paul-corbalan
/
python-pour-finance
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity
python-pour-finance/11-Quantopian-Avancé/00-Exemple-Pipeline.ipynb

720 lines
24 KiB
Plaintext
Raw Permalink Normal View History

Import
2023-08-21 15:12:19 +00:00
{
"cells": [
{
"cell_type": "markdown",
"metadata": {},
"source": [
"# Exemple Pipeline"
]
},
{
"cell_type": "code",
"execution_count": 1,
"metadata": {},
"outputs": [],
"source": [
"from quantopian.pipeline import Pipeline\n",
"from quantopian.research import run_pipeline\n",
"from quantopian.pipeline.data.builtin import USEquityPricing"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Obtenir les titres que nous voulons.\n",
"\n",
"### Les Q500US et Q1500US\n",
"\n",
"Ces groupes d'actions négociables sont appelés \"univers\", car toutes vos transactions utiliseront ces actions comme leur \"univers\" d'actions disponibles, ils ne négocieront rien en dehors de ces groupes."
]
},
{
"cell_type": "code",
"execution_count": 2,
"metadata": {},
"outputs": [],
"source": [
"from quantopian.pipeline.filters import Q1500US"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Le Q500US et le Q1500US présentent deux avantages principaux. Premièrement, ils réduisent considérablement le risque qu'un ordre ne soit pas exécuté. Deuxièmement, ils permettent des comparaisons plus significatives entre les stratégies car ils seront désormais utilisés comme univers standard pour les algorithmes."
]
},
{
"cell_type": "code",
"execution_count": 3,
"metadata": {},
"outputs": [],
"source": [
"universe = Q1500US()"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Filtrer davantage l'univers avec les classificateurs\n",
"\n",
"Ne prenons que des actions dans le secteur de l'énergie : https://www.quantopian.com/docs/data-reference/morningstar_fundamentals#industry-sector"
]
},
{
"cell_type": "code",
"execution_count": 4,
"metadata": {},
"outputs": [],
"source": [
"from quantopian.pipeline.data import morningstar"
]
},
{
"cell_type": "code",
"execution_count": 5,
"metadata": {},
"outputs": [],
"source": [
"sector = morningstar.asset_classification.morningstar_sector_code.latest"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Alternative:"
]
},
{
"cell_type": "code",
"execution_count": 6,
"metadata": {},
"outputs": [],
"source": [
"# from quantopian.pipeline.classifiers.morningstar import Sector\n",
"# morningstar_sector = Sector()"
]
},
{
"cell_type": "code",
"execution_count": 7,
"metadata": {},
"outputs": [],
"source": [
"energy_sector = sector.eq(309)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Masques de filtre\n",
"\n",
"Les masques peuvent également être appliqués aux méthodes qui renvoient des filtres comme top, bottom et le percentile_between.\n",
"\n",
"Les masques sont particulièrement utiles lorsque nous voulons appliquer un filtre dans les premières étapes d'un calcul combiné. Par exemple, supposons que nous voulions obtenir les 50 titres dont le prix à l'ouverture est le plus élevé et qui se situent également dans les 10 % supérieurs du volume en dollars. \n",
"\n",
