cleaning and correction on winning perc

data/fpmms.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:86135bb64013c54d5180c31fca13235943eb39571e760a695dac2aaa1e9cb1ce
-size 436427

data/markets_live_data.parquet

@@ -1,3 +0,0 @@
-version https://git-lfs.github.com/spec/v1
-oid sha256:69a3fffac1b1e11e818cdf3c709fd3006d6f93107df947693548a05bc66f337d
-size 145777

notebooks/winning_perc.ipynb

@@ -0,0 +1,400 @@
+{
+ "cells": [
+  {
+   "cell_type": "code",
+   "execution_count": 1,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import pandas as pd"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 2,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "all_trades = pd.read_parquet('../data/all_trades_profitability.parquet')"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 4,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "all_trades[\"creation_date\"] = all_trades[\"creation_timestamp\"].dt.date"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": 5,
+   "metadata": {},
+   "outputs": [
+    {
+     "name": "stderr",
+     "output_type": "stream",
+     "text": [
+      "/var/folders/gp/02mb1d514ng739czlxw1lhh00000gn/T/ipykernel_38171/1825242321.py:6: UserWarning: Converting to PeriodArray/Index representation will drop timezone information.\n",
+      "  all_trades[\"creation_timestamp\"].dt.to_period(\"W\").dt.strftime(\"%b-%d\")\n"
+     ]
+    }
+   ],
+   "source": [
+    "all_trades = all_trades.sort_values(\n",
+    "    by=\"creation_timestamp\", ascending=True\n",
+    ")\n",
+    "\n",
+    "all_trades[\"month_year_week\"] = (\n",
+    "    all_trades[\"creation_timestamp\"].dt.to_period(\"W\").dt.strftime(\"%b-%d\")\n",
+    ")"
+   ]
+  },
+  {
+   "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": 28,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th>month_year_week</th>\n",
+       "      <th>market_creator</th>\n",
+       "      <th>trader_address</th>\n",
+       "      <th>winning_trade</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>0</th>\n",
+       "      <td>Jul-21</td>\n",
+       "      <td>all</td>\n",
+       "      <td>0x95ecc70d9f4feb162ed9f41c4432d990c36c8f57</td>\n",
+       "      <td>33.333333</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>1</th>\n",
+       "      <td>Jul-21</td>\n",
+       "      <td>quickstart</td>\n",
+       "      <td>0x95ecc70d9f4feb162ed9f41c4432d990c36c8f57</td>\n",
+       "      <td>33.333333</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>2</th>\n",
+       "      <td>Jul-21</td>\n",
+       "      <td>quickstart</td>\n",
+       "      <td>0xf089874165be0377680683fd5187a058dea82683</td>\n",
+       "      <td>100.000000</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>3</th>\n",
+       "      <td>Jul-21</td>\n",
+       "      <td>all</td>\n",
+       "      <td>0xf089874165be0377680683fd5187a058dea82683</td>\n",
+       "      <td>100.000000</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>4</th>\n",
+       "      <td>Jul-21</td>\n",
+       "      <td>quickstart</td>\n",
+       "      <td>0x49f4e3d8edc85efda9b0a36d96e406a59b13fcc2</td>\n",
+       "      <td>50.000000</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "  month_year_week market_creator                              trader_address  \\\n",
+       "0          Jul-21            all  0x95ecc70d9f4feb162ed9f41c4432d990c36c8f57   \n",
+       "1          Jul-21     quickstart  0x95ecc70d9f4feb162ed9f41c4432d990c36c8f57   \n",
+       "2          Jul-21     quickstart  0xf089874165be0377680683fd5187a058dea82683   \n",
+       "3          Jul-21            all  0xf089874165be0377680683fd5187a058dea82683   \n",
