Stuff

app.py

-# from langchain_core.prompts import ChatPromptTemplate
-# from langchain_openai import ChatOpenAI
-# from langchain_core.messages import HumanMessage, SystemMessage, AIMessage
 import config
 import interface
-import initializer
-def main():    
-    if (config.INITIALIZE_ON_RUN):
-        initializer.run()
-    
-    gradio_app = interface.initialize()
-    gradio_app.launch(share = config.GRADIO_PUBLIC)
+from controllers import stocks
+import multiprocessing as mp
 
+print("Acquiring exchanges")
+try:
+    stocks.retrieve_exchanges()
+except:
+    print("Cannot retrieve exchanges from the API")
+    exit()
+else:
+    print("Exchange list updated!")
+
+# Due to API Limits, only Philippine stocks are considered
+print("Acquiring tickers")
+try:
+    tickers = stocks.retrieve_tickers("PSE")
+except:
+    print("Cannot retrieve tickers from the API")
+    exit()
+else:
+    print("Ticker list updated!")
+
+print("Acquiring stock data")
+processes = [mp.Process(target = stocks.retrieve_eod_historical_data, args = ("PSE", i["Code"])) for i in tickers]
+for p in processes:
+    p.start()
+for p in processes:
+    p.join()
+
+print("Initializing gradio app")
+gradio_app = interface.initialize()
+gradio_app.launch(share = config.GRADIO_PUBLIC)
 
 
 
-if __name__ == "__main__":
-    main()
\ No newline at end of file

config.py

@@ -4,7 +4,6 @@ BASE_DATA_PATH = 'data/'
 # Directories must end with a '/'
 TICKERS_DIRECTORY = 'tickers/'
 STOCKS_DIRECTORY = 'stocks/'
-FUSED_DATA_DIRECTORY = 'fused/'
 # Files may include its following path and must end with the respective extension
 EXCHANGES_FILEPATH = 'exchanges.json'
 
@@ -19,8 +18,7 @@ def get_tickers_directory() -> str:
 def get_stocks_directory() -> str:
     return BASE_DATA_PATH + STOCKS_DIRECTORY
 
-def get_fused_data_directory() -> str:
-    return BASE_DATA_PATH + FUSED_DATA_DIRECTORY
+
 
 def get_exchanges_filepath() -> str:
     return BASE_DATA_PATH + EXCHANGES_FILEPATH

controllers/rag.py

@@ -3,7 +3,6 @@ from chromadb.config import DEFAULT_TENANT, DEFAULT_DATABASE, Settings
 from langchain_core.documents import Document
 from langchain_openai import OpenAIEmbeddings
 from utility import stock_data_helper
-from langchain_text_splitters import RecursiveJsonSplitter
 import os
 try:
     import secret
@@ -13,6 +12,7 @@ except:
     
 def get_vectordb():
     embeddings = OpenAIEmbeddings(model="text-embedding-3-large", api_key = OPENAIKEY)
+    print("Loading local data")
     tickers = stock_data_helper.get_stock_codes("PSE")
     stocks = [stock_data_helper.get_stock_data("PSE", i['Code']) for i in tickers]
 
@@ -48,7 +48,7 @@ def get_vectordb():
         Name: "General/All stocks/tickers/companies in the philippines",
         Country: "Philippines"
         !HISTORICAL DATA:
-        {str(stock_data_helper.get_stock_codes)}
+        {str(stock_data_helper.get_stock_codes("PSE"))}
         """, 
         metadata ={
                 "Name": "General/All stocks/tickers/companies in the philippines",

initializer.py

-from controllers import stocks
-from utility import stock_data_helper
-import multiprocessing as mp
-def run():
-    try:
-        stocks.retrieve_exchanges()
-    except:
-        print("Cannot retrieve exchanges from the API")
-        exit()
-    else:
-        print("Exchange list updated!")
-
-    # Due to API Limits, only Philippine stocks are considered
-    try:
-        tickers = stocks.retrieve_tickers("PSE")
-    except:
-        print("Cannot retrieve tickers from the API")
-        exit()
-    else:
-        print("Ticker list updated!")
-
-    
-    tickers = stock_data_helper.get_stock_codes("PSE")
-    processes = [mp.Process(target = stocks.retrieve_eod_historical_data, args = ("PSE", i["Code"])) for i in tickers]
-    for p in processes:
-        p.start()
-    for p in processes:
-        p.join()
-    
-    # try:
-    #     stock_data_helper.create_fused_data("PSE")
-    # except Exception as exception:
-    #     print(exception)
-if __name__ == "__main__":
-    run()
\ No newline at end of file

interface.py

 import gradio as gr
 from controllers import rag, prompt
-vector_db = rag.get_vectordb()
+
+vector_db = None
 
 def get_prompt_response(query:str):
+    if vector_db == None:
+        vector_db = rag.get_vectordb()
     context = vector_db.similarity_search(query,k=10)
     ai_message = prompt.send_prompt(query, str(context))
     return ai_message.content

utility/stock_data_helper.py

@@ -5,10 +5,9 @@ import os
 EXCHANGES_FILEPATH = config.get_exchanges_filepath()
 TICKERS_DIRECTORY = config.get_tickers_directory()
 STOCKS_DIRECTORY = config.get_stocks_directory()
-FUSED_DATA_DIRECTORY = config.get_fused_data_directory()
 
 def get_stock_codes(exchange_code:str) -> list[str]:
-
+    print("Attempting to load local tickers")
     try:
         with open(TICKERS_DIRECTORY + f"{exchange_code}.json", 'r') as file:
             data = json.load(file)
@@ -20,7 +19,7 @@ def get_stock_codes(exchange_code:str) -> list[str]:
     return data
 
 def get_stock_data(exchange_code:str, stock_code:str):
-
+    print("Attempting to load local stock data")
     try:
         with open(STOCKS_DIRECTORY + f"{exchange_code}/{stock_code}.json", 'r') as file:
             data = json.load(file)
@@ -30,27 +29,3 @@ def get_stock_data(exchange_code:str, stock_code:str):
     
     return data
 
-def get_fused_data(exchange_code:str, stock_code:str):
-
-    try:
-        with open(FUSED_DATA_DIRECTORY + f"{exchange_code}/{stock_code}.json", 'r') as file:
-            data = json.load(file)
-    except:
-        print("File not found")
-        exit()
-    
-    return data
-
-def create_fused_data(exchange_code:str):
-    stock_codes = get_stock_codes(exchange_code)
-
-    if not os.path.exists(f"{FUSED_DATA_DIRECTORY}{exchange_code}/"):
-        os.makedirs(f"{FUSED_DATA_DIRECTORY}{exchange_code}/")
-
-    for stock in stock_codes:
-        stock['data'] = get_stock_data(exchange_code, stock['Code'])
-    
-        with open(f"{FUSED_DATA_DIRECTORY}{exchange_code}/" + f'{stock['Code']}.json', 'w') as filename:
-            json.dump(stock_codes, filename)
-        
-    return stock_codes
\ No newline at end of file