Update app.py

app.py

@@ -1,3 +1,4 @@
+#Loading packages
 import logging
 import os
 import time
@@ -9,7 +10,6 @@ import re
 import numpy as np
 import json
 import difflib
-from functools import lru_cache
 
 # Set up logging
 logging.basicConfig(level=logging.INFO, format="%(asctime)s - %(levelname)s - %(message)s")
@@ -46,14 +46,170 @@ if df is not None:
 else:
     df_yearly = None
 
-# Hardcoded fallback for recent periods if dataset is incomplete
-fallback_returns = {
-    (2020, 2022): 8.3,  # Average annual return based on external data
-    (2015, 2024): 12.2,
-    (2020, 2024): 10.5
+# Response cache with financial data entries
+response_cache = {
+    "hi": "Hello! I'm FinChat, your financial advisor. How can I help with investing?",
+    "hello": "Hello! I'm FinChat, your financial advisor. How can I help with investing?",
+    "hey": "Hi there! Ready to discuss investment goals with FinChat?",
+    "what is better individual stocks or etfs?": (
+        "Here’s a comparison of individual stocks vs. ETFs:\n"
+        "1. **Individual Stocks**: High returns possible (e.g., Apple up 80% in 2020) but riskier due to lack of diversification. Require active research.\n"
+        "2. **ETFs**: Diversify risk by tracking indices (e.g., SPY, S&P 500, ~12% avg. return 2015–2024). Lower fees and less research needed.\n"
+        "3. **Recommendation**: Beginners should start with ETFs; experienced investors may add stocks.\n"
+        "Consult a financial planner."
+    ),
+    "is $100 per month enough to invest?": (
+        "Yes, $100 per month is enough to start investing. Here’s why and how:\n"
+        "1. **Feasibility**: Brokerages like Fidelity have no minimums, and commission-free trading eliminates fees.\n"
+        "2. **Options**: Buy fractional shares of ETFs (e.g., SPY, ~$622/share in 2025) with $100.\n"
+        "3. **Strategy**: Use dollar-cost averaging to invest monthly, reducing market timing risks.\n"
+        "4. **Growth**: At 10% annual return, $100 monthly could grow to ~$41,000 in 20 years.\n"
+        "5. **Tips**: Ensure an emergency fund; diversify.\n"
+        "Consult a financial planner."
+    ),
+    "can i invest $100 a month?": (
+        "Yes, $100 a month is sufficient. Here’s how:\n"
+        "1. **Brokerage**: Open an account with Fidelity or Vanguard (no minimums).\n"
+        "2. **Investments**: Buy fractional shares of ETFs like SPY ($100 buys ~0.16 shares in 2025).\n"
+        "3. **Approach**: Use dollar-cost averaging for steady growth.\n"
+        "4. **Long-Term**: At 10% return, $100 monthly could reach ~$41,000 in 20 years.\n"
+        "5. **Tips**: Prioritize an emergency fund and diversify.\n"
+        "Consult a financial planner."
+    ),
+    "hi, give me step-by-step investing advice": (
+        "Here’s a step-by-step guide to start investing:\n"
+        "1. Open a brokerage account (e.g., Fidelity, Vanguard) if 18 or older.\n"
+        "2. Deposit an affordable amount, like $100, after an emergency fund.\n"
+        "3. Research and buy an ETF (e.g., SPY) using Yahoo Finance.\n"
+        "4. Monitor monthly and enable dividend reinvesting.\n"
+        "5. Use dollar-cost averaging ($100 monthly) to reduce risk.\n"
+        "6. Diversify across sectors.\n"
+        "Consult a financial planner."
+    ),
+    "hi, pretend you are a financial advisor. now tell me how can i start investing in stock market?": (
+        "Here’s a guide to start investing:\n"
+        "1. Learn from Investopedia or 'The Intelligent Investor.'\n"
+        "2. Set goals (e.g., retirement) and assess risk.\n"
+        "3. Choose a brokerage (Fidelity, Vanguard).\n"
+        "4. Start with ETFs (e.g., SPY) or mutual funds.\n"
+        "5. Use dollar-cost averaging ($100-$500 monthly).\n"
+        "6. Diversify and monitor.\n"
+        "Consult a financial planner."
+    ),
+    "do you have a list of companies you recommend?": (
+        "I can’t recommend specific companies without data. Try ETFs like SPY (S&P 500, ~12% avg. return 2015–2024) or QQQ (tech). "
+        "Research stocks like Apple (AAPL, ~80% return in 2020) or Johnson & Johnson on Yahoo Finance.\n"
+        "Consult a financial planner."
+    ),
+    "how do i start investing in stocks?": (
+        "Learn from Investopedia. Set goals and assess risk. Open a brokerage account (Fidelity, Vanguard) "
+        "and start with ETFs (e.g., SPY, ~12% avg. return 2015–2024). Consult a financial planner."
+    ),
+    "what's the difference between stocks and bonds?": (
+        "Stocks are company ownership with high risk and growth potential (e.g., S&P 500 ~12% avg. return 2015–2024). Bonds are loans to companies/governments "
+        "with lower risk and steady interest. Diversify for balance."
+    ),
+    "how much should i invest?": (
+        "Invest what you can afford after expenses and an emergency fund. Start with $100-$500 monthly "
+        "in ETFs like SPY (~12% avg. return 2015–2024). Consult a financial planner."
+    ),
+    "what is dollar-cost averaging?": (
