added assignment files

app.py

@@ -0,0 +1,138 @@
+import streamlit as st
+import torch
+from rag import FinancialChatbot
+import warnings
+from transformers import logging
+
+# Suppress warnings
+warnings.filterwarnings("ignore")
+logging.set_verbosity_error()
+
+torch.classes.__path__ = []
+
+# Use session state to persist the chatbot instance
+if "chatbot" not in st.session_state:
+    st.session_state.chatbot = FinancialChatbot(data_path="Nestle_Financtial_report_till2023.xlsx")
+
+def fetch_answer_from_backend(query):
+    """Calls the backend function to get an answer."""
+    return st.session_state.chatbot.get_answer(query)  # Use session state chatbot instance
+
+
+
+# Initialize Session State for Chat History
+if "chat_history" not in st.session_state:
+    st.session_state.chat_history = []
+
+if "loading" not in st.session_state:
+    st.session_state.loading = False
+
+# Layout and Title
+st.title("Financial RAG Chat Assistant")
+
+
+# st.markdown(
+#     """
+#     <style>
+#     body, .stApp {
+#         background: url('https://encrypted-tbn0.gstatic.com/images?q=tbn:ANd9GcQlSQ-kkQFs_BAOhel_cd_CkFoHN1V-5bUO7vv8NwRp0cjftW6Wm2utQSE&s=10');
+#         background-size: cover;
+#         # background-attachment: fixed;
+#         # background-position: center;
+#         # background-repeat: no-repeat;
+#     }
+#     # .stApp {
+#     #     background: transparent;
+#     # }
+#     # .stTextInput, .stTextArea, .stButton {
+#     #     background-color: rgba(0, 0, 0, 0.5);
+#     #     border-radius: 10px;
+#     #     color: #ffffff;
+#     # }
+#     # .stTextInput > div > input {
+#     #     color: #ffffff;
+#     # }
+#     # .stButton button {
+#     #     color: #ffffff;
+#     #     background-color: #4CAF50;
+#     #     border: none;
+#     #     border-radius: 10px;
+#     # }
+#     </style>
+#     """,
+#     unsafe_allow_html=True
+# )
+
+
+
+
+
+# Display Chat History
+for chat in st.session_state.chat_history:
+    # User's query on the right
+    st.markdown(
+        f"""
+        <div style='text-align: right; max-width: 75%; float: right; clear: both;'>
+            <div style='font-size: 13px; color: #4a4a4a; margin: 0 10px;'>You</div>
+            <div style='background-color: #d3d3d3; color: black; padding: 10px; border-radius: 10px; 
+                        box-shadow: 1px 1px 5px rgba(0, 0, 0, 0.1);'>
+                {chat['question']}
+            </div>
+        </div>
+        """,
+        unsafe_allow_html=True
+    )
+    
+    # Assistant's label and response on the left
+    if chat["answer"] is not None:
+        st.markdown(
+            f"""
+            <div style='text-align: left; max-width: 75%; float: left; clear: both;'>
+                <div style='font-size: 13px; color: #4a4a4a; margin: 0 10px;'>Assistant</div>
+                <div style='background-color: #add8e6; color: black; padding: 10px; border-radius: 10px; 
+                        box-shadow: 1px 1px 5px rgba(0, 0, 0, 0.1);'>
+                    {chat['answer']}
+                </div>
+            </div>
+            """,
+        unsafe_allow_html=True
+    )
+    
+    # Confidence Score (Below the answer)
+    if chat["confidence"] is not None:
+        st.markdown(
+            f"<div style='clear: both; color: #4a4a4a; font-size: 13px;'>Confidence: {chat['confidence'] * 100}%</div>",
+            unsafe_allow_html=True
+        )
+        st.divider()  # Adds a visual divider between Q&A pairs
+
+# User Input (Always at the Bottom)
+user_input = st.chat_input("Ask a financial question...")
