Merge pull request #132 from amor71/anchor-vwap

Anchor vwap

.gitignore

@@ -7,14 +7,14 @@
 /legacy/
 /tools/set_keys.bash
 /tools/fix_git_email.bash
-/analyis/.ipynb_checkpoints/
+/analysis/.ipynb_checkpoints/
 /liualgotrader/__pycache__/
 /tests/__pycache__/
 /.idea/
 /liualgotrader/scanners/__pycache__/
 /liualgotrader/common/__pycache__/
 /docs/build/
-/analyis/notebooks/.ipynb_checkpoints
+/analysis/notebooks/.ipynb_checkpoints
 
 # Byte-compiled / optimized / DLL files
 __pycache__/

analyis/notebooks/backtest_performance_analysis.ipynb → analysis/notebooks/backtest_performance_analysis.ipynb

@@ -13,7 +13,7 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "batch_id = \"8048cc68-9e69-490c-8e3b-c977453749f8\""
+    "batch_id = \"5ebd4bf8-728f-405c-b8f5-d18a280b98d9\""
    ]
   },
   {

analysis/notebooks/distributions.ipynb

@@ -0,0 +1,326 @@
+{
+ "cells": [
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## Take Action: Select schedule to analyze, and environment"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "start_date = \"2020-11-2\"\n",
+    "end_date = None # = now()!\n",
+    "env = \"PAPER\"  # values may be PAPER / PROD / BACKTEST"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "#### Imports"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "import asyncio\n",
+    "import json\n",
+    "import math\n",
+    "import sys\n",
+    "from datetime import date, datetime, timedelta\n",
+    "\n",
+    "import alpaca_trade_api as tradeapi\n",
+    "import matplotlib.pyplot as plt\n",
+    "import nest_asyncio\n",
+    "import numpy as np\n",
+    "import pandas as pd\n",
+    "import pytz\n",
+    "import requests\n",
+    "from dateutil import parser\n",
+    "from IPython.display import HTML, display, Markdown\n",
+    "from liualgotrader.analytics import analysis\n",
+    "from pytz import timezone\n",
+    "\n",
+    "import sklearn\n",
+    "from sklearn.cluster import Birch\n",
+    "print(f\"SKLEARN Version {sklearn.__version__}\")\n",
+    "\n",
+    "%matplotlib inline\n",
+    "\n",
+    "\n",
+    "nest_asyncio.apply()"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "#### Load trading day details"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "day_to_analyze = datetime.strptime(start_date, \"%Y-%m-%d\")\n",
+    "end_to_analyze = datetime.strptime(edn_date, \"%Y-%m-%d\") if end_date else datetime.now()\n",
+    "trades = analysis.load_trades(day_to_analyze, env, end_to_analyze)\n",
+    "algo_runs = analysis.load_runs(day_to_analyze, env, end_to_analyze)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "## Show trading session performance"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "symbol_name = []\n",
+    "counts = []\n",
+    "revenues = []\n",
+    "est = pytz.timezone(\"US/Eastern\")\n",
+    "batch_ids = trades.batch_id.unique().tolist()\n",
+    "# batch_ids.reverse()\n",
+    "\n",
+    "current_max = pd.options.display.max_rows\n",
+    "pd.set_option(\"display.max_rows\", None)\n",
+    "for batch_id in batch_ids:\n",
+    "    how_was_my_day = pd.DataFrame()\n",
+    "    how_was_my_day[\"symbol\"] = trades.loc[trades[\"batch_id\"] == batch_id][\n",
+    "        \"symbol\"\n",
+    "    ].unique()\n",
+    "    how_was_my_day[\"revenues\"] = how_was_my_day[\"symbol\"].apply(\n",
+    "        lambda x: analysis.calc_batch_revenue(x, trades, batch_id)\n",
+    "    )\n",
+    "    how_was_my_day[\"count\"] = how_was_my_day[\"symbol\"].apply(\n",
