "Plots the CDF of the action-histograms in the test user subset.\n",
    "\n",
    "Before running this experiment, make sure you have run the Stream meta-data collector (src/continual_ego4d/processing/run_summarize_user_streams.py) and set the resulting paths in this notebook.\n",
    "\"\"\"\n",
    "import datetime\n",
    "import os\n",
    "import pickle\n",
    "from collections import Counter\n",
    "\n",
    "import matplotlib.pyplot as plt\n",
    "import numpy as np\n",
    "import pandas as pd\n",
    "import seaborn as sns\n",
    "from IPython.core.interactiveshell import InteractiveShell\n",
    "\n",
    "InteractiveShell.ast_node_interactivity = \"all\" # Print all variables on their own lines\n",
    "pd.set_option('display.max_rows', 1000)\n",
    "\n",
    "# Use latex in mpl\n",
    "plt.rcParams['text.usetex'] = True\n",
    "plt.rcParams['text.latex.preamble'] = r'\\usepackage{amsmath}' #for \\text command\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "5653617d",
   "metadata": {},
   "outputs": [],
   "source": [
    "\"\"\"\n",
    "CONFIG: Add your config params here\n",
    "\"\"\"\n",
    "########### PLOT CONFIG ###############\n",
    "TRAIN_USERS_MODE=False\n",
    "\n",
    "if TRAIN_USERS_MODE:\n",
    "    # Path obtained by Stream meta-data collector (src/continual_ego4d/processing/run_summarize_user_streams.py)\n",
    "    FILE_TO_ANALYZE = \"/your/path/to/logs/2022-10-07_04-49-02_UIDa5c4c52b-a8d8-4155-b1f4-bed9cd82374e/dataset_entries_train_ego4d_LTA_train_usersplit_10users.ckpt\"\n",
    "\n",
    "else:\n",
    "    # Path obtained by Stream meta-data collector (src/continual_ego4d/processing/run_summarize_user_streams.py)\n",
    "    FILE_TO_ANALYZE = \"/your/path/to/logs/2022-10-07_04-33-34_UIDd679068a-dc6e-40ff-b146-70ffe0671a97/dataset_entries_test_ego4d_LTA_test_usersplit_40users.ckpt\"\n",
    "\n",
    "main_outdir = \"../imgs/label_histogram\" # Output path of imgs\n",
    "title = f\"LABEL_HISTOGRAM_{'TRAIN' if TRAIN_USERS_MODE else 'TEST'}\" # Title for the plot file"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "outputs": [],
   "source": [
    "\"\"\"Get dictionary with dataframe per user stream (includes all labels and other meta data).\"\"\"\n",
    "\n",
    "with open(FILE_TO_ANALYZE, 'rb') as f:\n",
    "    ds = pickle.load(f)\n",
    "\n",
    "\"\"\" Create DF \"\"\"\n",
    "def ds_to_user_dfs(ds):\n",
    "    ret = {}\n",
    "    for user, user_entries in ds.items():\n",
    "        # Do all for actions/verbs/nouns\n",
    "        user_df = pd.json_normalize(user_entries)  # Convert to DF\n",
    "\n",
    "        # Create action column\n",
    "        def label_fn(x):\n",
    "            assert len(x) == 2, \"Need two columns to merge\"\n",
    "            if not isinstance(x[0],list):\n",
    "                assert not isinstance(x[1],list)\n",
    "                return f\"{x[0]}-{x[1]}\"\n",
    "\n",
    "            return [f\"{l}-{r}\" for l,r in zip(x[0],x[1])]\n",
    "\n",
    "        user_df['action_label'] = user_df.loc[:,('verb_label','noun_label')].apply(label_fn,axis=1)\n",
    "        ret[user] = user_df\n",
    "        print(\"Created action_label column\")\n",
    "    return ret\n",
    "\n",
    "# MAIN SCRIPT\n",
    "dfs = ds_to_user_dfs(ds) # Preprocess dataframes"
   ],
   "metadata": {
    "collapsed": false,
    "pycharm": {
     "name": "#%%\n"
    }
   }
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "a598a363",
   "metadata": {
    "scrolled": false
   },
   "outputs": [],
   "source": [
    "parent_dirname = datetime.datetime.now().strftime('%Y-%m-%d_%H-%M-%S') + \"_\" + title\n",
    "parent_dirpath = os.path.join(main_outdir, parent_dirname)\n",
    "os.makedirs(parent_dirpath, exist_ok=True)\n",
