Upload 4 files

README.md

@@ -1,12 +1,9 @@
----
-title: Medical Qa Assistant
-emoji: 📚
-colorFrom: yellow
-colorTo: pink
-sdk: streamlit
-sdk_version: 1.44.1
-app_file: app.py
-pinned: false
----
+# 🧠 Medical QA Assistant
 
-Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference
+A Retrieval-Augmented Generation (RAG) app that answers medical questions using research paper snippets from the `pubmed_qa` dataset.
+
+- Built with: Streamlit, Hugging Face Transformers, FAISS
+- Embedding Model: BioBERT via `sentence-transformers`
+- Generator Model: BART
+
+Great for medical students or MedBot-style experimentation.

app.py

@@ -0,0 +1,14 @@
+import streamlit as st
+from rag_pipeline import load_rag_pipeline, ask_question
+
+st.set_page_config(page_title="Medical QA Assistant 💊", layout="centered")
+st.title("🧠 Medical QA Assistant")
+st.markdown("Ask medical questions and get answers based on PubMed research.")
+
+query = st.text_input("Enter your medical question:")
+if query:
+    with st.spinner("Searching..."):
+        pipe = load_rag_pipeline()
+        result = ask_question(pipe, query)
+        st.markdown("### 💡 Answer")
+        st.write(result)

rag_pipeline.py

@@ -0,0 +1,27 @@
+from transformers import pipeline
+from sentence_transformers import SentenceTransformer
+import faiss
+import numpy as np
+
+# Dummy knowledge base - replace with real embeddings in production
+knowledge_base = [
+    {"text": "Aspirin is used to reduce fever and relieve mild to moderate pain.", "embedding": None},
+    {"text": "Hypertension is a condition in which the blood pressure in the arteries is elevated.", "embedding": None},
+    {"text": "Diabetes is a chronic condition that affects how the body processes blood sugar.", "embedding": None},
+]
+
+def load_rag_pipeline():
+    embedder = SentenceTransformer("sentence-transformers/all-MiniLM-L6-v2")
+    for entry in knowledge_base:
+        entry["embedding"] = embedder.encode(entry["text"])
+    index = faiss.IndexFlatL2(len(knowledge_base[0]["embedding"]))
+    index.add(np.array([entry["embedding"] for entry in knowledge_base]))
+    return {"embedder": embedder, "index": index, "texts": [entry["text"] for entry in knowledge_base]}
+
+def ask_question(pipe, query):
+    query_vec = pipe["embedder"].encode(query)
+    D, I = pipe["index"].search(np.array([query_vec]), k=1)
+    context = pipe["texts"][I[0][0]]
+    generator = pipeline("text2text-generation", model="facebook/bart-large")
+    answer = generator(f"question: {query} context: {context}", max_length=100, do_sample=False)
+    return answer[0]['generated_text']

requirements.txt

@@ -0,0 +1,5 @@
+streamlit
+torch
+transformers
+sentence-transformers
+faiss-cpu