Merge pull request #4 from ryann-green/main

updating branch

predictions/main.py

@@ -320,7 +320,9 @@
     # total_results.append(all_results)
 
     increment_predictions(f'10_day_ahead_close/stock_performance/{date}/tickers','predictions/predictions_table.csv')
-    increment_non_trigger_evals((f'10_day_ahead_close/stock_performance/{date}/tickers','predictions/non_trigger_stocks.csv'))
+    increment_non_trigger_evals(f'10_day_ahead_close/stock_performance/{date}/tickers','predictions/non_trigger_stocks.csv')
+
+    # At this point we will run backtesting.py
 
 # pd.concat(total_results).to_csv(f'{new_folder_path}/total_results.csv')
 

predictions/utils/backtesting.py

@@ -3,23 +3,10 @@
 import pandas as pd
 import warnings
 import time
-from build_results import increment_backtest_results
+from build_results import increment_results
 
 warnings.filterwarnings("ignore")
 
-# steps to operationalize
-# 1. For stocks that don't meet the high or low triggers, add them to a separate csv with ticker and prediction date for future analysis
-# 2. Check to see the latest date for prediction date for each stock in the Summary results csv file. Find the results in the total_results csv that are greater than the ticker+pred_date from the summary file, and use those values to start this price evaluation script.
-# 3. Create different segments in sheets to evaluate success of stocks
-# 4. Identify top stock to trade on a day of the week based on best chance of profitability within 15 days
-    #segment by best combination of
-    # day of week success
-    # overall success rate
-    # predicted higher success rate
-    # predicted higher boolean flag .. validation whether or not my prediction prices actually give me the best chance of success based on subsequent high and low prices    
-# 5. Based on suggestion above, Figure out stock to trade on next possible trading day based on the latest daste in the "total results" csv file.
-
-
 # Input data for multiple stocks
 # stocks_data = pd.read_csv('total_results.csv')  
 stocks_data = pd.read_csv('predictions/non_trigger_stocks.csv')
@@ -121,7 +108,7 @@ def check_first_trigger(historical_data, ticker, buy_price, overall_success_rate
             "profit_pct_per_stock": profit_pct
         }
     else:
-        print(f"{(h_data['ticker'])} has no trigger price as of {last_date}")
+        print(f"{(ticker)} from {(last_date)} has no trigger price as of 1/9/2025")
 
         return None
 
@@ -141,6 +128,8 @@ def process_stocks(stocks_data, historical_data):
         ticker=stock[1]
         # print(ticker)
 
+        # if ticker in 'MA':
+
         ticker_data = historical_data.loc[historical_data['ticker']==ticker]
 
         # print(ticker_data)
@@ -190,13 +179,13 @@ def process_stocks(stocks_data, historical_data):
     historical_data = fetch_historical_data(stocks_data)
     processed_results,non_triggers = process_stocks(stocks_data, historical_data)
     # print(processed_results)
-    increment_backtest_results('predictions/backtest_results.csv',processed_results)
     results_df = pd.DataFrame(processed_results)
+    increment_results('predictions/backtest_results.csv',processed_results)
+    # print(results_df)
     # results_df.to_csv('summary_results.csv', index=False)
 
 
     # the below should get me a csv file of the trickers and dates that I'd need to check on the next run
     # will need to add a clause that ignores ticker from X period in the past
     non_triggers_df=pd.DataFrame(non_triggers)
-    # non_triggers_df.to_csv('non_trigger_stocks.csv', index=False)
-    # print(results_df)
+    non_triggers_df.to_csv('predictions/non_trigger_stocks.csv') 

predictions/utils/build_results.py

@@ -126,6 +126,7 @@ def increment_predictions(target,table):
 
 # run this everytime the predictions script is ran to add on to non_trigger_stocks for new predictions
 def increment_non_trigger_evals (target,table):
+    print('Running increment_non_trigger_evals function')
     folder = os.path.expanduser(f'~/Code/stock_predictions/{target}')
 
     #get the DataFrame of the file that I want to append the results to for the predictions
@@ -158,7 +159,7 @@ def increment_non_trigger_evals (target,table):
                 if os.path.isfile(file_path):
                     if  'all_results' in file_path :
 