+       "4          Jul-21     quickstart  0x49f4e3d8edc85efda9b0a36d96e406a59b13fcc2   \n",
+       "\n",
+       "   winning_trade  \n",
+       "0      33.333333  \n",
+       "1      33.333333  \n",
+       "2     100.000000  \n",
+       "3     100.000000  \n",
+       "4      50.000000  "
+      ]
+     },
+     "execution_count": 28,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "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": 30,
+   "metadata": {},
+   "outputs": [
+    {
+     "data": {
+      "text/html": [
+       "<div>\n",
+       "<style scoped>\n",
+       "    .dataframe tbody tr th:only-of-type {\n",
+       "        vertical-align: middle;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe tbody tr th {\n",
+       "        vertical-align: top;\n",
+       "    }\n",
+       "\n",
+       "    .dataframe thead th {\n",
+       "        text-align: right;\n",
+       "    }\n",
+       "</style>\n",
+       "<table border=\"1\" class=\"dataframe\">\n",
+       "  <thead>\n",
+       "    <tr style=\"text-align: right;\">\n",
+       "      <th></th>\n",
+       "      <th>month_year_week</th>\n",
+       "      <th>market_creator</th>\n",
+       "      <th>trader_address</th>\n",
+       "      <th>winning_trade</th>\n",
+       "    </tr>\n",
+       "  </thead>\n",
+       "  <tbody>\n",
+       "    <tr>\n",
+       "      <th>7</th>\n",
+       "      <td>Jul-21</td>\n",
+       "      <td>pearl</td>\n",
+       "      <td>0xe283e408c6017447da9fe092d52c386753699680</td>\n",
+       "      <td>0.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>29</th>\n",
+       "      <td>Jul-21</td>\n",
+       "      <td>pearl</td>\n",
+       "      <td>0x913dedfcfb335a49509b67acb3b1ab2612a5c0c9</td>\n",
+       "      <td>100.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>30</th>\n",
+       "      <td>Jul-21</td>\n",
+       "      <td>pearl</td>\n",
+       "      <td>0x1b9e28e7f817e1312636a485f31cca8a4be61fac</td>\n",
+       "      <td>0.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>33</th>\n",
+       "      <td>Jul-21</td>\n",
+       "      <td>pearl</td>\n",
+       "      <td>0xe0113a139f591efa8bf5e19308c7c27199682d77</td>\n",
+       "      <td>0.0</td>\n",
+       "    </tr>\n",
+       "    <tr>\n",
+       "      <th>37</th>\n",
+       "      <td>Jul-21</td>\n",
+       "      <td>pearl</td>\n",
+       "      <td>0x66a022b113b41e08d90cfd9468b8b6565d6ea995</td>\n",
+       "      <td>100.0</td>\n",
+       "    </tr>\n",
+       "  </tbody>\n",
+       "</table>\n",
+       "</div>"
+      ],
+      "text/plain": [
+       "   month_year_week market_creator                              trader_address  \\\n",
+       "7           Jul-21          pearl  0xe283e408c6017447da9fe092d52c386753699680   \n",
+       "29          Jul-21          pearl  0x913dedfcfb335a49509b67acb3b1ab2612a5c0c9   \n",
+       "30          Jul-21          pearl  0x1b9e28e7f817e1312636a485f31cca8a4be61fac   \n",
+       "33          Jul-21          pearl  0xe0113a139f591efa8bf5e19308c7c27199682d77   \n",
+       "37          Jul-21          pearl  0x66a022b113b41e08d90cfd9468b8b6565d6ea995   \n",
+       "\n",
+       "    winning_trade  \n",
+       "7             0.0  \n",
+       "29          100.0  \n",
+       "30            0.0  \n",
+       "33            0.0  \n",
+       "37          100.0  "
+      ]
+     },
+     "execution_count": 30,
+     "metadata": {},
+     "output_type": "execute_result"
+    }
+   ],
+   "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
+}

