{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "9a51a013-8798-4327-bd2b-5aaf0427cda4",
   "metadata": {},
   "source": [
    "# pvcaptest"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "50531b9f-821b-4f17-ae11-535b6241dee6",
   "metadata": {},
   "outputs": [],
   "source": [
    "import utils\n",
    "import pandas as pd\n",
    "\n",
    "from bokeh.plotting import figure, show\n",
    "from bokeh.models import HoverTool\n",
    "from bokeh.transform import jitter\n",
    "from bokeh.io import output_notebook"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "c987023b-ca0a-4ebf-9b9d-33793d3cf8f8",
   "metadata": {},
   "outputs": [],
   "source": [
    "output_notebook()"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "10d7f560-2768-42ac-a21a-258c6b2ffdbe",
   "metadata": {},
   "source": [
    "## ReadTheDocs"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "6c7a30e9-97a4-4344-b832-ba526d2836e9",
   "metadata": {},
   "outputs": [],
   "source": [
    "df = utils.get_rtd_analytics_data('pvcaptest')\n",
    "data_period = df['Date'].min().strftime('%Y-%m-%d') + ' to ' + df['Date'].max().strftime('%Y-%m-%d')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "60c947c9-f9c2-4ccf-b5e7-c9c9f3359e05",
   "metadata": {},
   "outputs": [],
   "source": [
    "total_by_version = df.groupby('Version')['Views'].sum().reset_index()\n",
    "\n",
    "p = figure(x_range=total_by_version['Version'], height=350, tooltips=[(\"Version\", \"@Version\"), (\"Views\", \"@Views\")],\n",
    "           title=f\"Page views by RTD version ({data_period})\",)\n",
    "p.vbar(x='Version', top='Views', width=0.75, source=total_by_version,\n",
    "       line_color='white')\n",
    "p.xaxis.major_label_orientation = 3.14/2\n",
    "p.yaxis.axis_label = 'Total page views'\n",
    "show(p)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "2104961b-ff35-43b3-bf0b-dfdf7c932c55",
   "metadata": {},
   "outputs": [],
   "source": [
    "daily_views = df.pivot_table(values='Views', index='Date', columns='Version', aggfunc='sum')[['stable', 'latest']].fillna(0)\n",
    "\n",
    "p = figure(x_axis_type=\"datetime\", height=350, title=f\"Daily views by RTD version\")\n",
    "hover_tool = HoverTool(tooltips=[('Date', '@x{%Y-%m-%d}'), ('Views', '@y')],\n",
    "                       formatters={'@x': 'datetime'})\n",
    "hover_tool.point_policy = 'snap_to_data'\n",
    "p.add_tools(hover_tool)\n",
    "p.line(daily_views.index, daily_views['stable'], legend_label='stable')\n",
    "p.line(daily_views.index, daily_views['latest'], legend_label='latest', color='#ff7f0e')\n",
    "p.yaxis.axis_label = 'Daily page views'\n",
    "show(p)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "2413e533-b4fa-4b2b-b308-a4bd9a4254d0",
   "metadata": {},
   "outputs": [],
   "source": [
    "df2 = df.loc[df['Version'] == 'latest', :].copy()\n",
    "pathviews = df2.groupby('Path')['Views'].sum().reset_index()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "29f9acd5-f445-4aa9-a0c8-aa80b7af611d",
   "metadata": {},
   "outputs": [],
   "source": [
    "n = 20\n",
    "subset = pathviews.sort_values('Views', ascending=False)[:n]\n",
    "p = figure(y_range=subset['Path'], height=400, tooltips=[(\"URL\", \"@Path\"), (\"Views\", \"@Views\")],\n",
    "           title=f\"Views by URL (Top {n}, {data_period})\")\n",
    "p.hbar(y='Path', right='Views', source=subset, height=0.75,\n",
    "       line_color='white')\n",
    "p.xaxis.axis_label = 'Total page views'\n",
    "show(p)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "b5d056f6-793c-4423-9598-c97349def706",
   "metadata": {},
   "outputs": [],
   "source": [
    "prefixes = ['/examples', ]\n",
    "pathviews['Prefix'] = pathviews['Path'].str.rsplit(\"/\", 1).str[0]\n",
    "groupviews = pathviews.loc[pathviews['Prefix'].isin(prefixes), :].groupby('Prefix')['Views'].sum().loc[prefixes].sort_values().reset_index()\n",
    "\n",
    "p = figure(x_range=groupviews['Prefix'], height=350, tooltips=[(\"Section\", \"@Prefix\"), (\"Views\", \"@Views\")],\n",
    "           title=f\"Page views by docs section ({data_period})\")\n",
    "p.vbar(x='Prefix', top='Views', width=0.75, source=groupviews)\n",
    "p.yaxis.axis_label = 'Total page views'\n",
    "show(p)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "b9cd913d-b115-4212-880a-4d86dbe05d62",
   "metadata": {},
   "outputs": [],
   "source": [
    "subset = pathviews.loc[pathviews['Prefix'].isin(prefixes), :]\n",
    "\n",
    "p = figure(x_range=prefixes, height=350, tooltips=[(\"URL\", \"@Path\"), (\"Views\", \"@Views\")],\n",