"Supposons que nous voulions ensuite obtenir le 90ème-100ème percentile de ces titres par le prix de clôture. Nous pouvons le faire avec les éléments suivants :"
]
},
{
"cell_type": "code",
"execution_count": 8,
"metadata": {},
"outputs": [],
"source": [
"from quantopian.pipeline.factors import SimpleMovingAverage, AverageDollarVolume"
]
},
{
"cell_type": "code",
"execution_count": 9,
"metadata": {},
"outputs": [],
"source": [
"# Facteur de volume en dollar - Dollar volume factor\n",
"dollar_volume = AverageDollarVolume(window_length=30)\n",
"\n",
"# Filtre sur les volumes importants en dollars - High dollar volume filter\n",
"high_dollar_volume = dollar_volume.percentile_between(90,100)\n",
"\n",
"# Filtre de prix supérieur à l'ouverture (titres à fort volume en dollars)\n",
"# Top open price filter (high dollar volume securities)\n",
"top_open_price = USEquityPricing.open.latest.top(50, mask=high_dollar_volume)\n",
"\n",
"# Filtre de prix de clôture du premier percentile (volume de dollars élevé, prix d'ouverture des 50 premiers)\n",
"# Top percentile close price filter (high dollar volume, top 50 open price)\n",
"high_close_price = USEquityPricing.close.latest.percentile_between(90, 100, mask=top_open_price)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Application des Filtres et des Facteurs\n",
"\n",
"Appliquons nos propres filtres, en suivant certains des exemples ci-dessus. Choisissons les titres suivants :\n",
"\n",
"* Titres de Q1500US\n",
"* Les actions qui sont dans le secteur de l'énergie\n",
"* Il doit s'agir d'actions fortement négociées sur le marché (en fonction du volume de dollars échangés, elles doivent faire partie des 5% les plus échangés)\n",
"\n",
"Ensuite, nous calculerons la différence en pourcentage comme nous l'avons fait précédemment. En utilisant ce pourcentage de différence, nous créerons une stratégie peu sophistiquée qui court-circuite tout ce qui présente un pourcentage de différence négatif (la différence entre la moyenne sur 10 jours et la moyenne sur 30 jours)."
]
},
{
"cell_type": "code",
"execution_count": 10,
"metadata": {},
"outputs": [],
"source": [
"def make_pipeline():\n",
" \n",
" # Univers Q1500US\n",
" base_universe = Q1500US()\n",
" \n",
" # Secteur Energie\n",
" energy_sector = sector.eq(309)\n",
" \n",
" # Masque de l'univers 1500US et du secteur Energie\n",
" base_energy = base_universe & energy_sector\n",
" \n",
" # Volume en dollars (30 jours)\n",
" dollar_volume = AverageDollarVolume(window_length=30)\n",
"\n",
" # 5% supérieurs en volume en dollars moyens\n",
" high_dollar_volume = dollar_volume.percentile_between(95,100)\n",
" \n",
" # Combiner les filtres\n",
" top_half_base_energy = base_energy & high_dollar_volume\n",
" \n",
" # 10 day mean close\n",
" mean_10 = SimpleMovingAverage(inputs=[USEquityPricing.close], window_length=10, mask=top_half_base_energy)\n",
"\n",
" # 30 day mean close\n",
" mean_30 = SimpleMovingAverage(inputs=[USEquityPricing.close], window_length=30, mask=top_half_base_energy)\n",
"\n",
" # Différence en pourcentage\n",