    "\n",
    "plt.rcParams['font.family'] = 'DeJavu Serif'\n",
    "plt.rcParams['font.serif'] = ['Times New Roman']\n",
    "plt.rcParams['font.size'] = 16\n",
    "\n",
    "fontsize = 16\n",
    "params = {'font.family':'DeJavu Serif','font.serif':'Times New Roman',\n",
    "          'axes.labelsize': fontsize+2,'axes.titlesize':fontsize, 'font.size': fontsize, 'legend.fontsize': fontsize, 'xtick.labelsize': fontsize, 'ytick.labelsize': fontsize}\n",
    "plt.rcParams.update(params)\n",
    "\n",
    "figsize = (6,4)\n",
    "figsize = (8,3)\n",
    "xlim=(-1,None)\n",
    "\n",
    "\n",
    "# LINE CONFIG\n",
    "line_colors = sns.color_palette(\"flare\", 10) \n",
    "cumul=True\n",
    "\n",
    "if cumul:\n",
    "    ylim=(None,None)\n",
    "else:\n",
    "    ylim=(-0.1,None)\n",
    "\n",
    "if TRAIN_USERS_MODE:\n",
    "    line_alpha=0.45\n",
    "else:\n",
    "    line_alpha=0.3\n",
    "\n",
    "\n",
    "# BARCHART CONFIG\n",
    "plot_yerror = None \n",
    "bar_color = 'black' #my_cmap.colors if bar_colors is None else bar_colors[plot_idx]\n",
    "legend_label = None\n",
    "bar_align='center'\n",
    "log=False\n",
    "legend_labels=None\n",
    "grid = False\n",
    "bar_alpha = 0.1\n",
    "bar_marker_offset=0\n",
    "bar_width=0.4\n",
    "\n",
    "class PlotEntry:\n",
    "    def __init__(self):\n",
    "        self.x_vals = None\n",
    "        self.y_vals = None\n",
    "        self.label_name = None\n",
    "\n",
    "\"\"\" Plot average histogram over user's actions/verbs/nouns.\"\"\"\n",
    "def plot_user_histogram_lines_and_avg_bars(dfs, label_col_name):\n",
    "    \"\"\"\"\"\"\n",
    "    nb_users = len(dfs)\n",
    "    fig, ax = plt.subplots(figsize=figsize, dpi=600)\n",
    "    \n",
    "    if 'action' in label_col_name:\n",
    "        bar_width=0.3\n",
    "        xlabel='action (high to low frequency per user)'\n",
    "        ylabel = r\"$\\text{CDF}_\\text{action}$\" if cumul else r\"$P_\\text{action}$\"\n",
    "        \n",
    "        \n",
    "    elif 'verb' in label_col_name:\n",
    "        bar_width=0.8\n",
    "        xlabel='verb (high to low frequency per user)'\n",
    "        ylabel = r\"$\\text{CDF}_\\text{verb}$\" if cumul else r\"$P_\\text{verb}$\"\n",
    "\n",
    "    elif 'noun' in label_col_name:\n",
    "        bar_width=0.8\n",
    "        xlabel='noun (high to low frequency per user)'\n",
    "        ylabel = r\"$\\text{CDF}_\\text{noun}$\" if cumul else r\"$P_\\text{noun}$\"\n",
    "\n",
    "    else:\n",
    "        raise ValueError()\n",
    "    \n",
    "    # Create per user a plot-entry\n",
    "    user_avg_dict_list = []\n",
    "    \n",
    "    # Iterate users, get per-users histogram and plot line\n",
    "    for user_idx, (user_id, user_df) in enumerate(dfs.items()):\n",
    "#         print(f\"Idx {user_idx}: User {user_id}\")\n",
    "        cnt = Counter(user_df[label_col_name].tolist())\n",
    "\n",
    "        col_sorted = sorted([(k,v) for k,v in cnt.items() ] ,key=lambda x: x[1], reverse=True)\n",
    "        vals_sorted = [x[0] for x in col_sorted] # Action/verb/noun ids\n",
    "        cnts_sorted = [x[1] for x in col_sorted] # Frequencies (counts)\n",
    "        \n",
    "        # Normalize counts\n",
    "        total_count = sum(cnts_sorted)\n",
    "        cnts_sorted_n = [float(x)/total_count for x in cnts_sorted]\n",
    "        \n",
    "        # Plot\n",
    "        y_vals= cnts_sorted_n\n",
    "        x_vals= list(range(len(cnts_sorted_n)))\n",
    "        \n",
    "        y_vals_for_avg = y_vals\n",
    "        if cumul:\n",
    "            y_vals_c =  [y_vals[0]]\n",
    "            for idx in range(1,len(y_vals)):\n",