+        "Dollar-cost averaging is investing a fixed amount regularly (e.g., $100 monthly) in ETFs, "
+        "reducing risk by spreading purchases over time."
+    ),
+    "give me few investing idea": (
+        "Here are investing ideas:\n"
+        "1. Open a brokerage account (e.g., Fidelity) if 18 or older.\n"
+        "2. Deposit $100 or what you can afford.\n"
+        "3. Buy a researched ETF (e.g., SPY, ~12% avg. return 2015–2024) or index fund.\n"
+        "4. Check regularly and enable dividend reinvesting.\n"
+        "5. Use dollar-cost averaging (e.g., monthly buys).\n"
+        "Consult a financial planner."
+    ),
+    "give me investing tips": (
+        "Here are investing tips:\n"
+        "1. Educate yourself with Investopedia or books.\n"
+        "2. Open a brokerage account (e.g., Vanguard).\n"
+        "3. Start small with ETFs like SPY (~12% avg. return 2015–2024).\n"
+        "4. Invest regularly using dollar-cost averaging.\n"
+        "5. Diversify to manage risk.\n"
+        "Consult a financial planner."
+    ),
+    "how to start investing": (
+        "Here’s how to start investing:\n"
+        "1. Study basics on Investopedia.\n"
+        "2. Open a brokerage account (e.g., Fidelity).\n"
+        "3. Deposit $100 or more after securing savings.\n"
+        "4. Buy an ETF like SPY (~12% avg. return 2015–2024) after research.\n"
+        "5. Invest monthly with dollar-cost averaging.\n"
+        "Consult a financial planner."
+    ),
+    "investing advice": (
+        "Here’s investing advice:\n"
+        "1. Learn basics from Investopedia.\n"
+        "2. Open a brokerage account (e.g., Vanguard).\n"
+        "3. Start with $100 in an ETF like SPY (~12% avg. return 2015–2024).\n"
+        "4. Use dollar-cost averaging for regular investments.\n"
+        "5. Monitor and diversify your portfolio.\n"
+        "Consult a financial planner."
+    ),
+    "steps to invest": (
+        "Here are steps to invest:\n"
+        "1. Educate yourself using Investopedia.\n"
+        "2. Open a brokerage account (e.g., Fidelity).\n"
+        "3. Deposit an initial $100 after savings.\n"
+        "4. Buy an ETF like SPY (~12% avg. return 2015–2024) after research.\n"
+        "5. Use dollar-cost averaging monthly.\n"
+        "Consult a financial planner."
+    ),
+    "what is the s&p 500 index fund average growth rate?": (
+        "The S&P 500 index fund’s average annual return is approximately 10–12% over the long term (1927–2025), including dividends, based on historical data. "
+        "For example, from 2015 to 2024, it averaged ~12.2% annually. Returns vary yearly due to market conditions. Consult a financial planner."
+    ),
+    "what was the s&p 500 return in 2020?": (
+        "The S&P 500 returned approximately 16.3% in 2020, including dividends, driven by recovery from the COVID-19 market crash."
+    ),
+    "what was the s&p 500 return in 2022?": (
+        "The S&P 500 returned approximately -18.1% in 2022, impacted by high inflation and interest rate hikes."
+    ),
+    "what is the average annual growth rate of the s&p 500 from 2000 to 2010?": (
+        "The S&P 500’s average annual growth rate from 2000 to 2010 was approximately 0.4%, including dividends, impacted by the dot-com crash and 2008 financial crisis."
+    ),
+    "what is the average annual growth rate of the s&p 500 from 2011 to 2016?": (
+        "The S&P 500’s average annual growth rate from 2011 to 2016 was approximately 12.7%, including dividends, driven by post-financial crisis recovery."
+    ),
+    "what was the average annual return of the s&p 500 between 2010 and 2020?": (
+        "The S&P 500’s average annual return from 2010 to 2020 was approximately 13.6%, including dividends, driven by post-financial crisis recovery."
+    ),
+    "what will my return be in 10 years if i invest $5000 into s&p 500 right now?": (
+        "Assuming a 10% average annual return, a $5,000 investment in the S&P 500 would grow to approximately $12,974 in 10 years with annual compounding. "
+        "This is based on the historical average return of 10–12% (1927–2025). Future returns vary and are not guaranteed. Consult a financial planner."
+    ),
+    "what was the 1-year average annual growth rate of the s&p 500 from 2020?": (
+        "The S&P 500 returned approximately 16.3% in 2020, including dividends, driven by recovery from the COVID-19 market crash."
+    ),
+    "what was the 3-year average annual growth rate of the s&p 500 from 2018?": (
+        "The S&P 500’s average annual growth rate from 2018 to 2020 was approximately 10.2%, including dividends, based on historical data."
+    ),
+    "what was the 5-year average annual growth rate of the s&p 500 from 2016?": (
+        "The S&P 500’s average annual growth rate from 2016 to 2020 was approximately 13.6%, including dividends, driven by strong market recovery."
+    ),
+    "what is the average return rate of the s&p 500 in the past 10 years?": (
+        "The S&P 500’s average annual return rate from 2015 to 2024 was approximately 12.2%, including dividends, based on historical data."
+    )
 }
 