+
+# If user inputs a question
+if user_input:
+    # Add question to chat history and show loading animation
+    st.session_state.chat_history.append({
+        "question": user_input,
+        "answer": None,  # Placeholder for the answer
+        "confidence": None  # Placeholder for the confidence score
+    })
+    st.session_state.loading = True
+    st.rerun()  # Refresh to display the question immediately
+
+# If loading, simulate the API call and update the last question's answer
+if st.session_state.loading:
+    with st.spinner("Fetching answer..."):
+        # Get the last question
+        last_question = st.session_state.chat_history[-1]["question"]
+
+        # API Call
+        answer, confidence = fetch_answer_from_backend(last_question)
+
+        # Update the last chat history item with the answer and confidence
+        st.session_state.chat_history[-1]["answer"] = answer
+        st.session_state.chat_history[-1]["confidence"] = confidence
+
+    # Stop loading and refresh to show the answer
+    st.session_state.loading = False
+    st.rerun()

index_map.pkl

@@ -0,0 +1,3 @@
+version https://git-lfs.github.com/spec/v1
+oid sha256:887ad3476658fdd9c5b5ddac6c642b3682beb5b1b67c262026fad69dbc64ddcd
+size 13547

index_map.txt

@@ -0,0 +1,250 @@
+Net Sales/Income from operations - year Dec '23 is: 19126.3
+Net Sales/Income from operations - year Dec '22 is: 16787.43
+Net Sales/Income from operations - year Dec '21 is: 14633.72
+Net Sales/Income from operations - year Dec '20 is: 13290.16
+Net Sales/Income from operations - year Dec '19 is: 12295.27
+Other Operating Income - year Dec '23 is: --
+Other Operating Income - year Dec '22 is: 77.63
+Other Operating Income - year Dec '21 is: 75.69
+Other Operating Income - year Dec '20 is: 59.87
+Other Operating Income - year Dec '19 is: 73.63
+Total Income From Operations - year Dec '23 is: 19126.3
+Total Income From Operations - year Dec '22 is: 16865.06
+Total Income From Operations - year Dec '21 is: 14709.41
+Total Income From Operations - year Dec '20 is: 13350.03
+Total Income From Operations - year Dec '19 is: 12368.9
+EXPENDITURE - year Dec '23 is: nan
+EXPENDITURE - year Dec '22 is: nan
+EXPENDITURE - year Dec '21 is: nan
+EXPENDITURE - year Dec '20 is: nan
+EXPENDITURE - year Dec '19 is: nan
+Consumption of Raw Materials - year Dec '23 is: 8054.95
+Consumption of Raw Materials - year Dec '22 is: 7652.11
+Consumption of Raw Materials - year Dec '21 is: 6154.1
+Consumption of Raw Materials - year Dec '20 is: 5554.24
+Consumption of Raw Materials - year Dec '19 is: 5150.3
+Purchase of Traded Goods - year Dec '23 is: 417.77
+Purchase of Traded Goods - year Dec '22 is: 317.13
+Purchase of Traded Goods - year Dec '21 is: 227.52
+Purchase of Traded Goods - year Dec '20 is: 189
+Purchase of Traded Goods - year Dec '19 is: 217.81
+Increase/Decrease in Stocks - year Dec '23 is: -40.01
+Increase/Decrease in Stocks - year Dec '22 is: -245.79
+Increase/Decrease in Stocks - year Dec '21 is: -62.7
+Increase/Decrease in Stocks - year Dec '20 is: -69.33
+Increase/Decrease in Stocks - year Dec '19 is: -144.19
+Power & Fuel - year Dec '23 is: --
+Power & Fuel - year Dec '22 is: --
+Power & Fuel - year Dec '21 is: --