-                        print(file_path)
+                        # print(file_path)
                         try:
                             new_data = pd.read_csv(file_path)[['ticker'
                                               ,'latest_close'
@@ -167,7 +168,7 @@ def increment_non_trigger_evals (target,table):
                                               ,'last_date_for_prediction'
                                               ,'adj_prediction_price_w_high_inc'
                                               ,'adj_prediction_higher'
-                                              ,'stop_loss '
+                                              ,'stop_loss'
                                               ]]
 
                             new_data['buy_price']=new_data['latest_close']
@@ -186,7 +187,7 @@ def increment_non_trigger_evals (target,table):
                                                       , "adj_prediction_price_w_high_inc": "adj_prediction_price"
                                                       , "adj_prediction_higher": "adj_price_higher"
                                                       , "stop_loss ": "stop_loss"}, inplace=True)
-                            print(new_data.columns)
+                            # print(new_data.columns)
 
 
                             new_date=max(new_data['last_date'])
@@ -200,7 +201,7 @@ def increment_non_trigger_evals (target,table):
                                 print(f"Length of new data: {len(new_data)}")
                                 print(f"Length of combined data after concat {len(pd.concat([incremented_summary,new_data]))}")
 
-                                pd.concat([incremented_summary,new_data]).to_csv(table)
+                                pd.concat([incremented_summary,new_data]).reset_index(drop=True).to_csv(table)
                                 # print(pd.concat([incremented_summary,new_data]))
 
                                 # print(f'Non-Triggered Stocks updated with data from {new_date}')
@@ -222,15 +223,34 @@ def increment_non_trigger_evals (target,table):
     return 'increment_non_trigger_evals'
 
 # run this in backtesting to increment to existing backtest results table
-def increment_backtest_results(old_results,new_results):
-    print('old results')
-    print(pd.read_csv(old_results))
+def increment_results(old_results_path,new_results):
 
-    print('new_results')
-    print(new_results)
+    print('------------------------------------------')
+
+    print('Running increment_results function')
+    print('Incrementing Backtesting Results')
+    print('Retrieving existing backtesting results')
+    old_backtest_results=pd.read_csv(old_results_path)
+    # old_max_date=max(old_backtest_results['pred_date'])
 
+    # print('new_results')
+    print('Retrieving new backtesting results')
+    new_backtest_results=pd.DataFrame(new_results)
+    # new_max_date=max(new_backtest_results['pred_date'])
+
+    # print(f"Current max date: {old_max_date}")
+    # print(f"New data date: {new_max_date}")
+    print(f"Length of old table: {len(old_backtest_results)}")
+    print(f"Length of new data: {len(new_backtest_results)}")
+    print(f"Length of combined data after concat {len(pd.concat([old_backtest_results,new_backtest_results]))}")
+
+    pd.concat([old_backtest_results,new_backtest_results]).to_csv(old_results_path)
+
+    return 'incremented_results'
+
+def print_incremented_summary (old_results,new_results):
 
-    return 'increment_backtest_resultss'
+    return "Incremented summary statistics"
 
 
 # using for testing functions
@@ -240,5 +260,5 @@ def increment_backtest_results(old_results,new_results):
 
 #     # print(pd.read_csv('predictions/non_trigger_stocks.csv'))
 #     incremented_table = increment_non_trigger_evals('10_day_ahead_close/stock_performance/2025-01-04/tickers','predictions/non_trigger_stocks.csv')
-    increment_backtest_results('predictions/backtest_results.csv')
-#     print(incremented_table)
+    # increment_results('predictions/backtest_results.csv')
+    print('Hi')

predictions/utils/get_data.py

@@ -6,8 +6,6 @@ def get_initial_df (ticker,start_date,end_date):
         data = yf.download(ticker, start=start_date, end=end_date)
         # Calculate leading technical indicators
         data['ALMA'] = ta.alma(data['Close'])
-        stoch_rsi = ta.stochrsi(data['Close'])
-        # stoch_rsi = ta.rsi(data['Close'])
 
         print('this is data close')
         print(data['Close'])