    "                y_vals_c.append(y_vals_c[-1]+y_vals[idx])\n",
    "            y_vals = y_vals_c\n",
    "        \n",
    "        color = line_colors[user_idx % len(line_colors)]\n",
    "        plt.plot(x_vals,y_vals,alpha=line_alpha,color=color)\n",
    "        \n",
    "        user_avg_dict_list.append({\n",
    "        'user_id':user_id,'y_vals':y_vals_for_avg\n",
    "        })\n",
    "\n",
    "#         print(f\"Freqs for {label_col_name}: {cnts_sorted}\")\n",
    "#         print(f\"Labels for {label_col_name}: {vals_sorted}\")\n",
    "\n",
    "    # Get avg of all user histograms + SE -> Barplot\n",
    "    # Get max length, pad others\n",
    "    max_val_len = max(len(d['y_vals']) for d in user_avg_dict_list)\n",
    "    \n",
    "    arrays_to_avg = []\n",
    "    for d in user_avg_dict_list:\n",
    "        zeros_to_append = max_val_len - len(d['y_vals'])\n",
    "        if zeros_to_append > 0:\n",
    "            d['y_vals'] = d['y_vals'] + [0]*zeros_to_append\n",
    "        arrays_to_avg.append(d['y_vals'])\n",
    "    \n",
    "    arrays_to_avg_np = np.asarray(arrays_to_avg)\n",
    "    \n",
    "    avg_yvals = np.mean(arrays_to_avg_np,axis=0) # keep 0-dim\n",
    "    avg_xvals = list(range(len(avg_yvals)))\n",
    "\n",
    "    # Take cumul of mean\n",
    "    if cumul:\n",
    "        avg_yvals_c =  [avg_yvals[0]]\n",
    "        for idx in range(1,len(avg_yvals)):\n",
    "            avg_yvals_c.append(avg_yvals_c[-1]+avg_yvals[idx])\n",
    "        avg_yvals = avg_yvals_c\n",
    "\n",
    "\n",
    "    plt.bar(avg_xvals, height=avg_yvals, yerr=plot_yerror, label=legend_label,\n",
    "            color=bar_color, align=bar_align, width=bar_width, log=log,\n",
    "            alpha=bar_alpha,\n",
    "#                linewidth=0.1,edgecolor='white'\n",
    "           )\n",
    "\n",
    "    x_center = [el+bar_marker_offset for el in avg_xvals]\n",
    "    plt.plot(x_center, avg_yvals, marker=\"v\",markersize=1.6,\n",
    "             linestyle=\"\", alpha=1, color=\"black\",label=r'avg. $\\mathcal{U}_\\text{test}$')\n",
    "    \n",
    "    leg = plt.legend(fontsize=18, markerscale=3, loc='lower right', prop={'size':18})\n",
    "\n",
    "    plt.axhline(y=1, color='black', linestyle=':',alpha=0.6,linewidth=0.4)\n",
    "    \n",
    "    # Hide the right and top spines\n",
    "    ax.spines.right.set_visible(False)\n",
    "    ax.spines.top.set_visible(False)\n",
    "    \n",
    "    # Only show ticks on the left and bottom spines\n",
    "    ax.yaxis.set_ticks_position('left')\n",
    "    ax.xaxis.set_ticks_position('bottom')\n",
    "\n",
    "    plt.xlim(*xlim)\n",
    "    plt.ylim(*ylim)\n",
    "    plt.xlabel(xlabel)\n",
    "    plt.ylabel(ylabel)\n",
    "#     print(title)\n",
    "#     plt.title(title)\n",
    "    plt.grid(grid, which='both')\n",
    "    fig.tight_layout() \n",
    "\n",
    "    # Save\n",
    "    if parent_dirpath is not None:\n",
    "        filename = f\"{title}_{xlabel}.pdf\"\n",
    "        filepath = os.path.join(parent_dirpath,filename)\n",
    "        fig.savefig(filepath, bbox_inches='tight')\n",
    "        print(f\"Saved plot: {filepath}\")\n",
    "\n",
    "    plt.show()\n",
    "    plt.clf()\n",
    "    \n",
    "\n",
    "\n",
    "# Grids of histograms\n",
    "for action_mode in  ['verb_label','noun_label','action_label']:\n",
    "    plot_user_histogram_lines_and_avg_bars(dfs, action_mode)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "7ed694f1",
   "metadata": {},
   "outputs": [],
   "source": []
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3 (ipykernel)",
   "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": "3.9.12"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}