+Power & Fuel - year Dec '20 is: --
+Power & Fuel - year Dec '19 is: --
+Employees Cost - year Dec '23 is: 1849.18
+Employees Cost - year Dec '22 is: 1629.87
+Employees Cost - year Dec '21 is: 1521.3
+Employees Cost - year Dec '20 is: 1500.95
+Employees Cost - year Dec '19 is: 1262.95
+Depreciation - year Dec '23 is: 428.91
+Depreciation - year Dec '22 is: 402.6
+Depreciation - year Dec '21 is: 390.19
+Depreciation - year Dec '20 is: 370.38
+Depreciation - year Dec '19 is: 316.36
+Excise Duty - year Dec '23 is: --
+Excise Duty - year Dec '22 is: --
+Excise Duty - year Dec '21 is: --
+Excise Duty - year Dec '20 is: --
+Excise Duty - year Dec '19 is: --
+Admin. And Selling Expenses - year Dec '23 is: --
+Admin. And Selling Expenses - year Dec '22 is: --
+Admin. And Selling Expenses - year Dec '21 is: --
+Admin. And Selling Expenses - year Dec '20 is: --
+Admin. And Selling Expenses - year Dec '19 is: --
+R & D Expenses - year Dec '23 is: --
+R & D Expenses - year Dec '22 is: --
+R & D Expenses - year Dec '21 is: --
+R & D Expenses - year Dec '20 is: --
+R & D Expenses - year Dec '19 is: --
+Provisions And Contingencies - year Dec '23 is: --
+Provisions And Contingencies - year Dec '22 is: 9.85
+Provisions And Contingencies - year Dec '21 is: -25.17
+Provisions And Contingencies - year Dec '20 is: 13.97
+Provisions And Contingencies - year Dec '19 is: 24.88
+Exp. Capitalised - year Dec '23 is: --
+Exp. Capitalised - year Dec '22 is: --
+Exp. Capitalised - year Dec '21 is: --
+Exp. Capitalised - year Dec '20 is: --
+Exp. Capitalised - year Dec '19 is: --
+Other Expenses - year Dec '23 is: 4373.42
+Other Expenses - year Dec '22 is: 3774.06
+Other Expenses - year Dec '21 is: 3302.82
+Other Expenses - year Dec '20 is: 2959.7
+Other Expenses - year Dec '19 is: 2992.85
+P/L Before Other Inc. , Int., Excpt. Items & Tax - year Dec '23 is: 4042.08
+P/L Before Other Inc. , Int., Excpt. Items & Tax - year Dec '22 is: 3325.23
+P/L Before Other Inc. , Int., Excpt. Items & Tax - year Dec '21 is: 3201.35
+P/L Before Other Inc. , Int., Excpt. Items & Tax - year Dec '20 is: 2831.12
+P/L Before Other Inc. , Int., Excpt. Items & Tax - year Dec '19 is: 2547.94
+Other Income - year Dec '23 is: 121.21
+Other Income - year Dec '22 is: 101
+Other Income - year Dec '21 is: 120.11
+Other Income - year Dec '20 is: 145.85
+Other Income - year Dec '19 is: 246.88
+P/L Before Int., Excpt. Items & Tax - year Dec '23 is: 4163.29
+P/L Before Int., Excpt. Items & Tax - year Dec '22 is: 3426.23
+P/L Before Int., Excpt. Items & Tax - year Dec '21 is: 3321.46
+P/L Before Int., Excpt. Items & Tax - year Dec '20 is: 2976.97
+P/L Before Int., Excpt. Items & Tax - year Dec '19 is: 2794.82
+Interest - year Dec '23 is: 119.29
+Interest - year Dec '22 is: 154.33
+Interest - year Dec '21 is: 201.19
+Interest - year Dec '20 is: 164.18
+Interest - year Dec '19 is: 119.83
+P/L Before Exceptional Items & Tax - year Dec '23 is: 4044
+P/L Before Exceptional Items & Tax - year Dec '22 is: 3271.9
+P/L Before Exceptional Items & Tax - year Dec '21 is: 3120.27
+P/L Before Exceptional Items & Tax - year Dec '20 is: 2812.79
+P/L Before Exceptional Items & Tax - year Dec '19 is: 2674.99
+Exceptional Items - year Dec '23 is: -5.71