" percent_difference = (mean_10 - mean_30) / mean_30\n",
" \n",
" # Liste des Shorts (position courte ou vente à découvert)\n",
" shorts = percent_difference < 0\n",
" \n",
" # Liste des Longs (position longue)\n",
" longs = percent_difference > 0\n",
" \n",
" # Masque final/filtre pour tout ce qui est en short ou en long\n",
" securities_to_trade = (shorts | longs)\n",
" \n",
" return Pipeline(\n",
" columns={\n",
" 'longs': longs,\n",
" 'shorts': shorts,\n",
" 'percent_diff':percent_difference\n",
" },\n",
" screen=securities_to_trade\n",
" )"
]
},
{
"cell_type": "code",
"execution_count": 11,
"metadata": {},
"outputs": [
{
"data": {
"application/vnd.jupyter.widget-view+json": {
"model_id": "",
"version_major": 2,
"version_minor": 0
},
"text/plain": []
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"text/html": [
"<b>Pipeline Execution Time:</b> 4.65 Seconds"
],
"text/plain": [
"<IPython.core.display.HTML object>"
]
},
"metadata": {},
"output_type": "display_data"
},
{
"data": {
"text/html": [
"<div>\n",
"<table border=\"1\" class=\"dataframe\">\n",
" <thead>\n",
" <tr style=\"text-align: right;\">\n",
" <th></th>\n",
" <th></th>\n",
" <th>longs</th>\n",
" <th>percent_diff</th>\n",
" <th>shorts</th>\n",
" </tr>\n",
" </thead>\n",
" <tbody>\n",
" <tr>\n",
" <th rowspan=\"38\" valign=\"top\">2015-05-05 00:00:00+00:00</th>\n",
" <th>Equity(216 [HES])</th>\n",
" <td>True</td>\n",
" <td>0.036528</td>\n",
" <td>False</td>\n",
" </tr>\n",
" <tr>\n",
" <th>Equity(448 [APA])</th>\n",
" <td>True</td>\n",
" <td>0.035926</td>\n",
" <td>False</td>\n",
" </tr>\n",
" <tr>\n",
" <th>Equity(455 [APC])</th>\n",
" <td>True</td>\n",
" <td>0.049153</td>\n",
" <td>False</td>\n",
" </tr>\n",
" <tr>\n",
" <th>Equity(858 [BHI])</th>\n",
" <td>True</td>\n",
" <td>0.033807</td>\n",
" <td>False</td>\n",
" </tr>\n",
" <tr>\n",
" <th>Equity(1746 [COG])</th>\n",
" <td>True</td>\n",
" <td>0.058078</td>\n",
" <td>False</td>\n",
" </tr>\n",
" <tr>\n",
" <th>Equity(2368 [DVN])</th>\n",
" <td>True</td>\n",
" <td>0.046264</td>\n",
" <td>False</td>\n",
" </tr>\n",
" <tr>\n",
" <th>Equity(2564 [EOG])</th>\n",
" <td>True</td>\n",
" <td>0.032102</td>\n",
" <td>False</td>\n",
" </tr>\n",
" <tr>\n",
" <th>Equity(2621 [VAL])</th>\n",
" <td>True</td>\n",
" <td>0.060197</td>\n",
" <td>False</td>\n",
" </tr>\n",
" <tr>\n",
" <th>Equity(3443 [HAL])</th>\n",
" <td>True</td>\n",
" <td>0.049257</td>\n",
" <td>False</td>\n",
" </tr>\n",
" <tr>\n",
" <th>Equity(3647 [HP])</th>\n",
" <td>True</td>\n",
" <td>0.040991</td>\n",
" <td>False</td>\n",
" </tr>\n",
" <tr>\n",
" <th>Equity(5035 [MRO])</th>\n",
" <td>True</td>\n",
" <td>0.061598</td>\n",
" <td>False</td>\n",
" </tr>\n",
" <tr>\n",
" <th>Equity(5213 [NBL])</th>\n",
" <td>True</td>\n",
" <td>0.010443</td>\n",
" <td>False</td>\n",
" </tr>\n",
" <tr>\n",
" <th>Equity(5214 [NBR])</th>\n",
" <td>True</td>\n",
" <td>0.064133</td>\n",
" <td>False</td>\n",
" </tr>\n",
" <tr>\n",
" <th>Equity(5249 [NE])</th>\n",
" <td>True</td>\n",
" <td>0.037559</td>\n",
" <td>False</td>\n",
" </tr>\n",
" <tr>\n",
" <th>Equity(5729 [OXY])</th>\n",
