{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "9fe02137-ba1c-4ced-9909-b5b7c39ed6d4",
   "metadata": {},
   "source": [
    "# Merging state/county/city polygons with party affiliation and landvote data"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "6452373c-d10f-472c-9174-cd05a3363587",
   "metadata": {},
   "outputs": [],
   "source": [
    "import ibis\n",
    "from ibis import _\n",
    "\n",
    "import streamlit as st\n",
    "import ibis.expr.datatypes as dt  # Make sure to import the necessary module\n",
    "\n",
    "conn = ibis.duckdb.connect(extensions=[\"spatial\"])\n",
    "\n",
    "landvote_url = \"https://huggingface.co/datasets/boettiger-lab/landvote/resolve/main/landvote_polygons.parquet\"\n",
    "# party_url = \"https://huggingface.co/datasets/boettiger-lab/landvote/resolve/main/party_polygons.parquet\"\n",
    "party_url = \"https://huggingface.co/datasets/boettiger-lab/landvote/resolve/main/party_polygons_all.parquet\"\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "dd9ffc64-b4cf-4e5b-9c96-703e91a77837",
   "metadata": {},
   "outputs": [],
   "source": [
    "landvote = (conn\n",
    "            .read_parquet(landvote_url)\n",
    "            .cast({\"geometry\": \"geometry\"})\n",
    "            .mutate(county = _.county.upper())\n",
    "            .mutate(municipal = _.municipal.upper())\n",
    "            .mutate(elect_year = _.year - _.year % 4) # get most recent election year \n",
    "            .cast({\"municipal\": \"string\",\"county\":\"string\"})\n",
    "            .mutate(municipal=ibis.case()\n",
    "                 .when(_.jurisdiction.isin(['State','County']), ibis.literal(\"-\")) \n",
    "                 .else_(_.municipal) \n",
    "                 .end()\n",
    "                 )\n",
    "            .mutate(county=ibis.case()\n",
    "                 .when(_.jurisdiction.isin(['State']), ibis.literal(\"-\"))\n",
    "                 .else_(ibis.case()\n",
    "                         .when(_.county.endswith('COUNTY'), _.county)\n",
    "                         .else_(_.county + ' COUNTY')\n",
    "                         .end())\n",
    "                 .end())\n",
    "           )\n",
    "\n",
    "party = (conn\n",
    "        .read_parquet(party_url)\n",
    "        .cast({\"geometry\": \"geometry\",\"municipal\":\"string\"})\n",
    "        .mutate(municipal=ibis.case()\n",
    "                .when(_.jurisdiction.isin(['State','County']), ibis.literal(\"-\")) \n",
    "                .else_(_.municipal) \n",
    "                .end()\n",
    "                )\n",
    "        .mutate(county=ibis.case()\n",
    "                .when(_.jurisdiction.isin(['State']), ibis.literal(\"-\")) \n",
    "                .else_(_.county) \n",
    "                .end()\n",
    "                )\n",
    "        )"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "8fc40e96-fffd-4b23-9963-c931fdce96f6",
   "metadata": {},
   "outputs": [],
   "source": [
    "votes = (landvote\n",
    "        .join(party,[\"state\",\"county\",\"municipal\",\"jurisdiction\",\"geometry\", _.elect_year == party[\"year\"]],how = \"left\")\n",
    "        .drop('elect_year','state_right','county_right','municipal_right','year_right',\"geometry_right\",\"jurisdiction_right\")\n",
    "        .mutate(municipal=ibis.case()\n",
    "                .when(_.municipal == ibis.literal(\"-\"), None) \n",
    "                .else_(_.municipal) \n",
    "                .end()\n",
    "                )\n",
    "        .mutate(county=ibis.case()\n",
    "                .when(_.county == ibis.literal(\"-\"), None) \n",
    "                .else_(_.county) \n",
    "                .end()\n",
    "                )\n",
    "        )"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "e80cfd2e-40eb-4065-9ae6-dcaf83319d9a",
   "metadata": {},
   "source": [
    "# Make PMTiles. Each jurisdiction type is its own layer"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "b1cd8f44-57fa-49a8-b438-f9e4aab747c5",
   "metadata": {},
   "outputs": [],
   "source": [
    "import subprocess\n",
    "import os\n",
    "from huggingface_hub import HfApi, login\n",
    "import streamlit as st\n",
    "\n",
    "login(st.secrets[\"HF_TOKEN\"])\n",
    "# api = HfApi(add_to_git_credential=False)\n",
    "api = HfApi()\n",
    "\n",
    "def hf_upload(file, repo_id):\n",
    "    info = api.upload_file(\n",
    "            path_or_fileobj=file,\n",
    "            path_in_repo=file,\n",
    "            repo_id=repo_id,\n",
    "            repo_type=\"dataset\",\n",
    "        )\n",
    "def generate_pmtiles(input_file, input_file2, input_file3, output_file, max_zoom=12):\n",
    "    # Ensure Tippecanoe is installed\n",
    "    if subprocess.call([\"which\", \"tippecanoe\"], stdout=subprocess.DEVNULL) != 0:\n",
    "        raise RuntimeError(\"Tippecanoe is not installed or not in PATH\")\n",
    "\n",
    "    # Construct the Tippecanoe command\n",
    "    command = [\n",
    "        \"tippecanoe\",\n",
    "        \"-o\", output_file,\n",
    "        \"-zg\",\n",
    "        \"--extend-zooms-if-still-dropping\",\n",
    "        \"--force\",\n",
    "        \"--projection\", \"EPSG:4326\",  \n",
    "        \"-L\",\"state:\"+input_file,\n",
    "        \"-L\",\"county:\"+input_file2,\n",
    "        \"-L\",\"municipal:\"+input_file3\n",
    "    ]\n",
    "    # Run Tippecanoe\n",
    "    try:\n",
    "        subprocess.run(command, check=True)\n",
    "        print(f\"Successfully generated PMTiles file: {output_file}\")\n",
    "    except subprocess.CalledProcessError as e:\n",
    "        print(f\"Error running Tippecanoe: {e}\")\n",
    "\n"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "7061577e-0632-4395-8ac5-241a1fab53b0",
   "metadata": {},
   "outputs": [],
   "source": [
    "gdf_state = votes.filter(_.jurisdiction == 'State').execute().set_crs(\"EPSG:4326\")\n",
    "gdf_state.to_file(\"votes_state.geojson\")\n",
    "\n",
    "gdf_county = votes.filter(_.jurisdiction == 'County').execute().set_crs(\"EPSG:4326\")\n",
    "gdf_county.to_file(\"votes_county.geojson\")\n",
    "\n",
    "gdf_city = votes.filter(_.jurisdiction == 'Municipal').execute().set_crs(\"EPSG:4326\")\n",
    "gdf_city.to_file(\"votes_municipal.geojson\")\n",
    "\n",
    "generate_pmtiles(\"votes_state.geojson\", \"votes_county.geojson\",\"votes_municipal.geojson\", \"votes.pmtiles\")\n",
    "hf_upload(\"votes.pmtiles\", \"boettiger-lab/landvote\")"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "f2979624-bcdf-4a8a-899a-c22fc3cdaf0e",
   "metadata": {},
   "outputs": [],
   "source": [
    "# save as parquet\n",
    "votes.execute().set_crs(\"EPSG:4326\").to_parquet(\"votes.parquet\")\n",
    "hf_upload(\"votes.parquet\", \"boettiger-lab/landvote\")\n"
   ]
  }
 ],
 "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": "3.12.7"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}