Update app.py

app.py

@@ -16,14 +16,52 @@ openai.api_key = os.getenv("sk-proj-MKLxeaKCwQdMz3SXhUTz_r_mE0zN6wEo032M7ZQV4O2E
 persist_directory = "./docs/chroma/"
 vectordb = Chroma(persist_directory=persist_directory, embedding_function=embedding_model)
 
+from langchain_community.vectorstores import Chroma
+from langchain.embeddings import HuggingFaceEmbeddings
+from langchain.schema import Document
+
+# Load the embedding model
+model_name = "intfloat/e5-small"
+embedding_model = HuggingFaceEmbeddings(model_name=model_name)
+
+# Define the ChromaDB persist directory
+persist_directory = "./docs/chroma/"
+
+# ✅ Load ChromaDB (or create if empty)
+vectordb = Chroma(persist_directory=persist_directory, embedding_function=embedding_model)
+
+# ✅ Check if documents exist
+if vectordb._collection.count() == 0:
+    print("⚠️ No documents found in ChromaDB. Re-indexing dataset...")
+
+    # Sample dataset (Replace with real RunGalileo dataset)
+    documents = [
+        Document(page_content="HVAC systems help regulate indoor temperature."),
+        Document(page_content="Chiller plants are used in large buildings for cooling."),
+        Document(page_content="BACnet is a common protocol for building automation."),
+        Document(page_content="Heat pumps are essential in modern energy-efficient HVAC designs."),
+        Document(page_content="Smart thermostats improve energy efficiency through AI-based control.")
+    ]
+
+    # ✅ Insert documents into ChromaDB
+    vectordb.add_documents(documents)
+
+    print("✅ Documents successfully indexed into ChromaDB.")
+else:
+    print(f"✅ ChromaDB contains {vectordb._collection.count()} documents.")
+
+
 # ✅ Function to Retrieve Top-K Relevant Documents
 def retrieve_documents(question, k=5):
     """Retrieve top K relevant documents from ChromaDB"""
     docs = vectordb.similarity_search(question, k=k)
+    
     if not docs:
-        return ["No relevant documents found."]
+        return ["⚠️ No relevant documents found. Try a different query."]
+    
     return [doc.page_content for doc in docs]
 
+
 # ✅ Function to Generate AI Response
 def generate_response(question, context):
     """Generate AI response using OpenAI GPT-4"""
@@ -61,7 +99,7 @@ iface = gr.Interface(
         gr.Textbox(label="Generated Response"),
         gr.Textbox(label="Retrieved Documents")
     ],
-    title="RAG-Based Question Answering System for RunGalileo",
+    title="RAG-Based Question Answering System ",
     description="Enter a question and retrieve relevant documents with AI-generated response."
 )