Create app.py

app.py

@@ -0,0 +1,59 @@
+import os
+import faiss
+import numpy as np
+import streamlit as st
+from groq import Groq
+from sentence_transformers import SentenceTransformer
+
+# Initialize FAISS and Model
+VECTOR_DB_PATH = "vector_database.faiss"
+EMBEDDING_MODEL = "sentence-transformers/all-MiniLM-L6-v2"
+
+# Initialize embedding model
+embedding_model = SentenceTransformer(EMBEDDING_MODEL)
+
+# Load FAISS Index
+def load_faiss():
+    if not os.path.exists(VECTOR_DB_PATH):
+        st.error("Vector database not found! Ensure FAISS index is created.")
+        return None
+    index = faiss.read_index(VECTOR_DB_PATH)
+    return index
+
+faiss_index = load_faiss()
+
+# GROQ API setup
+GROQ_API_KEY = "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"  # Set in Hugging Face secrets
+client = Groq(api_key=GROQ_API_KEY)
+
+MODEL_ID = "deepseek-r1-distill-llama-70b"
+
+# Function to get nearest neighbor from FAISS
+def search_faiss(query, top_k=3):
+    query_embedding = embedding_model.encode(query, convert_to_numpy=True).reshape(1, -1)
+    distances, indices = faiss_index.search(query_embedding, top_k)
+    return indices
+
+# Function to call DeepSeek model from GROQ
+def generate_response(context, query):
+    prompt = f"Use the following retrieved context to answer the question:\n\nContext:\n{context}\n\nQuestion: {query}\nAnswer:"
+    response = client.chat.completions.create(
+        model=MODEL_ID,
+        messages=[{"role": "user", "content": prompt}],
+    )
+    return response.choices[0].message.content
+
+# Streamlit UI
+st.title("💡 AI Chat with FAISS & GROQ")
+st.write("Ask a question and get responses based on stored knowledge!")
+
+query = st.text_input("🔍 Enter your query:")
+if query:
+    if faiss_index is None:
+        st.error("FAISS database not loaded. Please check deployment.")
+    else:
+        indices = search_faiss(query)
+        retrieved_context = "\n".join([f"Chunk {i}: Retrieved text" for i in indices[0]])
+        response = generate_response(retrieved_context, query)
+        st.write("### 🤖 AI Response:")
+        st.write(response)