" <td>True</td>\n",
" <td>0.029776</td>\n",
" <td>False</td>\n",
" </tr>\n",
" <tr>\n",
" <th>Equity(6928 [SLB])</th>\n",
" <td>True</td>\n",
" <td>0.046555</td>\n",
" <td>False</td>\n",
" </tr>\n",
" <tr>\n",
" <th>Equity(7244 [SWN])</th>\n",
" <td>True</td>\n",
" <td>0.070788</td>\n",
" <td>False</td>\n",
" </tr>\n",
" <tr>\n",
" <th>Equity(7612 [ANDV])</th>\n",
" <td>True</td>\n",
" <td>0.005997</td>\n",
" <td>False</td>\n",
" </tr>\n",
" <tr>\n",
" <th>Equity(7990 [VLO])</th>\n",
" <td>False</td>\n",
" <td>-0.017145</td>\n",
" <td>True</td>\n",
" </tr>\n",
" <tr>\n",
" <th>Equity(8214 [WMB])</th>\n",
" <td>True</td>\n",
" <td>0.018876</td>\n",
" <td>False</td>\n",
" </tr>\n",
" <tr>\n",
" <th>Equity(8347 [XOM])</th>\n",
" <td>True</td>\n",
" <td>0.017343</td>\n",
" <td>False</td>\n",
" </tr>\n",
" <tr>\n",
" <th>Equity(8461 [CHK])</th>\n",
" <td>True</td>\n",
" <td>0.014265</td>\n",
" <td>False</td>\n",
" </tr>\n",
" <tr>\n",
" <th>Equity(9038 [RIG])</th>\n",
" <td>True</td>\n",
" <td>0.048180</td>\n",
" <td>False</td>\n",
" </tr>\n",
" <tr>\n",
" <th>Equity(13176 [CAM])</th>\n",
" <td>True</td>\n",
" <td>0.082110</td>\n",
" <td>False</td>\n",
" </tr>\n",
" <tr>\n",
" <th>Equity(17436 [PXD])</th>\n",
" <td>True</td>\n",
" <td>0.010248</td>\n",
" <td>False</td>\n",
" </tr>\n",
" <tr>\n",
" <th>Equity(19249 [RRC])</th>\n",
" <td>True</td>\n",
" <td>0.087062</td>\n",
" <td>False</td>\n",
" </tr>\n",
" <tr>\n",
" <th>Equity(19336 [WFT])</th>\n",
" <td>True</td>\n",
" <td>0.049141</td>\n",
" <td>False</td>\n",
" </tr>\n",
" <tr>\n",
" <th>Equity(22784 [FTI])</th>\n",
" <td>True</td>\n",
" <td>0.054529</td>\n",
" <td>False</td>\n",
" </tr>\n",
" <tr>\n",
" <th>Equity(23112 [CVX])</th>\n",
" <td>True</td>\n",
" <td>0.018972</td>\n",
" <td>False</td>\n",
" </tr>\n",
" <tr>\n",
" <th>Equity(23998 [COP])</th>\n",
" <td>True</td>\n",
" <td>0.023902</td>\n",
" <td>False</td>\n",
" </tr>\n",
" <tr>\n",
" <th>Equity(24809 [NOV])</th>\n",
" <td>True</td>\n",
" <td>0.024940</td>\n",
" <td>False</td>\n",
" </tr>\n",
" <tr>\n",
" <th>Equity(25707 [WLL])</th>\n",
" <td>True</td>\n",
" <td>0.048205</td>\n",
" <td>False</td>\n",
" </tr>\n",
" <tr>\n",
" <th>Equity(33856 [CLR])</th>\n",
" <td>True</td>\n",
" <td>0.064304</td>\n",
" <td>False</td>\n",
" </tr>\n",
" <tr>\n",
" <th>Equity(34440 [CXO])</th>\n",
" <td>True</td>\n",
" <td>0.042184</td>\n",
" <td>False</td>\n",
" </tr>\n",
" <tr>\n",
" <th>Equity(39797 [OAS])</th>\n",
" <td>True</td>\n",
" <td>0.042388</td>\n",
" <td>False</td>\n",
" </tr>\n",
" <tr>\n",
" <th>Equity(40852 [KMI])</th>\n",
" <td>True</td>\n",
" <td>0.023016</td>\n",
" <td>False</td>\n",
" </tr>\n",
" <tr>\n",
" <th>Equity(41636 [MPC])</th>\n",
" <td>True</td>\n",
" <td>0.011952</td>\n",
" <td>False</td>\n",
" </tr>\n",
" <tr>\n",
" <th>Equity(42788 [PSX])</th>\n",
" <td>True</td>\n",
" <td>0.020911</td>\n",
" <td>False</td>\n",
" </tr>\n",
" </tbody>\n",
"</table>\n",
"</div>"
],
"text/plain": [
" longs percent_diff shorts\n",
"2015-05-05 00:00:00+00:00 Equity(216 [HES]) True 0.036528 False\n",