-# Load model and tokenizer at startup
+# Load persistent cache
+cache_file = "cache.json"
+try:
+    if os.path.exists(cache_file):
+        with open(cache_file, 'r') as f:
+            response_cache.update(json.load(f))
+        logger.info("Loaded persistent cache from cache.json")
+except Exception as e:
+    logger.warning(f"Failed to load cache.json: {e}")
+
+# Load model and tokenizer (use fine-tuned model if available)
 model_name = "./finetuned_model" if os.path.exists("./finetuned_model") else "distilgpt2"
 try:
     logger.info(f"Loading tokenizer for {model_name}")
@@ -61,14 +217,11 @@ try:
     tokenizer.pad_token = tokenizer.eos_token
     logger.info(f"Loading model {model_name}")
     with torch.inference_mode():
-        if os.path.exists("./finetuned_model/distilgpt2_traced.pt"):
-            model = torch.jit.load("./finetuned_model/distilgpt2_traced.pt")
-        else:
-            model = AutoModelForCausalLM.from_pretrained(model_name, torch_dtype=torch.float16, low_cpu_mem_usage=True).to(device)
-            sample_input = tokenizer("What was the average annual return of the S&P 500 between 2020 and 2022?", return_tensors="pt", padding=True, truncation=True)
-            sample_input = {k: v.to(device) for k, v in sample_input.items()}
-            model = torch.jit.trace(model, (sample_input["input_ids"], sample_input["attention_mask"]))
-            model.save("./finetuned_model/distilgpt2_traced.pt")
+        model = AutoModelForCausalLM.from_pretrained(
+            model_name,
+            torch_dtype=torch.float16,
+            low_cpu_mem_usage=True
+        ).to(device)
     logger.info(f"Successfully loaded model: {model_name}")
 except Exception as e:
     logger.error(f"Error loading model/tokenizer: {e}")
@@ -76,38 +229,38 @@ except Exception as e:
 
