Update app.py

app.py

@@ -13,7 +13,6 @@ from sklearn.metrics.pairwise import cosine_similarity
 MODEL_URL = "https://huggingface.co/melsonop/mistral-chat-pdf/resolve/main/mistral-7b-instruct-v0.1.Q4_K_M.gguf"
 MODEL_PATH = "models/mistral.gguf"
 
-# Download model if not exists
 if not os.path.exists(MODEL_PATH):
     os.makedirs("models", exist_ok=True)
     print("Downloading model...")
@@ -24,7 +23,6 @@ if not os.path.exists(MODEL_PATH):
                 f.write(chunk)
     print("Model download completed!")
 
-# Load model
 llm = Llama(
     model_path=MODEL_PATH,
     n_ctx=2048,
@@ -33,11 +31,10 @@ llm = Llama(
     verbose=False
 )
 
-# Load embedding model
 embed_model = SentenceTransformer("all-MiniLM-L6-v2")
 
 # -----------------------------------
-# PDF Processing
+# PDF + QA
 # -----------------------------------
 
 docs = []
@@ -52,39 +49,36 @@ def load_pdf(file):
 
     docs = text.split(". ")
     doc_embeddings = embed_model.encode(docs)
-    return "PDF loaded. You can now ask questions."
+    return "✅ PDF loaded. You can now ask questions."
 
-# -----------------------------------
-# Question Answering
-# -----------------------------------
-
-def answer_question(question):
+def answer_question(message, history):
     if not docs:
-        return "Please upload a PDF first."
-
-    question_embedding = embed_model.encode([question])
+        return "⚠️ Please upload and load a PDF first."
+    
+    question_embedding = embed_model.encode([message])
     similarities = cosine_similarity(question_embedding, doc_embeddings)[0]
     top_indices = similarities.argsort()[-3:][::-1]
     context = "\n".join([docs[i] for i in top_indices])
 
-    prompt = f"""You are a helpful assistant. Use the context below to answer the question.\n\nContext:\n{context}\n\nQuestion: {question}\nAnswer:"""
-
+    prompt = f"""You are a helpful assistant. Use the context below to answer the question.\n\nContext:\n{context}\n\nQuestion: {message}\nAnswer:"""
     output = llm(prompt=prompt, max_tokens=512, temperature=0.7)
     return output['choices'][0]['text'].strip()
 
 # -----------------------------------
-# Gradio UI
+# Gradio Interface
 # -----------------------------------
 
 with gr.Blocks() as demo:
-    gr.Markdown("# 📄 Chat with your PDF using Mistral GGUF")
+    gr.Markdown("## 🤖 Chat with your PDF using Mistral GGUF")
     
-    with gr.Row():
-        pdf_input = gr.File(label="Upload PDF", file_types=[".pdf"])
-        load_btn = gr.Button("Load PDF")
-    
-    chatbot = gr.ChatInterface(fn=answer_question)
-    
-    load_btn.click(load_pdf, inputs=[pdf_input], outputs=[chatbot.textbox])
+    pdf_file = gr.File(label="📄 Upload your PDF", file_types=[".pdf"])
+    load_button = gr.Button("🔄 Load PDF")
+
+    load_status = gr.Textbox(label="", interactive=False)
+
+    chat = gr.ChatInterface(fn=answer_question)
+
+    load_button.click(fn=load_pdf, inputs=pdf_file, outputs=load_status)
 
+# 💡 This must be at the top level for Spaces to detect:
 demo.launch()