+    "        lambda x: analysis.count_trades(x, trades, batch_id)\n",
+    "    )\n",
+    "\n",
+    "    if len(algo_runs.loc[algo_runs[\"batch_id\"] == batch_id].start_time) > 0:\n",
+    "        batch_time = (\n",
+    "            algo_runs.loc[algo_runs[\"batch_id\"] == batch_id]\n",
+    "            .start_time.min()\n",
+    "            .tz_localize(\"utc\")\n",
+    "            .astimezone(est)\n",
+    "        )\n",
+    "    else:\n",
+    "        continue\n",
+    "\n",
+    "    env = algo_runs[algo_runs[\"batch_id\"] == batch_id].algo_env.tolist()[0]\n",
+    "    win_ratio = round(\n",
+    "        1.0\n",
+    "        * len(how_was_my_day[how_was_my_day.revenues >= 0])\n",
+    "        / len(how_was_my_day[how_was_my_day.revenues < 0]),\n",
+    "        2,\n",
+    "    )\n",
+    "    revenues = round(sum(how_was_my_day[\"revenues\"]), 2)\n",
+    "    traded = len(how_was_my_day)\n",
+    "    print(\n",
+    "        f\"[{env}] {batch_id}\\n{batch_time}\\nTotal revenues=${revenues}\\nTotal traded:{traded} Win/Lose ration {win_ratio}\"\n",
+    "    )\n",
+    "\n",
+    "    display(\n",
+    "        Markdown(f\"{len(how_was_my_day[how_was_my_day.revenues >= 0])} **Winners**\")\n",
+    "    )\n",
+    "    display(how_was_my_day[how_was_my_day.revenues >= 0].sort_values('revenues'))\n",
+    "    display(Markdown(f\"{len(how_was_my_day[how_was_my_day.revenues < 0])} **Lossers**\"))\n",
+    "    display(how_was_my_day[how_was_my_day.revenues < 0].sort_values('revenues'))\n",
+    "\n",
+    "pd.set_option(\"display.max_rows\", current_max)"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "##  Time analysis"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "time_model = pd.DataFrame()\n",
+    "\n",
+    "time_model[\"symbol\"] = trades.symbol.unique()\n",
+    "time_model[\"time\"] = time_model[\"symbol\"].apply(\n",
+    "    lambda x: (\n",
+    "        pytz.utc.localize(\n",
+    "            pd.to_datetime(trades.loc[trades.symbol == x].tstamp.min())\n",
+    "        ).astimezone(est)\n",
+    "        - day_to_analyze.replace(hour=9, minute=30, tzinfo=est)\n",
+    "    ).total_seconds()\n",
+    ")\n",
+    "time_model[\"revenue\"] = time_model[\"symbol\"].apply(\n",
+    "    lambda x: sum(\n",
+    "        [analysis.calc_batch_revenue(x, trades, batch_id) for batch_id in batch_ids]\n",
+    "    )\n",
+    ")\n",
+    "time_model[\"algo\"] = time_model[\"symbol\"].apply(\n",
+    "    lambda x: trades.loc[trades.symbol == x].algo_name.unique().tolist()\n",
+    ")\n",
+    "time_model = time_model.sort_values(\"time\")\n",
+    "time_model[\"total\"] = time_model.revenue.cumsum()\n",
+    "time_model[\"reasons\"] = time_model[\"symbol\"].apply(\n",
+    "    lambda x: [\n",
+    "        json.loads(i)[\"sell\"]['reasons']\n",
+    "        for i in trades.loc[trades.symbol == x].indicators.tolist()\n",
+    "        if json.loads(i)[\"sell\"] and 'reasons' in json.loads(i)[\"sell\"]\n",
+    "    ]\n",
+    ")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "plt.plot(time_model.time.tolist(), time_model.total.tolist())\n",
+    "plt.title(\"Accumulative Revenues\")\n",
+    "plt.show()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "plt.scatter(\n",
+    "    time_model.time,\n",
+    "    time_model.revenue,\n",
+    ")\n",
+    "plt.title(\"Trade profit by time\")\n",
+    "plt.show()"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "c1 = []\n",
+    "c2 = []\n",
