profiling-viz-pairwise.ipynb

{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "b3dc3efb-3938-40f3-8e20-dddb285cd32c",
   "metadata": {},
   "source": [
    "# Notebook level metrics from `jdaviz_profiler`\n",
    "\n",
    "Read the metrics table:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "646b6274-1f8b-493f-afda-41250989b262",
   "metadata": {},
   "outputs": [],
   "source": [
    "import numpy as np\n",
    "import matplotlib.pyplot as plt\n",
    "from astropy.table import Table\n",
    "\n",
    "nb_path = '/Users/bmmorris/Downloads/metrics/2025_11_04/notebook_metrics_105948596227666.csv'\n",
    "notebook_metrics = Table.read(nb_path, delimiter=',', format='csv')"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6e4ec367-7731-43fd-b2a4-ad6a3d67fda6",
   "metadata": {},
   "source": [
    "Gather column names of metrics and parameters:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "3290905c-965f-4fbc-bb43-484a8ba7f9c5",
   "metadata": {},
   "outputs": [],
   "source": [
    "metrics = [name for name in notebook_metrics.colnames if name.endswith('metric')]\n",
    "params = [name for name in notebook_metrics.colnames if name.endswith('param')]"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "c0f89f0a-1aec-455b-b9e2-eaed8affe7ea",
   "metadata": {},
   "source": [
    "Exclude runs that did not complete:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "3e5e9a86-6172-4a68-a667-cc3316712b15",
   "metadata": {},
   "outputs": [],
   "source": [
    "timed_out = notebook_metrics['total_execution_time_metric'] > 300\n",
    "not_completed = notebook_metrics['total_cells'] != notebook_metrics['executed_cells']\n",
    "completed_runs = ~(timed_out | not_completed)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "a2ed7fb6-6876-4eda-941b-a771e25e5c92",
   "metadata": {},
   "source": [
    "Make pairwise plots for these parameters, over each of the metrics:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "7b386962-7657-4efc-a61f-97cc314ba0e6",
   "metadata": {},
   "outputs": [],
   "source": [
    "pair_params = [\n",
    "    'image_pixel_side_param',\n",
    "    'viewport_pixel_size_param',\n",
    "    'n_images_param',\n",
    "    'image_compression_param',\n",
    "]"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "3eaa7915-04a5-412a-ac09-ad0f891bc499",
   "metadata": {},
   "source": [
    "Make pairwise plots for all metrics _except_ the \"min\" metrics:"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "295d3b45-a8f4-4b82-b881-9769a7ffc7e8",
   "metadata": {},
   "outputs": [],
   "source": [
    "for metric in metrics:\n",
    "    if '_min_' in metric:\n",
    "        continue\n",
    "\n",
    "    fig, ax = plt.subplots(\n",
    "        len(pair_params), len(pair_params), figsize=(12, 12),\n",
    "        sharex='col', sharey='row', layout='constrained'\n",
    "    )\n",
    "    \n",
    "    vmin, vmax = np.percentile(np.array(notebook_metrics[metric]), [15, 85])\n",
    "    cbar_limits = dict(\n",
    "        vmin=vmin,\n",
    "        vmax=vmax\n",
    "    )\n",
    "    for i, param_i in enumerate(pair_params):\n",
    "        for j, param_j in enumerate(pair_params):\n",
    "    \n",
    "            x = np.sort(list(set(notebook_metrics[completed_runs][param_i])))\n",
    "            y = np.sort(list(set(notebook_metrics[completed_runs][param_j])))\n",
    "\n",
    "            grid = np.zeros((len(x), len(y)))\n",
    "            for k in range(len(x)):\n",
    "                for ell in range(len(y)):\n",
    "                    condition = (\n",
    "                        (x[k] == notebook_metrics[param_i]) & \n",
    "                        (y[ell] == notebook_metrics[param_j])\n",
    "                    )\n",
    "                    grid[k, ell] = np.median(notebook_metrics[metric][condition])\n",
    "    \n",
    "            if i != j:\n",
    "                cax = ax[i, j].pcolormesh(\n",
    "                    y, x, grid, shading='nearest',\n",
    "                    alpha=0.1 if i == j else 1,\n",
    "                    **cbar_limits\n",
    "                )\n",
    "            if not isinstance(x[0], str):\n",
    "                ax[i, j].set(\n",
    "                    ylim=[x.min(), x.max()],\n",
    "                )\n",
    "    \n",
    "            if not isinstance(y[0], str):\n",
    "                ax[i, j].set(\n",
    "                    xlim=[y.min(), y.max()]\n",
    "                )\n",
    "    \n",
    "            if j == 0:        \n",
    "                ax[i, j].set(\n",
    "                    ylabel=param_i,\n",
    "                )\n",
    "                    \n",
    "            if i == len(pair_params) - 1:\n",
    "                ax[i, j].set(\n",
    "                    xlabel=param_j,\n",
    "                )\n",
    "            \n",
    "            if 'image_compression_param' not in [param_i, param_j]:\n",
    "                ax[i, j].set(\n",
    "                    xscale='log',\n",
    "                    yscale='log'\n",
    "                )\n",
    "    \n",
    "    fig.colorbar(cax, ax=ax[0, -1], label=metric)\n",
    "    plt.show()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "8746ca8d-be93-4ade-a16e-9c815a1f3364",
   "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": "3.12.9"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}