{
 "cells": [
  {
   "cell_type": "markdown",
   "id": "3faa9d74",
   "metadata": {},
   "source": [
    "# 🎲 Betting odds with `sportsdataverse-py`\n",
    "\n",
    "Welcome! In a few lines of Python you're about to pull **live betting odds**\n",
    "from a whole market of sportsbooks — moneylines, spreads, totals, player\n",
    "props, scores, even point-in-time history. `sportsdataverse.odds` wraps\n",
    "[The Odds API](https://the-odds-api.com) v4 and hands you back tidy **polars**\n",
    "DataFrames that are ready to model. 🚀\n",
    "\n",
    "If you've used the R package [oddsapiR](https://oddsapir.sportsdataverse.org),\n",
    "the `toa_*` names will feel right at home. Let's dive in!"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ed7715ad",
   "metadata": {},
   "source": [
    "## 🧰 The toolbox\n",
    "\n",
    "Every function returns a tidy **polars** `DataFrame` by default — pass\n",
    "`return_as_pandas=True` for pandas, or `return_parsed=False` for the raw JSON.\n",
    "Here's the whole kit (click any name for the full reference):\n",
    "\n",
    "| Function | What it gives you | Quota |\n",
    "|---|---|---|\n",
    "| [`toa_sports`](../odds/reference/additional.md#toa_sports) | Every in-season sport/league key (the `sport=` value) | 🆓 free |\n",
    "| [`toa_sports_odds`](../odds/reference/additional.md#toa_sports_odds) | **Current odds** for a sport — one row per outcome | 💳 paid |\n",
    "| [`toa_event_odds`](../odds/reference/additional.md#toa_event_odds) | Odds for a **single game**, including player props | 💳 paid |\n",
    "| [`toa_event_markets`](../odds/reference/additional.md#toa_event_markets) | Which markets a game has on offer | 🆓 free |\n",
    "| [`toa_sports_scores`](../odds/reference/additional.md#toa_sports_scores) | Live + recently-completed **scores** | 🆓 free |\n",
    "| [`toa_sports_events`](../odds/reference/additional.md#toa_sports_events) | Upcoming + live **event list** (grab `event_id`s here) | 🆓 free |\n",
    "| [`toa_sports_participants`](../odds/reference/additional.md#toa_sports_participants) | Teams / participants for a sport | 🆓 free |\n",
    "| [`toa_sports_odds_history`](../odds/reference/additional.md#toa_sports_odds_history) | **Historical** odds snapshot (paid plans) | 💳 paid |\n",
    "| [`toa_sports_events_history`](../odds/reference/additional.md#toa_sports_events_history) | Historical event snapshot | 💳 paid |\n",
    "| [`toa_event_odds_history`](../odds/reference/additional.md#toa_event_odds_history) | Historical single-game odds | 💳 paid |\n",
    "| [`toa_usage`](../odds/reference/additional.md#toa_usage) | Your remaining quota (reads cached headers) | 🆓 free |"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "591a97bc",
   "metadata": {},
   "source": [
    "## 🔑 Setup\n",
    "\n",
    "```sh\n",
    "pip install sportsdataverse\n",
    "```\n",
    "\n",
    "The Odds API needs a key — grab a free one at\n",
    "[the-odds-api.com](https://the-odds-api.com/#get-access). Set it once as the\n",
    "`ODDS_API_KEY` environment variable (the same name `oddsapiR` uses) or pass\n",
    "`api_key=` to any call. The live cells below run only when a key is present,\n",
    "so this page is happy either way. 😊"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "6d9843e2",
   "metadata": {},
   "outputs": [],
   "source": [
    "import os\n",
    "import polars as pl\n",
    "import sportsdataverse.odds as odds\n",
    "\n",
    "HAS_KEY = bool(os.environ.get(\"ODDS_API_KEY\"))\n",
    "print(\"ODDS_API_KEY set:\", HAS_KEY, \"— live cells will\" + (\"\" if HAS_KEY else \" NOT\") + \" run\")"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "95ae4526",
   "metadata": {},
   "source": [
    "## 🗂️ What's on the board?\n",
    "\n",
    "Start with [`toa_sports`](../odds/reference/additional.md#toa_sports) — it lists every sport/league key,\n",
    "and it's **free** (doesn't touch your quota). The `key` column is what you\n",
    "pass as `sport=` everywhere else."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "aaf33df5",
   "metadata": {},
   "outputs": [],
   "source": [
    "if HAS_KEY:\n",
    "    sports = odds.toa_sports(all_sports=True)\n",
    "    out = sports.select([c for c in [\"key\", \"group\", \"title\", \"active\"] if c in sports.columns]).head(12)\n",
    "else:\n",
