Update app.py

app.py

@@ -1,45 +1,70 @@
 import gradio as gr
 import openai
+import os
 from langchain.embeddings import HuggingFaceEmbeddings
 from langchain_community.vectorstores import Chroma
+from langchain.schema import Document
 
-# Set API Key
-openai.api_key = "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"
-
-# Load embedding model
+# ✅ Load the Sentence Transformer Embedding Model
 model_name = "intfloat/e5-small"
 embedding_model = HuggingFaceEmbeddings(model_name=model_name)
 
-# Load ChromaDB
+# ✅ Set up OpenAI API Key (Replace with your own API key)
+openai.api_key = os.getenv("OPENAI_API_KEY")  # Add in Hugging Face Secrets
+
+# ✅ Load ChromaDB with RunGalileo Dataset
 persist_directory = "./docs/chroma/"
 vectordb = Chroma(persist_directory=persist_directory, embedding_function=embedding_model)
 
-# Define RAG function
+# ✅ 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 [doc.page_content for doc in docs]
+
+# ✅ Function to Generate AI Response
+def generate_response(question, context):
+    """Generate AI response using OpenAI GPT-4"""
+    if not context or "No relevant documents found." in context:
+        return "No relevant context available. Try a different query."
+
+    full_prompt = f"Context: {context}\n\nQuestion: {question}"
+
+    try:
+        response = openai.ChatCompletion.create(
+            model="gpt-4",
+            messages=[
+                {"role": "system", "content": "You are an AI assistant that answers user queries based on the given context."},
+                {"role": "user", "content": full_prompt}
+            ],
+            max_tokens=300,
+            temperature=0.7
+        )
+        return response["choices"][0]["message"]["content"].strip()
+    except Exception as e:
+        return f"Error generating response: {str(e)}"
+
+# ✅ Full RAG Pipeline
 def rag_pipeline(question):
-    """Retrieve relevant documents and generate AI response"""
-    retrieved_docs = vectordb.similarity_search(question, k=5)
-    context = " ".join([doc.page_content for doc in retrieved_docs])
-    
-    # Generate AI response
-    full_prompt = f"Context: {context}\\n\\nQuestion: {question}"
-    response = openai migrate.create(
-        model="gpt-4",
-        messages=[{"role": "user", "content": full_prompt}],
-        max_tokens=300,
-        temperature=0.7
-    )
-    
-    return response['choices'][0]['message']['content'].strip(), retrieved_docs
-
-# Gradio UI
+    retrieved_docs = retrieve_documents(question, k=5)
+    context = " ".join(retrieved_docs)
+    response = generate_response(question, context)
+    return response, "\n\n".join(retrieved_docs)
+
+# ✅ Gradio UI Interface
 iface = gr.Interface(
     fn=rag_pipeline,
     inputs=gr.Textbox(label="Enter your question"),
-    outputs=[gr.Textbox(label="Generated Response"), gr.Textbox(label="Retrieved Documents")],
-    title="RAG-Based Question Answering System",
+    outputs=[
+        gr.Textbox(label="Generated Response"),
+        gr.Textbox(label="Retrieved Documents")
+    ],
+    title="RAG-Based Question Answering System for RunGalileo",
     description="Enter a question and retrieve relevant documents with AI-generated response."
 )
 
-# Launch Gradio app
+# ✅ Launch the Gradio App
 if __name__ == "__main__":
     iface.launch()