Update app.py

app.py

@@ -1,100 +1,104 @@
-import os
-import re
 import gradio as gr
-import numpy as np
-import faiss
-# Import the library
+import re
 from youtube_transcript_api import YouTubeTranscriptApi
 from sentence_transformers import SentenceTransformer
 from langchain_text_splitters import RecursiveCharacterTextSplitter
+import numpy as np
+import faiss
 from groq import Groq
+import os
 
 # ===============================
-# CONFIG & INITIALIZATION
+# MODULE 1: IMPORTS & GLOBALS
 # ===============================
 
-# Get API Key from Environment Variables (Set this in HF Space Secrets)
-GROQ_API_KEY = os.getenv("GROQ_API_KEY")
-groq_client = Groq(api_key=GROQ_API_KEY) if GROQ_API_KEY else None
-
-# Load embedding model
+# Initialize Models
 embedding_model = SentenceTransformer("all-MiniLM-L6-v2")
+groq_client = Groq(api_key=os.environ.get("GROQ_API_KEY"))
 
 # Global Storage
 vector_store = None
 chunks_store = []
 
 # ===============================
-# HELPER FUNCTIONS
+# MODULE 2: TRANSCRIPT FUNCTION
 # ===============================
 
 def extract_video_id(url):
-    """Extracts the 11-character YouTube video ID from various URL formats."""
-    regex = r"(?:v=|\/|be\/)([0-9A-Za-z_-]{11}).*"
-    match = re.search(regex, url)
-    if match:
-        return match.group(1)
-    return None
+    """Extract video ID from YouTube URL"""
+    patterns = [
+        r'(?:youtube\.com\/watch\?v=)([\w-]+)',
+        r'(?:youtu\.be\/)([\w-]+)',
+        r'(?:youtube\.com\/embed\/)([\w-]+)',
+        r'(?:youtube\.com\/v\/)([\w-]+)'
+    ]
+    
+    for pattern in patterns:
+        match = re.search(pattern, url)
+        if match:
+            return match.group(1)
+    raise ValueError("Invalid YouTube URL")
 
 def get_transcript(url):
-    """
-    Fetch transcript using the correct static method.
-    """
+    """Fetch transcript using YouTubeTranscriptApi"""
     try:
         video_id = extract_video_id(url)
-        if not video_id:
-            return "ERROR: Invalid YouTube URL. Could not find Video ID."
-        
-        # FIX: Calling the static method directly on the class
-        # We also try to fetch English by default or the first available
-        transcript_list = YouTubeTranscriptApi.get_transcript(video_id)
-        
-        full_text = " ".join([item['text'] for item in transcript_list])
+        transcript_data = YouTubeTranscriptApi.get_transcript(video_id)
+        full_text = " ".join([item['text'] for item in transcript_data])
         return full_text
     except Exception as e:
-        return f"ERROR: Could not retrieve transcript. (Details: {str(e)})"
+        return f"ERROR: {str(e)}"
+
+# ===============================
+# MODULE 3: VECTOR DATABASE
+# ===============================
 
 def process_transcript(transcript):
+    """Convert transcript → chunks → embeddings → FAISS index"""
     global vector_store, chunks_store
     
-    # Split text into manageable chunks
-    splitter = RecursiveCharacterTextSplitter(chunk_size=600, chunk_overlap=60)
+    # Step 1: Split text into chunks
+    splitter = RecursiveCharacterTextSplitter(
+        chunk_size=500,
+        chunk_overlap=50
+    )
     chunks = splitter.split_text(transcript)
     
-    # Create embeddings
+    # Step 2: Convert chunks → embeddings
     embeddings = embedding_model.encode(chunks)
     
-    # Initialize FAISS Index
+    # Step 3: Store in FAISS
     dimension = embeddings.shape[1]
     index = faiss.IndexFlatL2(dimension)
-    index.add(np.array(embeddings).astype('float32'))
+    index.add(np.array(embeddings))
     
-    # Store globally for retrieval
+    # Save globally
     vector_store = index
     chunks_store = chunks
 
+# ===============================
+# MODULE 4: RETRIEVAL
+# ===============================
+
 def retrieve_context(query, top_k=3):
-    if vector_store is None:
-        return ""
-    
+    """Retrieve most relevant chunks using similarity search"""
     query_embedding = embedding_model.encode([query])
-    distances, indices = vector_store.search(np.array(query_embedding).astype('float32'), top_k)
-    
-    # Fetch matching chunks
-    retrieved_chunks = [chunks_store[i] for i in indices[0] if i != -1]
+    distances, indices = vector_store.search(np.array(query_embedding), top_k)
+    retrieved_chunks = [chunks_store[i] for i in indices[0]]
     return "\n\n".join(retrieved_chunks)
 
+# ===============================
+# MODULE 5: LLM (GROQ)
+# ===============================
+
 def generate_answer(query):
-    if not groq_client:
-        return "Error: Groq API Key is not set in Hugging Face Secrets."
-    
+    """Full RAG pipeline: Query → Retrieve → Augment → Generate"""
     context = retrieve_context(query)
-    if not context:
-        return "I don't have any context from the video yet. Please process a video first."
-
+    
     prompt = f"""
-You are a professional AI Assistant. Use the provided context from a YouTube video to answer the user's question. 
-If the answer isn't in the context, say you don't know based on the video.
+You are a helpful AI assistant.
+
+Use ONLY the context below to answer.
 