 # Pre-tokenize prompt prefix
 prompt_prefix = (
-    "You are FinChat, a financial advisor with expertise in stock market performance. Provide concise, accurate answers with historical data for S&P 500 queries. "
-    "For period-specific queries, use precise year ranges and calculate average annual returns. For investment return queries, use compound interest calculations "
-    "based on historical averages. Avoid repetition and ensure answers are relevant.\n\n"
+    "You are FinChat, a financial advisor with expertise in stock market performance. Provide detailed, numbered list advice with clear reasoning for investing prompts, "
+    "including precise historical data when relevant (e.g., S&P 500 returns for specific years or periods). For investment return queries, use compound interest calculations "
+    "based on historical averages. Avoid repetition and incomplete answers. Explain why each step or choice is beneficial.\n\n"
     "Example 1:\n"
     "Q: What is the S&P 500’s average annual return?\n"
-    "A: The S&P 500’s average annual return is ~10–12% over the long term (1927–2025), including dividends.\n\n"
+    "A: The S&P 500’s average annual return is ~10–12% over the long term (1927–2025), including dividends.\n"
+    "1. This reflects historical data adjusted for inflation and dividends.\n"
+    "2. Returns vary yearly (e.g., 16.3% in 2020) due to market conditions.\n"
+    "3. ETFs like SPY track this index for broad market exposure.\n\n"
     "Example 2:\n"
     "Q: What will $5,000 be worth in 10 years if invested in the S&P 500?\n"
-    "A: Assuming a 10% average annual return, a $5,000 investment in the S&P 500 would grow to approximately $12,974 in 10 years with annual compounding.\n\n"
+    "A: Assuming a 10% average annual return, a $5,000 investment in the S&P 500 would grow to approximately $12,974 in 10 years with annual compounding.\n"
+    "1. This uses the historical average return of 10–12% (1927–2025).\n"
+    "2. Future returns vary and are not guaranteed.\n\n"
     "Example 3:\n"
-    "Q: What was the average annual return of the S&P 500 between 2020 and 2022?\n"
-    "A: The S&P 500’s average annual return from 2020 to 2022 was approximately 8.3%, including dividends, with significant volatility due to the COVID-19 recovery and 2022 bear market.\n\n"
+    "Q: What was the average annual return of the S&P 500 between 2010 and 2020?\n"
+    "A: The S&P 500’s average annual return from 2010 to 2020 was approximately 13.6%, including dividends.\n"
+    "1. This period includes strong recovery post-financial crisis.\n"
+    "2. Dividends contribute significantly to total returns.\n\n"
     "Q: "
 )
-prefix_tokens = tokenizer(prompt_prefix, return_tensors="pt", truncation=True, max_length=512)["input_ids"].to(device)
+prefix_tokens = tokenizer(prompt_prefix, return_tensors="pt", truncation=True, max_length=512).to(device)
 
-# Substring matching for cache with exact year matching
-@lru_cache(maxsize=100)
-def get_closest_cache_key(message):
+# Substring matching for cache with fuzzy matching
+def get_closest_cache_key(message, cache_keys):
     message = message.lower().strip()
-    year_match = re.search(r'(\d{4})\s*(?:and|to|-|–)\s*(\d{4})', message)
-    if year_match:
-        start_year, end_year = year_match.groups()
-        for key in response_cache.keys():
-            if f"{start_year} and {end_year}" in key or f"{start_year} to {end_year}" in key or f"{start_year}–{end_year}" in key:
-                return key
-    matches = difflib.get_close_matches(message, response_cache.keys(), n=1, cutoff=0.7)
+    matches = difflib.get_close_matches(message, cache_keys, n=1, cutoff=0.8)
     return matches[0] if matches else None
 
