Update app.py

app.py

@@ -1,15 +1,15 @@
-import os
 import gradio as gr
 from langchain_community.vectorstores import FAISS
 from langchain_community.embeddings import HuggingFaceEmbeddings
 from langchain_community.document_loaders import PyPDFLoader
-from langchain.text_splitter import RecursiveCharacterTextSplitter
+from langchain_text_splitters import RecursiveCharacterTextSplitter
 from langchain_community.llms import HuggingFaceHub
 
+
 # -----------------------------
-# 1. LOAD & PROCESS PDF
+# LOAD & SPLIT PDF
 # -----------------------------
-def load_pdf(file):
+def process_pdf(file):
     loader = PyPDFLoader(file.name)
     pages = loader.load()
 
@@ -23,7 +23,7 @@ def load_pdf(file):
 
 
 # -----------------------------
-# 2. BUILD VECTOR DATABASE
+# CREATE VECTOR DB
 # -----------------------------
 def create_db(docs):
     embeddings = HuggingFaceEmbeddings(
@@ -35,16 +35,15 @@ def create_db(docs):
 
 
 # -----------------------------
-# 3. QA ENGINE (FIXED RETRIEVAL)
+# QUESTION ANSWERING
 # -----------------------------
 def ask_question(file, question):
     if file is None:
         return "Please upload a PDF first."
 
-    docs = load_pdf(file)
+    docs = process_pdf(file)
     db = create_db(docs)
 
-    # Better retrieval (IMPORTANT FIX)
     retrieved_docs = db.similarity_search(question, k=5)
 
     context = "\n\n".join([doc.page_content for doc in retrieved_docs])
@@ -53,12 +52,10 @@ def ask_question(file, question):
         return "No relevant information found in the document."
 
     prompt = f"""
-You are a precise assistant that answers ONLY from the provided medical report.
+You are a helpful assistant. Answer ONLY using the given context.
 
-Rules:
-- If answer is not in context, say "Not found in report"
-- Do not guess
-- Be concise
+If the answer is not present, reply:
+"Not found in report"
 
 Context:
 {context}
@@ -69,30 +66,27 @@ Question:
 Answer:
 """
 
-    # Using free HuggingFace model (works on HF Spaces CPU)
     llm = HuggingFaceHub(
         repo_id="google/flan-t5-base",
         model_kwargs={"temperature": 0.3, "max_length": 256}
     )
 
-    response = llm.invoke(prompt)
-    return response
+    result = llm.invoke(prompt)
+    return result
 
 
 # -----------------------------
-# 4. GRADIO UI
+# GRADIO UI
 # -----------------------------
-with gr.Blocks(title="Medical PDF RAG App") as app:
-
-    gr.Markdown("# 🧠 Medical PDF Question Answering (RAG)")
+with gr.Blocks() as app:
+    gr.Markdown("# 🧠 Medical Report Q&A (RAG)")
 
     file = gr.File(label="Upload PDF")
-    question = gr.Textbox(label="Ask a Question")
+    question = gr.Textbox(label="Ask your question")
 
     btn = gr.Button("Get Answer")
     output = gr.Textbox(label="Answer")
 
     btn.click(fn=ask_question, inputs=[file, question], outputs=output)
 
-
 app.launch()