    "    out = \"set ODDS_API_KEY to run: odds.toa_sports(all_sports=True)\"\n",
    "out"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "48cc8642",
   "metadata": {},
   "source": [
    "## 💰 The main event: live odds\n",
    "\n",
    "[`toa_sports_odds`](../odds/reference/additional.md#toa_sports_odds) is the workhorse. It returns **long\n",
    "format** — one row per *event × bookmaker × market × outcome* — which is\n",
    "exactly the shape you want for filtering and modelling. Knobs:\n",
    "\n",
    "- `regions` — bookmaker regions: `us`, `us2`, `uk`, `eu`, `au` (comma-separate to mix).\n",
    "- `markets` — `h2h` (moneyline), `spreads`, `totals`, `outrights`, … (comma-separated).\n",
    "- `odds_format` — `american` or `decimal`.\n",
    "- `bookmakers` — pin specific books (takes precedence over `regions`)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "be784b2d",
   "metadata": {},
   "outputs": [],
   "source": [
    "if HAS_KEY:\n",
    "    board = odds.toa_sports_odds(sport=\"americanfootball_nfl\", regions=\"us\", markets=\"h2h,spreads\")\n",
    "    keep = [\"home_team\", \"away_team\", \"bookmaker_key\", \"market_key\", \"outcome_name\", \"outcome_point\", \"outcome_price\"]\n",
    "    out = board.select([c for c in keep if c in board.columns]).head(10)\n",
    "else:\n",
    "    board = None\n",
    "    out = \"set ODDS_API_KEY to run: odds.toa_sports_odds(sport='americanfootball_nfl', regions='us')\"\n",
    "out"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9c12c1ae",
   "metadata": {},
   "source": [
    "## 🍳 Cookbook: common odds tasks\n",
    "\n",
    "Because everything is one tidy long frame, the fun stuff is just a few polars\n",
    "expressions away. Twelve recipes you'll reach for constantly — every live\n",
    "cell is key-guarded, so the page renders fine with or without a key."
   ]
  },
  {
   "cell_type": "markdown",
   "id": "465e0be0",
   "metadata": {},
   "source": [
    "### Recipe 1 — Best available moneyline (line shopping 🛒)\n",
    "\n",
    "For each team, find the **highest** moneyline price across every book — and\n",
    "which book is offering it. Sort by price descending, group, take the top."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "5d0b11a5",
   "metadata": {},
   "outputs": [],
   "source": [
    "if HAS_KEY and board is not None:\n",
    "    h2h = board.filter(pl.col(\"market_key\") == \"h2h\")\n",
    "    best = (\n",
    "        h2h.sort(\"outcome_price\", descending=True)\n",
    "        .group_by([\"home_team\", \"away_team\", \"outcome_name\"], maintain_order=True)\n",
    "        .agg(pl.first(\"outcome_price\").alias(\"best_price\"), pl.first(\"bookmaker_key\").alias(\"best_book\"))\n",
    "    )\n",
    "    out = best.head(10)\n",
    "else:\n",
    "    out = \"needs ODDS_API_KEY\"\n",
    "out"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "48e4db65",
   "metadata": {},
   "source": [
    "### Recipe 2 — Spreads & totals for a slate 📋\n",
    "\n",
    "Ask for `markets=\"spreads,totals\"` and the `outcome_point` column carries the\n",
    "line (the spread number / the over-under total)."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "dbc9a5cb",
   "metadata": {},
   "outputs": [],
   "source": [
    "if HAS_KEY:\n",
    "    st = odds.toa_sports_odds(sport=\"americanfootball_nfl\", regions=\"us\", markets=\"spreads,totals\")\n",
    "    out = (\n",
    "        st.filter(pl.col(\"bookmaker_key\") == st[\"bookmaker_key\"][0])\n",
    "        .select([\"home_team\", \"away_team\", \"market_key\", \"outcome_name\", \"outcome_point\", \"outcome_price\"])\n",
    "        .head(10)\n",
    "        if st.height else \"no spreads/totals on the board right now\"\n",
    "    )\n",
    "else:\n",
    "    out = \"needs ODDS_API_KEY\"\n",
    "out"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "53319952",
   "metadata": {},
   "source": [
    "### Recipe 3 — Just one book 🎯\n",
    "\n",
    "Pin a single sportsbook with `bookmakers=`. Great for tracking *your* book's\n",
    "line without paying for a whole region."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "6cc8a030",
   "metadata": {},
   "outputs": [],
   "source": [
    "if HAS_KEY:\n",
    "    dk = odds.toa_sports_odds(sport=\"americanfootball_nfl\", bookmakers=\"draftkings\", markets=\"h2h\")\n",