+Exceptional Items - year Dec '22 is: --
+Exceptional Items - year Dec '21 is: -236.5
+Exceptional Items - year Dec '20 is: --
+Exceptional Items - year Dec '19 is: --
+P/L Before Tax - year Dec '23 is: 4038.29
+P/L Before Tax - year Dec '22 is: 3271.9
+P/L Before Tax - year Dec '21 is: 2883.77
+P/L Before Tax - year Dec '20 is: 2812.79
+P/L Before Tax - year Dec '19 is: 2674.99
+Tax - year Dec '23 is: 1039.62
+Tax - year Dec '22 is: 865.45
+Tax - year Dec '21 is: 738.91
+Tax - year Dec '20 is: 730.36
+Tax - year Dec '19 is: 705.44
+P/L After Tax from Ordinary Activities - year Dec '23 is: 2998.67
+P/L After Tax from Ordinary Activities - year Dec '22 is: 2406.45
+P/L After Tax from Ordinary Activities - year Dec '21 is: 2144.86
+P/L After Tax from Ordinary Activities - year Dec '20 is: 2082.43
+P/L After Tax from Ordinary Activities - year Dec '19 is: 1969.55
+Prior Year Adjustments - year Dec '23 is: --
+Prior Year Adjustments - year Dec '22 is: --
+Prior Year Adjustments - year Dec '21 is: --
+Prior Year Adjustments - year Dec '20 is: --
+Prior Year Adjustments - year Dec '19 is: --
+Extra Ordinary Items - year Dec '23 is: --
+Extra Ordinary Items - year Dec '22 is: --
+Extra Ordinary Items - year Dec '21 is: --
+Extra Ordinary Items - year Dec '20 is: --
+Extra Ordinary Items - year Dec '19 is: --
+Net Profit/(Loss) For the Period - year Dec '23 is: 2998.67
+Net Profit/(Loss) For the Period - year Dec '22 is: 2406.45
+Net Profit/(Loss) For the Period - year Dec '21 is: 2144.86
+Net Profit/(Loss) For the Period - year Dec '20 is: 2082.43
+Net Profit/(Loss) For the Period - year Dec '19 is: 1969.55
+Equity Share Capital - year Dec '23 is: 96.42
+Equity Share Capital - year Dec '22 is: 96.42
+Equity Share Capital - year Dec '21 is: 96.42
+Equity Share Capital - year Dec '20 is: 96.42
+Equity Share Capital - year Dec '19 is: 96.42
+Reserves Excluding Revaluation Reserves - year Dec '23 is: --
+Reserves Excluding Revaluation Reserves - year Dec '22 is: 2362.75
+Reserves Excluding Revaluation Reserves - year Dec '21 is: 1988.06
+Reserves Excluding Revaluation Reserves - year Dec '20 is: 1922.92
+Reserves Excluding Revaluation Reserves - year Dec '19 is: 1822.45
+Equity Dividend Rate (%) - year Dec '23 is: --
+Equity Dividend Rate (%) - year Dec '22 is: 2200
+Equity Dividend Rate (%) - year Dec '21 is: 2000
+Equity Dividend Rate (%) - year Dec '20 is: 2000
+Equity Dividend Rate (%) - year Dec '19 is: 3420
+EPS Before Extra Ordinary - year Dec '23 is: nan
+EPS Before Extra Ordinary - year Dec '22 is: nan
+EPS Before Extra Ordinary - year Dec '21 is: nan
+EPS Before Extra Ordinary - year Dec '20 is: nan
+EPS Before Extra Ordinary - year Dec '19 is: nan
+Basic EPS - year Dec '23 is: 31.1
+Basic EPS - year Dec '22 is: 249.58
+Basic EPS - year Dec '21 is: 222.46
+Basic EPS - year Dec '20 is: 215.98
+Basic EPS - year Dec '19 is: 204.28
+Diluted EPS - year Dec '23 is: 31.1
+Diluted EPS - year Dec '22 is: --
+Diluted EPS - year Dec '21 is: 222.46
+Diluted EPS - year Dec '20 is: 215.98
+Diluted EPS - year Dec '19 is: 204.28
+EPS After Extra Ordinary - year Dec '23 is: nan
+EPS After Extra Ordinary - year Dec '22 is: nan
+EPS After Extra Ordinary - year Dec '21 is: nan