 Context:
 {context}
@@ -102,22 +106,22 @@ Context:
 Question:
 {query}
 
-Answer:
+Answer clearly and accurately:
 """
     
     response = groq_client.chat.completions.create(
         model="llama-3.3-70b-versatile",
         messages=[{"role": "user", "content": prompt}]
     )
+    
     return response.choices[0].message.content
 
 # ===============================
-# UI LOGIC
+# MODULE 6: PROCESS PIPELINE
 # ===============================
 
-def process_video_ui(url):
-    if not url:
-        return "Please enter a valid URL", "❌ No URL"
+def process_video(url):
+    """Full pipeline: URL → Transcript → Embeddings → Ready for chat"""
     
     transcript = get_transcript(url)
     
@@ -125,46 +129,75 @@ def process_video_ui(url):
         return transcript, "❌ Failed to fetch transcript"
     
     process_transcript(transcript)
-    return transcript[:1500] + "...", "✅ Video processed! You can now chat."
-
-def chat_with_video_ui(user_query, history):
-    if not user_query:
-        return history, ""
     
+    status = "✅ Video processed! You can now chat."
+    
+    return transcript, status
+
+# ===============================
+# MODULE 7: CHAT FUNCTION
+# ===============================
+
+def chat_with_video(user_query, history):
+    """Handles chat interaction"""
     if vector_store is None:
-        history.append((user_query, "⚠️ Please process a video in the first tab before chatting."))
-        return history, ""
+        return history + [(user_query, "⚠️ Please process a video first.")]
     
     answer = generate_answer(user_query)
     history.append((user_query, answer))
-    return history, ""
-
-# ===============================
-# GRADIO INTERFACE
-# ===============================
-
-with gr.Blocks(theme=gr.themes.Soft()) as app:
-    gr.Markdown("# 🎥 YouTube RAG AI Expert")
-    gr.Markdown("Transcribe any YouTube video and chat with its content using Llama 3.3 & FAISS.")
-    
-    with gr.Tabs():
-        with gr.Tab("1. Load Video"):
-            url_input = gr.Textbox(label="YouTube Link", placeholder="https://www.youtube.com/watch?v=...")
-            process_btn = gr.Button("Transcribe & Index Video", variant="primary")
-            with gr.Row():
-                status_output = gr.Textbox(label="Status")
-                transcript_preview = gr.Textbox(label="Transcript Preview", lines=8)
-            
-            process_btn.click(process_video_ui, inputs=url_input, outputs=[transcript_preview, status_output])
-            
-        with gr.Tab("2. Chat with AI"):
-            chatbot = gr.Chatbot(height=500)
-            with gr.Row():
-                msg = gr.Textbox(label="Your Question", placeholder="What are the key takeaways?", scale=4)
-                submit = gr.Button("Ask", variant="primary", scale=1)
-            
-            submit.click(chat_with_video_ui, inputs=[msg, chatbot], outputs=[chatbot, msg])
-            msg.submit(chat_with_video_ui, inputs=[msg, chatbot], outputs=[chatbot, msg])
-
-if __name__ == "__main__":
-    app.launch()
+    return history
+
+# ===============================
+# MODULE 8: GRADIO UI
+# ===============================
+
+with gr.Blocks() as app:
+    
+    gr.Markdown("# 🎥 YouTube RAG Q&A System")
+    gr.Markdown("Process a YouTube video and chat with it using AI")
+    
+    # TAB 1: VIDEO PROCESSING
+    with gr.Tab("📥 Process Video"):
+        
+        url_input = gr.Textbox(
+            label="Enter YouTube URL",
+            placeholder="https://www.youtube.com/watch?v=..."
+        )
+        
+        process_btn = gr.Button("▶️ Transcribe & Process")
+        
+        transcript_output = gr.Textbox(
+            label="Transcript",
+            lines=10
+        )
+        
+        status_output = gr.Textbox(
+            label="Status"
+        )
+        
+        process_btn.click(
+            fn=process_video,
+            inputs=url_input,
+            outputs=[transcript_output, status_output]
+        )
+    
+    # TAB 2: CHAT
+    with gr.Tab("💬 Chat with Video"):
+        
+        chatbot = gr.Chatbot(height=400)
+        
+        user_input = gr.Textbox(
+            label="Ask a question about the video"
+        )
+        
+        send_btn = gr.Button("Send")
+        
+        send_btn.click(
+            fn=chat_with_video,
+            inputs=[user_input, chatbot],
+            outputs=chatbot
+        )
+
+# CHANGE 1: Add server parameters for Hugging Face
+# CHANGE 2: Remove debug=True
+app.launch(server_name="0.0.0.0", server_port=7860)