scripts/metrics.py

@@ -76,28 +76,6 @@ def compute_trader_metrics_by_market_creator(
     return metrics
 
 
-def compute_winning_metric_per_trader_per_market_creator(
-    trader_address: str, week_traders_data: pd.DataFrame, market_creator: str = "all"
-) -> float:
-    assert "market_creator" in week_traders_data.columns
-    filtered_traders_data = week_traders_data.loc[
-        week_traders_data["trader_address"] == trader_address
-    ]
-    if market_creator != "all":  # compute only for the specific market creator
-        filtered_traders_data = filtered_traders_data.loc[
-            filtered_traders_data["market_creator"] == market_creator
-        ]
-        if len(filtered_traders_data) == 0:
-            tqdm.write(f"No data. Skipping market creator {market_creator}")
-            return None  # No Data
-    winning_perc = (
-        filtered_traders_data["winning_trade"].sum()
-        / filtered_traders_data["winning_trade"].count()
-        * 100.0
-    )
-    return winning_perc
-
-
 def merge_trader_metrics(
     trader: str, weekly_data: pd.DataFrame, week: str
 ) -> pd.DataFrame:
@@ -165,45 +143,21 @@ def merge_trader_metrics_by_type(
     return result
 
 
-def merge_winning_metrics_by_trader(
-    trader: str, weekly_data: pd.DataFrame, week: str
-) -> pd.DataFrame:
-    trader_metrics = []
-    # computation as specification 1 for all market creators
-    winning_metrics_all = {}
-    winning_metric_all = compute_winning_metric_per_trader_per_market_creator(
-        trader, weekly_data, market_creator="all"
-    )
-    winning_metrics_all["winning_perc"] = winning_metric_all
-    winning_metrics_all["month_year_week"] = week
-    winning_metrics_all["market_creator"] = "all"
-    trader_metrics.append(winning_metrics_all)
-
-    # computation as specification 1 for quickstart markets
-    winning_metrics_qs = {}
-    winning_metric = compute_winning_metric_per_trader_per_market_creator(
-        trader, weekly_data, market_creator="quickstart"
-    )
-    if winning_metric:
-        winning_metrics_qs["winning_perc"] = winning_metric
-        winning_metrics_qs["month_year_week"] = week
-        winning_metrics_qs["market_creator"] = "quickstart"
-        trader_metrics.append(winning_metrics_qs)
-
-    # computation as specification 1 for pearl markets
-    winning_metrics_pearl = {}
-    winning_metric = compute_winning_metric_per_trader_per_market_creator(
-        trader, weekly_data, market_creator="pearl"
+def win_metrics_trader_level(weekly_data):
+    winning_trades = (
+        weekly_data.groupby(
+            ["month_year_week", "market_creator", "trader_address"], sort=False
+        )["winning_trade"].sum()
+        / weekly_data.groupby(
+            ["month_year_week", "market_creator", "trader_address"], sort=False
+        )["winning_trade"].count()
+        * 100
     )
-    if winning_metric:
-        winning_metrics_pearl["winning_perc"] = winning_metric
-        winning_metrics_pearl["month_year_week"] = week
-        winning_metrics_pearl["market_creator"] = "pearl"
-        trader_metrics.append(winning_metrics_pearl)
-
-    result = pd.DataFrame.from_dict(trader_metrics, orient="columns")
-    # tqdm.write(f"Total length of all winning metrics for this week = {len(result)}")
-    return result
+    # winning_trades is a series, give it a dataframe
+    winning_trades = winning_trades.reset_index()
+    winning_trades.columns = winning_trades.columns.astype(str)
+    winning_trades.rename(columns={"winning_trade": "winning_perc"}, inplace=True)
+    return winning_trades
 
 
 def compute_weekly_metrics_by_market_creator(
@@ -248,16 +202,13 @@ def compute_winning_metrics_by_trader(
     trader_agents_data: pd.DataFrame,
 ) -> pd.DataFrame:
     """Function to compute the winning metrics at the trader level per week and with different market creators"""
-    contents = []
-    all_weeks = list(trader_agents_data.month_year_week.unique())
-    for week in all_weeks:
-        weekly_data = trader_agents_data.loc[
-            trader_agents_data["month_year_week"] == week
-        ]
-        print(f"Computing weekly metrics for week ={week} by trader type")
-        # traverse each trader agent
-        traders = list(weekly_data.trader_address.unique())
-        for trader in tqdm(traders, desc=f"Trader' metrics", unit="metrics"):
-            contents.append(merge_winning_metrics_by_trader(trader, weekly_data, week))
-    print("End computing all weekly winning metrics by trader")
-    return pd.concat(contents, ignore_index=True)
+    market_all = trader_agents_data.copy(deep=True)
+    market_all["market_creator"] = "all"
+
+    # merging both dataframes
+    final_traders = pd.concat([market_all, trader_agents_data], ignore_index=True)
+    final_traders = final_traders.sort_values(by="creation_timestamp", ascending=True)
+
+    winning_df = win_metrics_trader_level(final_traders)
+    winning_df.head()
+    return winning_df