+EPS After Extra Ordinary - year Dec '20 is: nan
+EPS After Extra Ordinary - year Dec '19 is: nan
+Basic EPS - year Dec '23 is: 31.1
+Basic EPS - year Dec '22 is: 249.58
+Basic EPS - year Dec '21 is: 222.46
+Basic EPS - year Dec '20 is: 215.98
+Basic EPS - year Dec '19 is: 204.28
+Diluted EPS - year Dec '23 is: 31.1
+Diluted EPS - year Dec '22 is: --
+Diluted EPS - year Dec '21 is: 222.46
+Diluted EPS - year Dec '20 is: 215.98
+Diluted EPS - year Dec '19 is: 204.28
+Public Share Holding - year Dec '23 is: nan
+Public Share Holding - year Dec '22 is: nan
+Public Share Holding - year Dec '21 is: nan
+Public Share Holding - year Dec '20 is: nan
+Public Share Holding - year Dec '19 is: nan
+No Of Shares (Crores) - year Dec '23 is: --
+No Of Shares (Crores) - year Dec '22 is: --
+No Of Shares (Crores) - year Dec '21 is: --
+No Of Shares (Crores) - year Dec '20 is: --
+No Of Shares (Crores) - year Dec '19 is: --
+Share Holding (%) - year Dec '23 is: --
+Share Holding (%) - year Dec '22 is: --
+Share Holding (%) - year Dec '21 is: --
+Share Holding (%) - year Dec '20 is: --
+Share Holding (%) - year Dec '19 is: --
+Promoters and Promoter Group Shareholding - year Dec '23 is: nan
+Promoters and Promoter Group Shareholding - year Dec '22 is: nan
+Promoters and Promoter Group Shareholding - year Dec '21 is: nan
+Promoters and Promoter Group Shareholding - year Dec '20 is: nan
+Promoters and Promoter Group Shareholding - year Dec '19 is: nan
+a) Pledged/Encumbered - year Dec '23 is: nan
+a) Pledged/Encumbered - year Dec '22 is: nan
+a) Pledged/Encumbered - year Dec '21 is: nan
+a) Pledged/Encumbered - year Dec '20 is: nan
+a) Pledged/Encumbered - year Dec '19 is: nan
+- Number of shares (Crores) - year Dec '23 is: --
+- Number of shares (Crores) - year Dec '22 is: --
+- Number of shares (Crores) - year Dec '21 is: --
+- Number of shares (Crores) - year Dec '20 is: --
+- Number of shares (Crores) - year Dec '19 is: --
+- Per. of shares (as a % of the total sh. of prom. and promoter group) - year Dec '23 is: --
+- Per. of shares (as a % of the total sh. of prom. and promoter group) - year Dec '22 is: --
+- Per. of shares (as a % of the total sh. of prom. and promoter group) - year Dec '21 is: --
+- Per. of shares (as a % of the total sh. of prom. and promoter group) - year Dec '20 is: --
+- Per. of shares (as a % of the total sh. of prom. and promoter group) - year Dec '19 is: --
+- Per. of shares (as a % of the total Share Cap. of the company) - year Dec '23 is: --
+- Per. of shares (as a % of the total Share Cap. of the company) - year Dec '22 is: --
+- Per. of shares (as a % of the total Share Cap. of the company) - year Dec '21 is: --
+- Per. of shares (as a % of the total Share Cap. of the company) - year Dec '20 is: --
+- Per. of shares (as a % of the total Share Cap. of the company) - year Dec '19 is: --
+b) Non-encumbered - year Dec '23 is: nan
+b) Non-encumbered - year Dec '22 is: nan
+b) Non-encumbered - year Dec '21 is: nan
+b) Non-encumbered - year Dec '20 is: nan
+b) Non-encumbered - year Dec '19 is: nan
+- Number of shares (Crores) - year Dec '23 is: --
+- Number of shares (Crores) - year Dec '22 is: --
+- Number of shares (Crores) - year Dec '21 is: --
