streamline ml training better

app/cron_trend_analysis.py

@@ -9,7 +9,7 @@ import pandas as pd
 from tqdm import tqdm
 import concurrent.futures
 import re
-
+import subprocess
 
 async def save_json(symbol, data):
     with open(f"json/trend-analysis/{symbol}.json", 'w') as file:
@@ -89,6 +89,14 @@ async def process_symbol(ticker, start_date, end_date, crypto_symbols):
         print(e)
 
 async def run():
+
+    #Train first model
+    try:
+        print('training...')
+        subprocess.run(["python3", "ml_models/classification.py", "--train"], check=True)
+    except subprocess.CalledProcessError as e:
+        print(f"Error running classification.py: {e}")
+
     con = sqlite3.connect('stocks.db')
     etf_con = sqlite3.connect('etf.db')
     crypto_con = sqlite3.connect('crypto.db')

app/ml_models/classification.py

@@ -20,6 +20,15 @@ import aiohttp
 import pickle
 import time
 
+import argparse
+
+# Set up argument parser
+parser = argparse.ArgumentParser(description="Train and test process script.")
+parser.add_argument('--train', action='store_true', help="Set to True to run training")
+    
+# Parse the arguments
+args = parser.parse_args()
+
 
 async def download_data(ticker, start_date, end_date, nth_day):
     try:
@@ -189,11 +198,14 @@ async def test_process(nth_day):
 
 
 async def main():
-    
-    for nth_day in [5,20,60]:
+    for nth_day in [5, 20, 60]:
         await train_process(nth_day)
-    
     await test_process(nth_day=5)
 
-# Run the main function
-#asyncio.run(main())
+if __name__ == "__main__":
+    
+    # Run main if --train is set to True
+    if args.train:
+        asyncio.run(main())
+    else:
+        print("Training not initiated. Pass --train True to start training.")

app/primary_cron_job.py

@@ -548,6 +548,15 @@ def run_earnings():
     ]
     run_command(command)
 
+def run_fomc_impact():
+    run_command(["python3", "cron_fomc_impact.py"])
+    command = [
+        "sudo", "rsync", "-avz", "-e", "ssh",
+        "/root/backend/app/json/fomc-impact",
+        f"root@{useast_ip_address}:/root/backend/app/json"
+    ]
+    run_command(command)
+
 def run_economy_indicator():
     run_command(["python3", "cron_economic_indicator.py"])
     command = [
@@ -557,6 +566,33 @@ def run_economy_indicator():
     ]
     run_command(command)
 
+def run_trend_analysis():
+    run_command(["python3", "cron_trend_analysis.py"])
+    command = [
+        "sudo", "rsync", "-avz", "-e", "ssh",
+        "/root/backend/app/json/trend-analysis",
+        f"root@{useast_ip_address}:/root/backend/app/json"
+    ]
+    run_command(command)
+
+def run_sentiment_analysis():
+    run_command(["python3", "cron_sentiment_analysis.py"])
+    command = [
+        "sudo", "rsync", "-avz", "-e", "ssh",
+        "/root/backend/app/json/sentiment-analysis",
+        f"root@{useast_ip_address}:/root/backend/app/json"
+    ]
+    run_command(command)
+
+def run_price_analysis():
+    run_command(["python3", "cron_price_analysis.py"])
+    command = [
+        "sudo", "rsync", "-avz", "-e", "ssh",
+        "/root/backend/app/json/price-analysis",
+        f"root@{useast_ip_address}:/root/backend/app/json"
+    ]
+    run_command(command)
+
 # Create functions to run each schedule in a separate thread
 def run_threaded(job_func):
     job_thread = threading.Thread(target=job_func)
@@ -577,6 +613,7 @@ schedule.every().day.at("07:30").do(run_threaded, run_financial_statements).tag(
 schedule.every().day.at("08:00").do(run_threaded, run_economy_indicator).tag('economy_indicator_job')
 schedule.every().day.at("08:00").do(run_threaded, run_cron_insider_trading).tag('insider_trading_job')
 schedule.every().day.at("08:30").do(run_threaded, run_dividends).tag('dividends_job')
+schedule.every().day.at("08:30").do(run_threaded, run_fomc_impact).tag('fomc_impact_job')
 schedule.every().day.at("09:00").do(run_threaded, run_congress_trading).tag('congress_job')
 schedule.every().day.at("10:00").do(run_threaded, run_shareholders).tag('shareholders_job')
 schedule.every().day.at("10:30").do(run_threaded, run_sec_filings).tag('sec_filings_job')
@@ -598,6 +635,9 @@ schedule.every().day.at("14:00").do(run_threaded, run_cron_sector).tag('sector_j
 schedule.every(2).days.at("01:00").do(run_threaded, run_market_maker).tag('markt_maker_job')
 schedule.every(2).days.at("08:30").do(run_threaded, run_financial_score).tag('financial_score_job')
 schedule.every().saturday.at("05:00").do(run_threaded, run_ownership_stats).tag('ownership_stats_job')
+schedule.every().saturday.at("08:00").do(run_threaded, run_trend_analysis).tag('trend_analysis_job')
+schedule.every().saturday.at("08:00").do(run_threaded, run_sentiment_analysis).tag('sentiment_analysis_job')
+schedule.every().saturday.at("08:00").do(run_threaded, run_price_analysis).tag('price_analysis_job')
 
 
 schedule.every(5).minutes.do(run_threaded, run_cron_market_movers).tag('market_movers_job')