    "    out = dk.select([\"home_team\", \"away_team\", \"outcome_name\", \"outcome_price\"]).head() if dk.height else \"no lines yet\"\n",
    "else:\n",
    "    out = \"needs ODDS_API_KEY\"\n",
    "out"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "01b31c29",
   "metadata": {},
   "source": [
    "### Recipe 4 — Implied probability & the hold 🧮\n",
    "\n",
    "American moneyline prices convert to **implied win probability** with a tiny\n",
    "formula. Add up both sides and the excess over 100% is the book's *hold* (the\n",
    "vig). Pure polars math on the frame you already pulled — no extra API call."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "e3794088",
   "metadata": {},
   "outputs": [],
   "source": [
    "if HAS_KEY and board is not None:\n",
    "    h2h = board.filter(pl.col(\"market_key\") == \"h2h\")\n",
    "    devig = (\n",
    "        h2h.with_columns(\n",
    "            pl.when(pl.col(\"outcome_price\") < 0)\n",
    "            .then(-pl.col(\"outcome_price\") / (-pl.col(\"outcome_price\") + 100))\n",
    "            .otherwise(100 / (pl.col(\"outcome_price\") + 100))\n",
    "            .alias(\"implied_prob\")\n",
    "        )\n",
    "        .group_by([\"home_team\", \"away_team\", \"bookmaker_key\"], maintain_order=True)\n",
    "        .agg(pl.sum(\"implied_prob\").alias(\"market_total\"))\n",
    "        .with_columns(((pl.col(\"market_total\") - 1) * 100).round(2).alias(\"hold_pct\"))\n",
    "        .sort(\"hold_pct\")\n",
    "    )\n",
    "    out = devig.head(10)\n",
    "else:\n",
    "    out = \"needs ODDS_API_KEY\"\n",
    "out"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "931743d5",
   "metadata": {},
   "source": [
    "### Recipe 5 — Find the biggest favorite on the board 🐻\n",
    "\n",
    "Sort the moneyline outcomes by price ascending — the most negative number is\n",
    "the heaviest chalk on the slate. A classic \"find the X\" one-liner."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "e8821c26",
   "metadata": {},
   "outputs": [],
   "source": [
    "if HAS_KEY and board is not None:\n",
    "    faves = (\n",
    "        board.filter(pl.col(\"market_key\") == \"h2h\")\n",
    "        .sort(\"outcome_price\")\n",
    "        .select([\"home_team\", \"away_team\", \"outcome_name\", \"outcome_price\", \"bookmaker_key\"])\n",
    "        .head(5)\n",
    "    )\n",
    "    out = faves\n",
    "else:\n",
    "    out = \"needs ODDS_API_KEY\"\n",
    "out"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "700ec4e6",
   "metadata": {},
   "source": [
    "### Recipe 6 — Consensus over/under per game 📊\n",
    "\n",
    "Books disagree by a half-point here and there. Take the **median** total\n",
    "across every book to get a stable market consensus for each matchup."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "d601111e",
   "metadata": {},
   "outputs": [],
   "source": [
    "if HAS_KEY:\n",
    "    tot = odds.toa_sports_odds(sport=\"americanfootball_nfl\", regions=\"us\", markets=\"totals\")\n",
    "    if tot.height:\n",
    "        consensus = (\n",
    "            tot.filter(pl.col(\"outcome_name\") == \"Over\")\n",
    "            .group_by([\"home_team\", \"away_team\"], maintain_order=True)\n",
    "            .agg(\n",
    "                pl.median(\"outcome_point\").alias(\"consensus_total\"),\n",
    "                pl.col(\"bookmaker_key\").n_unique().alias(\"n_books\"),\n",
    "            )\n",
    "            .sort(\"consensus_total\", descending=True)\n",
    "        )\n",
    "        out = consensus.head(10)\n",
    "    else:\n",
    "        out = \"no totals on the board right now\"\n",
    "else:\n",
    "    out = \"needs ODDS_API_KEY\"\n",
    "out"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "f0eb8625",
   "metadata": {},
   "source": [
    "### Recipe 7 — Just today's slate ⏰\n",
    "\n",
    "Narrow the pull to a time window with `commence_time_from` / `commence_time_to`\n",
    "(ISO-8601, UTC). Here: only games kicking off in the next 24 hours."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "ed2e7d51",
   "metadata": {},
   "outputs": [],
   "source": [
    "from datetime import datetime, timedelta, timezone\n",
    "\n",
    "if HAS_KEY:\n",
    "    now = datetime.now(timezone.utc)\n",
    "    fmt = \"%Y-%m-%dT%H:%M:%SZ\"\n",
    "    today = odds.toa_sports_odds(\n",