+- Number of shares (Crores) - year Dec '20 is: --
+- Number of shares (Crores) - year Dec '19 is: --
+- Per. of shares (as a % of the total sh. of prom. and promoter group) - year Dec '23 is: --
+- Per. of shares (as a % of the total sh. of prom. and promoter group) - year Dec '22 is: --
+- Per. of shares (as a % of the total sh. of prom. and promoter group) - year Dec '21 is: --
+- Per. of shares (as a % of the total sh. of prom. and promoter group) - year Dec '20 is: --
+- Per. of shares (as a % of the total sh. of prom. and promoter group) - year Dec '19 is: --
+- Per. of shares (as a % of the total Share Cap. of the company) - year Dec '23 is: --
+- Per. of shares (as a % of the total Share Cap. of the company) - year Dec '22 is: --
+- Per. of shares (as a % of the total Share Cap. of the company) - year Dec '21 is: --
+- Per. of shares (as a % of the total Share Cap. of the company) - year Dec '20 is: --
+- Per. of shares (as a % of the total Share Cap. of the company) - year Dec '19 is: --
+Notes - year Dec '23 is: |202312
+Notes - year Dec '22 is: |202212
+Notes - year Dec '21 is: |202112
+Notes - year Dec '20 is: |202012
+Notes - year Dec '19 is: |201912

rag.py

@@ -0,0 +1,147 @@
+import torch
+import pandas as pd
+import faiss
+import numpy as np
+import re
+import os
+from sentence_transformers import SentenceTransformer
+from transformers import AutoModelForCausalLM, AutoTokenizer
+
+class FinancialChatbot:
+    def __init__(self, data_path, model_name="all-MiniLM-L6-v2", qwen_model_name="Qwen/Qwen2-0.5B-Instruct"):
+        self.device = "cpu"
+        self.data_path = data_path  # Store data path
+
+        # Load SBERT for embeddings
+        self.sbert_model = SentenceTransformer(model_name, device=self.device)
+        self.sbert_model = self.sbert_model.half()
+
+        # Load Qwen model for text generation
+        self.qwen_model = AutoModelForCausalLM.from_pretrained(
+            qwen_model_name, torch_dtype=torch.float16, trust_remote_code=True
+        ).to(self.device)
+
+        self.qwen_tokenizer = AutoTokenizer.from_pretrained(qwen_model_name, trust_remote_code=True)
+
+        # Load or create FAISS index
+        self.load_or_create_index()
+
+    import os  # Import os for file checks
+
+    def load_or_create_index(self):
+        """Loads FAISS index and index_map if they exist, otherwise creates new ones."""
+        if os.path.exists("financial_faiss.index") and os.path.exists("index_map.txt"):
+            try:
+                self.faiss_index = faiss.read_index("financial_faiss.index")
+                with open("index_map.txt", "r", encoding="utf-8") as f:
+                    self.index_map = {i: line.strip() for i, line in enumerate(f)}
+                print("FAISS index and index_map loaded successfully.")
+            except Exception as e:
+                print(f"Error loading FAISS index: {e}. Recreating index...")
+                self.create_faiss_index()
+        else:
+            print("FAISS index or index_map not found. Creating a new one...")
+            self.create_faiss_index()
+
+
+    def create_faiss_index(self):
+        """Creates a FAISS index from the provided Excel file."""