" Equity(448 [APA]) True 0.035926 False\n",
" Equity(455 [APC]) True 0.049153 False\n",
" Equity(858 [BHI]) True 0.033807 False\n",
" Equity(1746 [COG]) True 0.058078 False\n",
" Equity(2368 [DVN]) True 0.046264 False\n",
" Equity(2564 [EOG]) True 0.032102 False\n",
" Equity(2621 [VAL]) True 0.060197 False\n",
" Equity(3443 [HAL]) True 0.049257 False\n",
" Equity(3647 [HP]) True 0.040991 False\n",
" Equity(5035 [MRO]) True 0.061598 False\n",
" Equity(5213 [NBL]) True 0.010443 False\n",
" Equity(5214 [NBR]) True 0.064133 False\n",
" Equity(5249 [NE]) True 0.037559 False\n",
" Equity(5729 [OXY]) True 0.029776 False\n",
" Equity(6928 [SLB]) True 0.046555 False\n",
" Equity(7244 [SWN]) True 0.070788 False\n",
" Equity(7612 [ANDV]) True 0.005997 False\n",
" Equity(7990 [VLO]) False -0.017145 True\n",
" Equity(8214 [WMB]) True 0.018876 False\n",
" Equity(8347 [XOM]) True 0.017343 False\n",
" Equity(8461 [CHK]) True 0.014265 False\n",
" Equity(9038 [RIG]) True 0.048180 False\n",
" Equity(13176 [CAM]) True 0.082110 False\n",
" Equity(17436 [PXD]) True 0.010248 False\n",
" Equity(19249 [RRC]) True 0.087062 False\n",
" Equity(19336 [WFT]) True 0.049141 False\n",
" Equity(22784 [FTI]) True 0.054529 False\n",
" Equity(23112 [CVX]) True 0.018972 False\n",
" Equity(23998 [COP]) True 0.023902 False\n",
" Equity(24809 [NOV]) True 0.024940 False\n",
" Equity(25707 [WLL]) True 0.048205 False\n",
" Equity(33856 [CLR]) True 0.064304 False\n",
" Equity(34440 [CXO]) True 0.042184 False\n",
" Equity(39797 [OAS]) True 0.042388 False\n",
" Equity(40852 [KMI]) True 0.023016 False\n",
" Equity(41636 [MPC]) True 0.011952 False\n",
" Equity(42788 [PSX]) True 0.020911 False"
]
},
"execution_count": 11,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"result = run_pipeline(make_pipeline(), '2015-05-05', '2015-05-05')\n",
"result"
]
},
{
"cell_type": "code",
"execution_count": 12,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"<class 'pandas.core.frame.DataFrame'>\n",
"MultiIndex: 38 entries, (2015-05-05 00:00:00+00:00, Equity(216 [HES])) to (2015-05-05 00:00:00+00:00, Equity(42788 [PSX]))\n",
"Data columns (total 3 columns):\n",
"longs 38 non-null bool\n",
"percent_diff 38 non-null float64\n",
"shorts 38 non-null bool\n",
"dtypes: bool(2), float64(1)\n",
"memory usage: 684.0+ bytes\n"
]
}
],
"source": [
"result.info()"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"# Exécution de cette stratégie dans l'IDE Quantopian"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"from quantopian.algorithm import attach_pipeline,pipeline_output\n",
"from quantopian.pipeline import Pipeline\n",
"from quantopian.pipeline.data.builtin import USEquityPricing\n",
"from quantopian.pipeline.factors import AverageDollarVolume,SimpleMovingAverage\n",
"from quantopian.pipeline.filters.morningstar import Q1500US\n",
"from quantopian.pipeline.data import morningstar\n",
"\n",
"def initialize(context):\n",
" \n",
" schedule_function(my_rebalance,date_rules.week_start(),time_rules.market_open(hours=1))\n",
" \n",
" my_pipe = make_pipeline()\n",
" attach_pipeline(my_pipe,'my_pipeline')\n",