    "        sport=\"americanfootball_nfl\",\n",
    "        regions=\"us\",\n",
    "        markets=\"h2h\",\n",
    "        commence_time_from=now.strftime(fmt),\n",
    "        commence_time_to=(now + timedelta(hours=24)).strftime(fmt),\n",
    "    )\n",
    "    out = (\n",
    "        today.select([\"commence_time\", \"home_team\", \"away_team\"]).unique(maintain_order=True).head(10)\n",
    "        if today.height else \"nothing kicks off in the next 24h\"\n",
    "    )\n",
    "else:\n",
    "    out = \"set ODDS_API_KEY to run the commence-time filter recipe\"\n",
    "out"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4390a342",
   "metadata": {},
   "source": [
    "### Recipe 8 — Player props for one game 🎯\n",
    "\n",
    "Event-level markets (player props!) live on [`toa_event_odds`](../odds/reference/additional.md#toa_event_odds).\n",
    "Grab an `event_id` from [`toa_sports_events`](../odds/reference/additional.md#toa_sports_events), then ask\n",
    "for a prop market like `player_pass_tds` or `player_anytime_td`."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "8d7ea073",
   "metadata": {},
   "outputs": [],
   "source": [
    "if HAS_KEY:\n",
    "    events = odds.toa_sports_events(sport=\"americanfootball_nfl\", return_parsed=False)\n",
    "    if events:\n",
    "        eid = events[0][\"id\"]\n",
    "        props = odds.toa_event_odds(sport=\"americanfootball_nfl\", event_id=eid, markets=\"player_pass_tds\")\n",
    "        out = props.select([c for c in [\"outcome_name\", \"outcome_description\", \"outcome_point\", \"outcome_price\"]\n",
    "                             if c in props.columns]).head()\n",
    "    else:\n",
    "        out = \"no upcoming NFL events right now\"\n",
    "else:\n",
    "    out = \"set ODDS_API_KEY to run the player-props recipe\"\n",
    "out"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "0c064903",
   "metadata": {},
   "source": [
    "### Recipe 9 — Which markets does a game offer? 🗃️\n",
    "\n",
    "Not sure which props are even available? [`toa_event_markets`](../odds/reference/additional.md#toa_event_markets)\n",
    "lists every market on offer per book — and it's **free**. Count them up to\n",
    "see which sportsbook posts the deepest menu."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "66fe8113",
   "metadata": {},
   "outputs": [],
   "source": [
    "if HAS_KEY:\n",
    "    events = odds.toa_sports_events(sport=\"americanfootball_nfl\", return_parsed=False)\n",
    "    if events:\n",
    "        eid = events[0][\"id\"]\n",
    "        mk = odds.toa_event_markets(sport=\"americanfootball_nfl\", event_id=eid)\n",
    "        out = (\n",
    "            mk.group_by(\"bookmaker_key\").agg(pl.col(\"market_key\").n_unique().alias(\"n_markets\"))\n",
    "            .sort(\"n_markets\", descending=True).head(10)\n",
    "            if mk.height else \"no markets posted for this event yet\"\n",
    "        )\n",
    "    else:\n",
    "        out = \"no upcoming NFL events right now\"\n",
    "else:\n",
    "    out = \"set ODDS_API_KEY to run the event-markets recipe\"\n",
    "out"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "6cace678",
   "metadata": {},
   "source": [
    "### Recipe 10 — Recent finals & margin of victory 🏁\n",
    "\n",
    "[`toa_sports_scores`](../odds/reference/additional.md#toa_sports_scores) returns live + recently\n",
    "completed games (`days_from=1..3`, **free**). Keep the completed ones and\n",
    "show the final scoreline — handy for grading bets after the fact."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "fd7350bb",
   "metadata": {},
   "outputs": [],
   "source": [
    "if HAS_KEY:\n",
    "    sc = odds.toa_sports_scores(sport=\"americanfootball_nfl\", days_from=3)\n",
    "    keep = [c for c in [\"completed\", \"home_team\", \"away_team\", \"scores\", \"last_update\"] if c in sc.columns]\n",
    "    if sc.height and \"completed\" in sc.columns:\n",
    "        out = sc.filter(pl.col(\"completed\")).select(keep).head(10)\n",
    "    else:\n",
    "        out = sc.select(keep).head(10) if sc.height else \"no recent scores right now\"\n",
    "else:\n",
    "    out = \"set ODDS_API_KEY to run: odds.toa_sports_scores(sport='americanfootball_nfl', days_from=3)\"\n",
    "out"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "fe6b1338",
   "metadata": {},
   "source": [
    "### Recipe 11 — Tour several leagues at once 🔁\n",
    "\n",
    "The `sport=` key is the only thing that changes between leagues, so one loop\n",
    "counts the upcoming events across a handful of them. `toa_sports_events` is\n",
    "**free**, so this sweep costs you nothing."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "1abfd12d",
   "metadata": {},
   "outputs": [],
   "source": [
    "if HAS_KEY:\n",
    "    keys = [\"americanfootball_nfl\", \"basketball_nba\", \"icehockey_nhl\", \"baseball_mlb\"]\n",
    "    rows = []\n",
    "    for k in keys:\n",
    "        evs = odds.toa_sports_events(sport=k, return_parsed=False)\n",
    "        rows.append({\"sport\": k, \"upcoming_events\": len(evs) if isinstance(evs, list) else 0})\n",
    "    out = pl.DataFrame(rows).sort(\"upcoming_events\", descending=True)\n",
    "else:\n",
    "    out = \"set ODDS_API_KEY to tour leagues with odds.toa_sports_events(sport=...)\"\n",
    "out"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "ad77d5b4",
   "metadata": {},
   "source": [
    "### Recipe 12 — Who's in the league? (participants 👥)\n",
    "\n",
    "[`toa_sports_participants`](../odds/reference/additional.md#toa_sports_participants) lists every team /\n",
    "participant for a sport — the lookup table you join odds against by name.\n",
    "Also **free**."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "30871f49",
   "metadata": {},
   "outputs": [],
   "source": [
    "if HAS_KEY:\n",
    "    parts = odds.toa_sports_participants(sport=\"americanfootball_nfl\")\n",
    "    keep = [c for c in [\"full_name\", \"id\", \"abbreviation\"] if c in parts.columns]\n",
    "    out = parts.select(keep if keep else parts.columns).head(10) if parts.height else \"no participants listed\"\n",
    "else:\n",
    "    out = \"set ODDS_API_KEY to run: odds.toa_sports_participants(sport='americanfootball_nfl')\"\n",
    "out"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "4074bd00",
   "metadata": {},
   "source": [
    "## ⛽ Mind your quota\n",
    "\n",
    "Paid calls cost credits (every 10 bookmakers × market ≈ 1 credit). After any\n",
    "call, [`toa_usage`](../odds/reference/additional.md#toa_usage) reads the most recent\n",
    "`x-requests-remaining` / `x-requests-used` headers **without spending a\n",
    "request** — handy to drop at the end of a script."
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "id": "36d377ce",
   "metadata": {},
   "outputs": [],
   "source": [
    "odds.toa_usage() if HAS_KEY else \"set ODDS_API_KEY to track quota with odds.toa_usage()\""
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9be208ba",
   "metadata": {},
   "source": [
    "## ⏳ Time travel: historical odds\n",
    "\n",
    "On a paid plan you can pull point-in-time snapshots — perfect for *closing\n",
    "line value* studies. Pass a `date=` ISO-8601 timestamp; the snapshot is\n",
    "unwrapped to the same long format and every row is stamped with the snapshot\n",
    "time.\n",
    "\n",
    "| Function | Snapshot of… |\n",
    "|---|---|\n",
    "| [`toa_sports_odds_history`](../odds/reference/additional.md#toa_sports_odds_history) | a whole sport's odds at `date` |\n",
    "| [`toa_sports_events_history`](../odds/reference/additional.md#toa_sports_events_history) | the events at `date` |\n",
    "| [`toa_event_odds_history`](../odds/reference/additional.md#toa_event_odds_history) | one game's odds at `date` |\n",
    "\n",
    "```python\n",
    "odds.toa_sports_odds_history(sport=\"americanfootball_nfl\", date=\"2023-11-29T22:45:00Z\")\n",
    "```"
   ]
  },
  {
   "cell_type": "markdown",
   "id": "9e6ac402",
   "metadata": {},
   "source": [
    "## 🎉 Where to next\n",
    "\n",
    "- Pass `return_as_pandas=True` for a pandas frame, or `return_parsed=False` for raw JSON.\n",
    "- Full reference: the **Betting → Odds** section in the sidebar.\n",
    "- R user? The same surface lives in [oddsapiR](https://oddsapir.sportsdataverse.org).\n",
    "\n",
    "Happy modelling — may your closing line value be ever positive! 📈"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "Python 3",
   "language": "python",
   "name": "python3"
  },
  "language_info": {
   "name": "python"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 5
}