+    "c3 = []\n",
+    "c4 = []\n",
+    "c5 = []\n",
+    "symbols = trades[\"symbol\"].unique()\n",
+    "for symbol in symbols:\n",
+    "    for algo in trades.loc[trades.symbol == symbol].algo_name.unique():\n",
+    "        position = 0\n",
+    "        for i, row in (\n",
+    "            trades.loc[(trades.symbol == symbol) & (trades.algo_name == algo)]\n",
+    "            .sort_values(\"tstamp\")\n",
+    "            .iterrows()\n",
+    "        ):\n",
+    "            if not position:\n",
+    "                c1.append(symbol)\n",
+    "                c2.append(algo)\n",
+    "                c3.append(row[\"tstamp\"])\n",
+    "                c4.append(0.0)\n",
+    "                c5.append(0.0)\n",
+    "                delta = 0.0\n",
+    "            position += row[\"qty\"] * (\n",
+    "                (1 if row[\"operation\"] == \"sell\" and row[\"qty\"] > 0 else -1)\n",
+    "            )\n",
+    "            delta += (\n",
+    "                float(row[\"price\"])\n",
+    "                * row[\"qty\"]\n",
+    "                * (1 if row[\"operation\"] == \"sell\" and row[\"qty\"] > 0 else -1)\n",
+    "            )\n",
+    "            c4[-1] = delta\n",
+    "            c5[-1] = round((row[\"tstamp\"] - c3[-1]).total_seconds(), 3)\n",
+    "\n",
+    "trades_data = pd.DataFrame(\n",
+    "    data={\"symbol\": c1, \"algo\": c2, \"when\": c3, \"revenue\": c4, \"duration\": c5}\n",
+    ").sort_values(\"when\")"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "for algo in trades_data.algo.unique():\n",
+    "    plt.plot(\n",
+    "        trades_data.loc[trades_data.algo == algo].when.tolist(),\n",
+    "        trades_data.loc[trades_data.algo == algo].revenue.cumsum().tolist(),\n",
+    "        label=algo,\n",
+    "    )\n",
+    "plt.title(\"Accumulative Revenue by Algo\")\n",
+    "plt.legend()\n",
+    "plt.show()\n",
+    "\n",
+    "for algo in trades_data.algo.unique():\n",
+    "    plt.scatter(\n",
+    "        trades_data.loc[trades_data.algo == algo].when.tolist(),\n",
+    "        trades_data.loc[trades_data.algo == algo].revenue.tolist(),\n",
+    "        label=algo,\n",
+    "    )\n",
+    "plt.title(\"Trades Profit by Algo over time\")\n",
+    "plt.legend()\n",
+    "plt.show()"
+   ]
+  },
+  {
+   "cell_type": "markdown",
+   "metadata": {},
+   "source": [
+    "### Revenue Distribution"
+   ]
+  },
+  {
+   "cell_type": "code",
+   "execution_count": null,
+   "metadata": {},
+   "outputs": [],
+   "source": [
+    "for algo in trades_data.algo.unique():\n",
+    "    trades_data.loc[trades_data.algo == algo].revenue.hist(bins=10, alpha=0.5, label=algo)\n",
+    "plt.legend()\n"
+   ]
+  }
+ ],
+ "metadata": {
+  "kernelspec": {
+   "display_name": "Python 3",
+   "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": "3.8.0"
+  }
+ },
+ "nbformat": 4,
+ "nbformat_minor": 2
+}

analyis/notebooks/portfolio_performance_analysis.ipynb → analysis/notebooks/portfolio_performance_analysis.ipynb

@@ -13,7 +13,7 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "day_to_analyze = \"2020-11-18\"\n",
+    "day_to_analyze = \"2020-11-12\"\n",
     "env = \"PAPER\"  # values may be PAPER / PROD / BACKTEST"
    ]
   },
@@ -48,8 +48,8 @@
     "from liualgotrader.analytics import analysis\n",
     "from pytz import timezone\n",
     "\n",
-    "%matplotlib inline\n",
     "\n",
+    "%matplotlib inline\n",
     "\n",
     "nest_asyncio.apply()"
    ]