 # Parse period from user input
 def parse_period(query):
-    # Match specific year ranges (e.g., "between 2020 and 2022", "2020–2022")
+    # Match specific year ranges (e.g., "between 2010 and 2020", "2000–2008")
     match = re.search(r'(?:between|from)\s*(\d{4})\s*(?:and|to|-|–)\s*(\d{4})', query, re.IGNORECASE)
     if match:
         start_year, end_year = map(int, match.groups())
@@ -118,7 +271,7 @@ def parse_period(query):
         duration, start_year = map(int, match.groups())
         end_year = start_year + duration - 1
         return start_year, end_year, duration
-    # Match general duration queries (e.g., "past 5 years", "3-year growth rate")
+    # Match general duration queries (e.g., "past 10 years", "3-year growth rate")
     match = re.search(r'past\s*(\d+)-year|\b(\d+)-year.*(?:return|growth\s*rate)', query, re.IGNORECASE)
     if match:
         duration = int(match.group(1) or match.group(2))
@@ -130,13 +283,6 @@ def parse_period(query):
 
 # Calculate average growth rate
 def calculate_growth_rate(start_year, end_year, duration=None):
-    if (start_year, end_year) in fallback_returns:
-        avg_return = fallback_returns[(start_year, end_year)]
-        if duration:
-            response = f"The S&P 500’s {duration}-year average annual return from {start_year} to {end_year} was approximately {avg_return:.1f}%, including dividends."
-        else:
-            response = f"The S&P 500’s average annual return from {start_year} to {end_year} was approximately {avg_return:.1f}%, including dividends."
-        return avg_return, response
     if df_yearly is None or start_year is None or end_year is None:
         return None, "Data not available or invalid period."
     df_period = df_yearly[(df_yearly['Year'] >= start_year) & (df_yearly['Year'] <= end_year)]
@@ -180,7 +326,8 @@ def chat_with_model(user_input, history=None, is_processing=False):
 
         # Normalize and check cache
         cache_key = user_input.lower().strip()
-        closest_key = get_closest_cache_key(cache_key)
+        cache_keys = list(response_cache.keys())
+        closest_key = cache_key if cache_key in response_cache else get_closest_cache_key(cache_key, cache_keys)
         if closest_key:
             logger.info(f"Cache hit for: {closest_key}")
             response = response_cache[closest_key]
@@ -237,8 +384,7 @@ def chat_with_model(user_input, history=None, is_processing=False):
         # Construct prompt
         full_prompt = prompt_prefix + user_input + "\nA:"
         try:
-            inputs = tokenizer(full_prompt, return_tensors="pt", padding=True, truncation=True)
-            inputs = {k: v.to(device) for k, v in inputs.items()}
+            inputs = tokenizer(full_prompt, return_tensors="pt", truncation=True, max_length=512).to(device)
         except Exception as e:
             logger.error(f"Error tokenizing input: {e}")
             response = f"Error: Failed to process input: {str(e)}"
@@ -254,7 +400,14 @@ def chat_with_model(user_input, history=None, is_processing=False):
         with torch.inference_mode():
             logger.info("Generating response with model")
             gen_start_time = time.time()
-            outputs = model(inputs["input_ids"], inputs["attention_mask"])
+            outputs = model.generate(
+                **inputs,
+                max_new_tokens=40,  # Reduced for faster inference
+                min_length=20,
+                do_sample=False,
+                repetition_penalty=2.0,
+                pad_token_id=tokenizer.eos_token_id
+            )
             gen_end_time = time.time()
             logger.info(f"Generation time: {gen_end_time - gen_start_time:.2f} seconds")
         response = tokenizer.decode(outputs[0], skip_special_tokens=True)
@@ -288,7 +441,7 @@ def chat_with_model(user_input, history=None, is_processing=False):
 # Save cache on exit
 def save_cache():
     try:
-        with open("cache.json", "w") as f:
+        with open(cache_file, 'w') as f:
             json.dump(response_cache, f, indent=2)
         logger.info("Saved cache to cache.json")
     except Exception as e: