{
 "cells": [
  {
   "cell_type": "code",
   "execution_count": 3,
   "metadata": {},
   "outputs": [],
   "source": [
    "import pandas as pd\n",
    "# import sys\n",
    "# sys.path.append('..')\n",
    "# from scripts.metrics import compute_weekly_metrics_by_market_creator"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 4,
   "metadata": {},
   "outputs": [],
   "source": [
    "all_trades = pd.read_parquet('../data/all_trades_profitability.parquet')"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 5,
   "metadata": {},
   "outputs": [
    {
     "data": {
      "text/plain": [
       "Index(['trader_address', 'market_creator', 'trade_id', 'creation_timestamp',\n",
       "       'title', 'market_status', 'collateral_amount', 'outcome_index',\n",
       "       'trade_fee_amount', 'outcomes_tokens_traded', 'current_answer',\n",
       "       'is_invalid', 'winning_trade', 'earnings', 'redeemed',\n",
       "       'redeemed_amount', 'num_mech_calls', 'mech_fee_amount', 'net_earnings',\n",
       "       'roi', 'staking', 'nr_mech_calls'],\n",
       "      dtype='object')"
      ]
     },
     "execution_count": 5,
     "metadata": {},
     "output_type": "execute_result"
    }
   ],
   "source": [
    "all_trades.columns"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 6,
   "metadata": {},
   "outputs": [],
   "source": [
    "def compute_winning_metric_per_trader_per_market_creator(\n",
    "    trader_address: str, week_traders_data: pd.DataFrame, market_creator: str = \"all\"\n",
    ") -> float:\n",
    "    assert \"market_creator\" in week_traders_data.columns\n",
    "    filtered_traders_data = week_traders_data.loc[\n",
    "        week_traders_data[\"trader_address\"] == trader_address\n",
    "    ]\n",
    "    if market_creator != \"all\":  # compute only for the specific market creator\n",
    "        filtered_traders_data = filtered_traders_data.loc[\n",
    "            filtered_traders_data[\"market_creator\"] == market_creator\n",
    "        ]\n",
    "        if len(filtered_traders_data) == 0:\n",
    "            return None  # No Data\n",
    "    winning_perc = (\n",
    "        filtered_traders_data[\"winning_trade\"].sum()\n",
    "        / filtered_traders_data[\"winning_trade\"].count()\n",
    "        * 100.0\n",
    "    )\n",
    "    return winning_perc"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 7,
   "metadata": {},
   "outputs": [],
   "source": [
    "def merge_winning_metrics_by_trader(\n",
    "    trader: str, weekly_data: pd.DataFrame, week: str\n",
    ") -> pd.DataFrame:\n",
    "    trader_metrics = []\n",
    "    # computation as specification 1 for all market creators\n",
    "    winning_metrics_all = {}\n",
    "    winning_metric_all = compute_winning_metric_per_trader_per_market_creator(\n",
    "        trader, weekly_data, market_creator=\"all\"\n",
    "    )\n",
    "    winning_metrics_all[\"winning_perc\"] = winning_metric_all\n",
    "    winning_metrics_all[\"month_year_week\"] = week\n",
    "    winning_metrics_all[\"market_creator\"] = \"all\"\n",
    "    trader_metrics.append(winning_metrics_all)\n",
    "    if week == \"Jul-21\":\n",
    "        print(f\"trader = {trader}, win_perc for all ={winning_metric_all}\")\n",
    "\n",
    "    # computation as specification 1 for quickstart markets\n",
    "    winning_metrics_qs = {}\n",
    "    winning_metric = compute_winning_metric_per_trader_per_market_creator(\n",
    "        trader, weekly_data, market_creator=\"quickstart\"\n",
    "    )\n",
    "    if winning_metric:\n",
    "        winning_metrics_qs[\"winning_perc\"] = winning_metric\n",
    "        winning_metrics_qs[\"month_year_week\"] = week\n",
    "        winning_metrics_qs[\"market_creator\"] = \"quickstart\"\n",
    "        trader_metrics.append(winning_metrics_qs)\n",
    "\n",
    "    # computation as specification 1 for pearl markets\n",
    "    winning_metrics_pearl = {}\n",
    "    winning_metric = compute_winning_metric_per_trader_per_market_creator(\n",
    "        trader, weekly_data, market_creator=\"pearl\"\n",
    "    )\n",
    "    if winning_metric:\n",
    "        winning_metrics_pearl[\"winning_perc\"] = winning_metric\n",
    "        winning_metrics_pearl[\"month_year_week\"] = week\n",
    "        winning_metrics_pearl[\"market_creator\"] = \"pearl\"\n",
    "        trader_metrics.append(winning_metrics_pearl)\n",
    "\n",
    "    result = pd.DataFrame.from_dict(trader_metrics, orient=\"columns\")\n",
    "    # tqdm.write(f\"Total length of all winning metrics for this week = {len(result)}\")\n",
    "    return result"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 27,
   "metadata": {},
   "outputs": [],
   "source": [
    "def win_metrics_trader_level(weekly_data):\n",
    "    winning_trades = (\n",
    "        weekly_data.groupby([\"month_year_week\", \"market_creator\",\"trader_address\"], sort=False)[\n",
    "            \"winning_trade\"\n",
    "        ].sum()\n",
    "        / weekly_data.groupby([\"month_year_week\", \"market_creator\",\"trader_address\"], sort=False)[\n",
    "            \"winning_trade\"\n",
    "        ].count()\n",
    "        * 100\n",
    "    )\n",
    "    # winning_trades is a series, give it a dataframe\n",
    "    winning_trades = winning_trades.reset_index()\n",
    "    winning_trades.columns = winning_trades.columns.astype(str)\n",
    "    winning_trades.columns = [\"month_year_week\", \"market_creator\", \"trader_address\", \"winning_trade\"]\n",
    "    winning_trades.rename(columns={\"winning_trade\": \"winning_perc\"})\n",
    "    return winning_trades"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "from tqdm import tqdm\n",
    "\n",
    "market_all = all_trades.copy(deep=True)\n",
    "market_all[\"market_creator\"] = \"all\"\n",
    "\n",
    "# merging both dataframes\n",
    "final_traders = pd.concat([market_all, all_trades], ignore_index=True)\n",
    "final_traders = final_traders.sort_values(\n",
    "        by=\"creation_timestamp\", ascending=True)\n",
    "\n",
    "\n",
    "winning_df = win_metrics_trader_level(final_traders)\n",
    "winning_df.head()"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "winning_df = compute_winning_metrics_by_trader(all_trades)"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": 29,
   "metadata": {},
   "outputs": [],
   "source": [
    "winning_pearl = winning_df.loc[winning_df[\"market_creator\"]==\"pearl\"]"
   ]
  },
  {
   "cell_type": "code",
   "execution_count": null,
   "metadata": {},
   "outputs": [],
   "source": [
    "winning_pearl.head()"
   ]
  }
 ],
 "metadata": {
  "kernelspec": {
   "display_name": "hf_dashboards",
   "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": "3.12.2"
  }
 },
 "nbformat": 4,
 "nbformat_minor": 2
}