    "           title=f\"Page views by docs page ({data_period})\")\n",
    "p.scatter(x=jitter('Prefix', width=0.1, range=p.x_range, distribution='normal'),\n",
    "          y='Views', source=subset)\n",
    "p.yaxis.axis_label = 'Total page views'\n",
    "show(p)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "a3ef457c-5c98-4250-9507-bff27d257b69",
   "metadata": {},
   "outputs": [],
   "source": [
    "for prefix in prefixes:\n",
    "    subset = pathviews.loc[pathviews['Prefix'] == prefix, :]\n",
    "    subset = subset.loc[~subset['Path'].str.endswith('/index.html'), :]\n",
    "    subset = subset.sort_values('Views', ascending=False)[:n]\n",
    "    subset['Path_Clean'] = subset['Path'].str.replace('^'+prefix+'/', '', regex=True)\n",
    "\n",
    "    height = (400 // 5) * len(subset)\n",
    "    extra = f\"Top {n}, \" if len(subset) == n else ''\n",
    "    p = figure(y_range=subset['Path_Clean'], width=750, height=height, tooltips=[(\"URL\", \"@Path\"), (\"Views\", \"@Views\")],\n",
    "               title=f\"{prefix}: Views by URL ({extra}{data_period})\")\n",
    "    p.hbar(y='Path_Clean', right='Views', source=subset, height=0.75,\n",
    "           line_color='white')\n",
    "    p.xaxis.axis_label = 'Total page views'\n",
    "    show(p)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "5451b58c-6550-41c1-849d-afb5bc1260a1",
   "metadata": {},
   "outputs": [],
   "source": [
    "release_dates = {\n",
    "    'v0.10.0': '2021-07-25',\n",
    "}\n",
    "stable = df.loc[df['Version'] == 'stable', :].copy()\n",
    "stable['Real_Version'] = '?'\n",
    "for version, date in release_dates.items():\n",
    "    stable.loc[stable['Date'] >= date, 'Real_Version'] = version\n",
    "\n",
    "pivot = stable.pivot_table(values='Views', index='Date', columns='Real_Version', aggfunc='sum')\n",
    "realigned = pd.DataFrame({\n",
    "    version: pd.Series(pivot.loc[date:, version].values, index=(pivot.loc[date:, :].index - pd.to_datetime(date)).days)\n",
    "    for version, date in release_dates.items()\n",
    "}).cumsum()\n",
    "\n",
    "tab10 = ['#1f77b4', '#ff7f0e', '#2ca02c', '#d62728', '#9467bd', '#8c564b', '#e377c2', '#7f7f7f', '#bcbd22', '#17becf']\n",
    "\n",
    "p = figure(height=350, title=f\"Cumulative views since release date ({data_period})\")\n",
    "hover_tool = HoverTool(tooltips=[('Days since release', '@x'), ('Cumulative Views', '@y')])\n",
    "hover_tool.point_policy = 'snap_to_data'\n",
    "p.add_tools(hover_tool)\n",
    "\n",
    "for i, column in enumerate(realigned.columns):\n",
    "    label = column\n",
    "    if column == 'v0.1.1':\n",
    "        label = label + ' (partial)'\n",
    "    p.line(realigned.index, realigned[column], legend_label=label, color=tab10[i])\n",
    "\n",
    "p.yaxis.axis_label = 'Total page views'\n",
    "p.xaxis.axis_label = 'Days since release'\n",
    "p.legend.location = \"top_left\"\n",
    "show(p)"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "23e08435-cb15-4c27-93db-2c7a9617a77f",
   "metadata": {},
   "source": [
    "## GitHub"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "e9959e5d-5244-448b-a120-08d038b5c804",
   "metadata": {},
   "outputs": [],
   "source": [
    "gh_proj_path = 'pvcaptest/pvcaptest'"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "0a7abd61-24ed-46e1-9686-4af7333fbf76",
   "metadata": {},
   "outputs": [],
   "source": [
    "gh = utils.get_github_stars(gh_proj_path)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "8c18098d-7eb0-4a4b-b807-4bcbe74d8f48",
   "metadata": {},
   "outputs": [],
   "source": [
    "p = utils.plot_github_stars_timeseries(gh)\n",
    "show(p)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "f8de7c4e-b91e-4c6d-971f-3ff0cf6a3a5d",
   "metadata": {},
   "outputs": [],
   "source": [
    "contributor_ts, _ = utils.get_github_contributor_timeseries(gh_proj_path)\n",
    "p = utils.plot_github_contributors_timeseries(contributor_ts)\n",
    "show(p)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "a3aa29df-57d2-442d-bc0f-857b0904efa0",
   "metadata": {},
   "outputs": [],
   "source": [
    "contributors = utils.get_github_contributors(gh_proj_path)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "805f57a0-2301-4fb8-8e17-92c038d06c90",
   "metadata": {},
   "outputs": [],
   "source": [
    "mosaic = utils.make_github_contributors_mosaic(contributors)\n",
    "mosaic"
   ]
  }
 ],
 "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": "3.7.11"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}