" \n",
"def my_rebalance(context,data):\n",
" for security in context.portfolio.positions:\n",
" if security not in context.longs and security not in context.shorts and data.can_trade(security):\n",
" order_target_percent(security,0)\n",
" \n",
" for security in context.longs:\n",
" if data.can_trade(security):\n",
" order_target_percent(security,context.long_weight)\n",
"\n",
" for security in context.shorts:\n",
" if data.can_trade(security):\n",
" order_target_percent(security,context.short_weight)\n",
"\n",
"\n",
"\n",
"\n",
"def my_compute_weights(context):\n",
" \n",
" if len(context.longs)==0:\n",
" long_weight = 0\n",
" else:\n",
" long_weight = 0.5 / len(context.longs)\n",
" \n",
" if len(context.shorts)==0:\n",
" short_weight = 0\n",
" else:\n",
" short_weight = -0.5 / len(context.shorts)\n",
" \n",
" return (long_weight,short_weight)\n",
"\n",
"\n",
"\n",
"\n",
"\n",
"\n",
"def before_trading_start(context,data):\n",
" context.output = pipeline_output('my_pipeline')\n",
" \n",
" # LONG\n",
" context.longs = context.output[context.output['longs']].index.tolist()\n",
" \n",
" # SHORT\n",
" context.shorts = context.output[context.output['shorts']].index.tolist()\n",
"\n",
"\n",
" context.long_weight,context.short_weight = my_compute_weights(context)\n",
"\n",
"\n",
"\n",
"def make_pipeline():\n",
" \n",
" # Univers Q1500US\n",
" base_universe = Q1500US()\n",
" \n",
" # Secteur Energie\n",
" sector = morningstar.asset_classification.morningstar_sector_code.latest\n",
" energy_sector = sector.eq(309)\n",
" \n",
" # Masque de l'univers 1500US et du secteur Energie\n",
" base_energy = base_universe & energy_sector\n",
" \n",
" # Volume en dollars (30 jours)\n",
" dollar_volume = AverageDollarVolume(window_length=30)\n",
" \n",
" # 5% supérieurs en volume en dollars moyens\n",
" high_dollar_volume = dollar_volume.percentile_between(95,100)\n",
" \n",
" # Combiner les filtres\n",
" top_five_base_energy = base_energy & high_dollar_volume\n",
" \n",
" # 10 day mean close\n",
" mean_10 = SimpleMovingAverage(inputs=[USEquityPricing.close],window_length=10,mask=top_five_base_energy)\n",
" \n",
" # 30 day mean close\n",
" mean_30 = SimpleMovingAverage(inputs=[USEquityPricing.close],window_length=30,mask=top_five_base_energy)\n",
" \n",
" # Différence en pourcentage\n",
" percent_difference = (mean_10-mean_30)/mean_30\n",
" \n",
" # Liste des Shorts (position courte ou vente à découvert)\n",
" shorts = percent_difference < 0\n",
" \n",
" # Liste des Longs (position longue)\n",
" longs = percent_difference > 0\n",
" \n",
" # Masque final/filtre pour tout ce qui est en short ou en long\n",
" securities_to_trade = (shorts | longs)\n",
" \n",
" # Retourne le Pipeline\n",
" return Pipeline(columns={\n",
" 'longs':longs,\n",
" 'shorts':shorts,\n",
" 'perc_diff':percent_difference\n",
" },screen=securities_to_trade)"
]
}
],
"metadata": {
"kernelspec": {
"display_name": "Python 3.5",
"language": "python",
"name": "py35"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.5.9"
}
},
"nbformat": 4,
"nbformat_minor": 2
}