+        df = pd.read_excel(self.data_path)
+        sentences = []
+        self.index_map = {}  # Initialize index_map
+
+        for row_idx, row in df.iterrows():
+            for col in df.columns[1:]:  # Ignore the first column (assumed to be labels)
+                sentence = f"{row[df.columns[0]]} - year {col} is: {row[col]}"
+                sentences.append(sentence)
+                self.index_map[len(self.index_map)] = sentence  # Store mapping
+
+        # Encode the sentences into embeddings
+        embeddings = self.sbert_model.encode(sentences, convert_to_numpy=True)
+
+        # Create FAISS index (FlatL2 for simplicity)
+        self.faiss_index = faiss.IndexFlatL2(embeddings.shape[1])
+        self.faiss_index.add(embeddings)
+
+        # Save index and index map
+        faiss.write_index(self.faiss_index, "financial_faiss.index")
+        with open("index_map.txt", "w", encoding="utf-8") as f:
+            for sentence in self.index_map.values():
+                f.write(sentence + "\n")
+
+    def query_faiss(self, query, top_k=3):
+        """Retrieves the top_k closest sentences from FAISS index."""
+        query_embedding = self.sbert_model.encode([query], convert_to_numpy=True)
+        distances, indices = self.faiss_index.search(query_embedding, top_k)
+
+        results = [self.index_map[idx] for idx in indices[0] if idx in self.index_map]
+        confidences = [1 - (dist / (np.max(distances[0]) or 1)) for dist in distances[0]]
+
+        return results, confidences
+
+    def moderate_query(self, query):
+        """Blocks inappropriate queries containing restricted words."""
+        BLOCKED_WORDS = re.compile(r"\b(hack|bypass|illegal|exploit|scam|kill|laundering|murder|suicide|self-harm)\b", re.IGNORECASE)
+        return not bool(BLOCKED_WORDS.search(query))
+
+    def generate_answer(self, context, question):
+        messages = [
+            {"role": "system", "content": "You are a financial assistant. Answer only finance-related questions. If the question is not related to finance, reply: 'I'm sorry, but I can only answer financial-related questions.' If the user greets you (e.g., 'Hello', 'Hi', 'Good morning'), respond politely with 'Hello! How can I assist you today?'."},
+            {"role": "user", "content": f"{question} - related contect extracted form db {context}"}
+        ]
+
+        # Use Qwen's chat template
+        input_text = self.qwen_tokenizer.apply_chat_template(
+            messages, tokenize=False, add_generation_prompt=True
+        )
+
+        # Tokenize and move input to device
+        inputs = self.qwen_tokenizer([input_text], return_tensors="pt").to(self.device)
+        self.qwen_model.config.pad_token_id = self.qwen_tokenizer.eos_token_id
+
+        # Generate response
+        outputs = self.qwen_model.generate(
+            inputs.input_ids,
+            max_new_tokens=50,
+            pad_token_id=self.qwen_tokenizer.eos_token_id,
+        )
+
+        # Extract only the newly generated part
+        generated_ids = outputs[:, inputs.input_ids.shape[1]:]  # Remove prompt part
+        response = self.qwen_tokenizer.batch_decode(generated_ids, skip_special_tokens=True)[0]
+
+        return response
+
+
+
+    def get_answer(self, query):
+        """Main function to process a user query and return an answer."""
+        
+        # Check if query is appropriate
+        if not self.moderate_query(query):
+            return "Inappropriate request.", 0.0
+
+        # Retrieve relevant documents and their confidence scores
+        retrieved_docs, confidences = self.query_faiss(query)
+        if not retrieved_docs:
+            return "No relevant information found.", 0.0
+
+        # Combine retrieved documents as context
+        context = " ".join(retrieved_docs)
+        avg_confidence = round(sum(confidences) / len(confidences), 2)
+
+        # Generate model response
+        model_response = self.generate_answer(context, query)
+
+        # Extract only the relevant part of the response
+        model_response = model_response.strip()
+        
+        # Ensure only the actual answer is returned
+        if model_response.lower() in ["i don't know", "no relevant information found"]:
+            return "I don't know.", avg_confidence
+        #print(avg_confidence)
+        if avg_confidence == 0.0:
+            return "Not relevant ", avg_confidence
+
+        
+        return model_response, avg_confidence

requirements.txt

@@ -0,0 +1,4 @@
+faiss-cpu
+sentence-transformers
+accelerate>=0.26.0
+openpyxl