Update app.py

app.py

@@ -1,75 +1,89 @@
+import os
 import gradio as gr
-from huggingface_hub import InferenceClient
-
 from langchain_community.vectorstores import FAISS
-from langchain_community.document_loaders import PyPDFLoader
 from langchain_community.embeddings import HuggingFaceEmbeddings
-from langchain.chains import ConversationalRetrievalChain
+from langchain_community.document_loaders import PyPDFLoader
 from langchain_huggingface import HuggingFaceEndpoint
+from langchain.chains import ConversationalRetrievalChain
 
+# Global retrieval chain + history
 retrieval_chain = None
 chat_history = []
 
+# Utility to fetch token (prefer user > fallback to env)
+def get_hf_token(user_token: str | None = None) -> str | None:
+    return user_token.strip() if user_token and user_token.strip() else os.getenv("HUGGINGFACEHUB_API_TOKEN")
+
 
-def process_pdf(file, hf_token):
+# Step 1: Process PDF
+def process_pdf(file, token):
     global retrieval_chain
 
+    hf_token = get_hf_token(token)
+    if not hf_token:
+        return "❌ No Hugging Face API token provided."
+
     # Load and split PDF
     loader = PyPDFLoader(file.name)
     documents = loader.load()
 
-    # Embed with sentence transformers
+    # Embed documents
     embeddings = HuggingFaceEmbeddings(model_name="sentence-transformers/all-MiniLM-L6-v2")
     vectorstore = FAISS.from_documents(documents, embeddings)
 
-    # Use HuggingFaceEndpoint instead of HuggingFaceHub
+    retriever = vectorstore.as_retriever()
+
+    # Build retrieval chain with DeepSeek model
     llm = HuggingFaceEndpoint(
-        repo_id="deepseek-ai/DeepSeek-R1-0528",
+        repo_id="deepseek-ai/deepseek-llm-R1-0528",
         huggingfacehub_api_token=hf_token,
-        task="text-generation",
-    )
-
-    retrieval_chain = ConversationalRetrievalChain.from_llm(
-        llm=llm,
-        retriever=vectorstore.as_retriever()
     )
+    retrieval_chain = ConversationalRetrievalChain.from_llm(llm, retriever=retriever)
 
-    return "PDF processed. You can now ask questions!"
+    return "✅ PDF processed. You can now ask questions!"
 
 
-def respond(message, history: list[dict[str, str]], hf_token: gr.OAuthToken):
+# Step 2: Respond to user questions
+def respond(message, history, token):
     global retrieval_chain, chat_history
 
-    if retrieval_chain is None:
-        return "Please upload a PDF first."
+    hf_token = get_hf_token(token)
+    if not hf_token:
+        return "❌ No Hugging Face API token provided.", history
 
-    # Use invoke() instead of deprecated __call__
-    result = retrieval_chain.invoke({
-        "question": message,
-        "chat_history": chat_history
-    })
+    if retrieval_chain is None:
+        return "⚠️ Please upload and process a PDF first.", history
 
+    # Run query against retriever
+    result = retrieval_chain.invoke({"question": message, "chat_history": chat_history})
     answer = result["answer"]
+
     chat_history.append((message, answer))
-    return answer
+    return answer, chat_history
 
 
+# Gradio UI
 with gr.Blocks() as demo:
-    with gr.Sidebar():
-        hf_login = gr.LoginButton()
-        pdf_upload = gr.File(label="Upload PDF", file_types=[".pdf"])
-        status_box = gr.Textbox(label="Status", interactive=False)
-
-    chatbot = gr.ChatInterface(
-        respond,
-        type="messages"
+    with gr.Row():
+        pdf_input = gr.File(label="Upload PDF", file_types=[".pdf"])
+        token_input = gr.Textbox(label="HuggingFace Token (optional)", type="text")
+        process_btn = gr.Button("Process PDF")
+
+    chatbot = gr.Chatbot(label="Chat with your PDF")
+    msg = gr.Textbox(label="Ask a question")
+
+    process_btn.click(
+        fn=lambda file, token: process_pdf(file, token),
+        inputs=[pdf_input, token_input],
+        outputs=[]
     )
 
-    pdf_upload.upload(
-        fn=lambda file, token: process_pdf(file, token.token),
-        inputs=[pdf_upload, hf_login],
-        outputs=[status_box]
+    msg.submit(
+        fn=lambda message, history, token: respond(message, history, token),
+        inputs=[msg, chatbot, token_input],
+        outputs=[chatbot, chatbot]
     )
 
+
 if __name__ == "__main__":
-    demo.launch()
+    demo.launch()