Update app.py

app.py

@@ -1,52 +1,66 @@
-from fastapi import FastAPI, Request
-from fastapi.responses import HTMLResponse, JSONResponse
-from fastapi.templating import Jinja2Templates
+import gradio as gr
 from transformers import AutoTokenizer, AutoModelForQuestionAnswering, pipeline
 
-app = FastAPI(title="QA Dashboard Pro")
-
-# Load your fine-tuned model
+# ✅ Load your fine-tuned model
+# Use your Hugging Face model repo name (e.g., "MedhaCodes/qna_finetuned_model")
+# or use "saved_model" if running locally
 MODEL_PATH = "MedhaCodes/qna_finetuned_model"
+
+# Load the model and tokenizer
 qa_pipeline = pipeline(
     "question-answering",
     model=AutoModelForQuestionAnswering.from_pretrained(MODEL_PATH),
     tokenizer=AutoTokenizer.from_pretrained(MODEL_PATH)
 )
 
-templates = Jinja2Templates(directory="templates")
-
-@app.get("/", response_class=HTMLResponse)
-async def home(request: Request):
-    return templates.TemplateResponse("index.html", {"request": request})
-
-@app.post("/predict")
-async def predict(request: Request):
-    data = await request.json()
-    context = data.get("context")
-    questions_text = data.get("question")
-
+# Function to process multiple questions
+def answer_multiple_questions(context, questions_text):
     if not context or not questions_text:
-        return JSONResponse({"error": "Please provide both context and question"}, status_code=400)
+        return "⚠️ Please provide both context and questions."
 
-    # Handle multiple questions line-by-line
+    # Split questions by line
     questions = [q.strip() for q in questions_text.strip().split("\n") if q.strip()]
     answers = []
 
     for i, q in enumerate(questions, start=1):
-        try:
-            result = qa_pipeline(question=q, context=context)
-            answer = result["answer"]
-            score = result["score"]
-            answers.append({
-                "question": q,
-                "answer": answer,
-                "score": round(score, 4)
-            })
-        except Exception as e:
-            answers.append({
-                "question": q,
-                "answer": f"Error: {str(e)}",
-                "score": 0
-            })
-
-    return {"results": answers}
+        result = qa_pipeline(question=q, context=context)
+        answer = result["answer"]
+        score = result["score"]
+        answers.append(f"Q{i}: {q}\nAnswer: {answer}\nConfidence: {score:.4f}\n{'-'*60}")
+
+    return "\n".join(answers)
+
+# Gradio Interface
+interface = gr.Interface(
+    fn=answer_multiple_questions,
+    inputs=[
+        gr.Textbox(
+            label="📘 Context (Paragraph/Text)",
+            lines=20,
+            placeholder="Enter or paste a long text context here..."
+        ),
+        gr.Textbox(
+            label="❓ Questions (one per line)",
+            lines=10,
+            placeholder="Type each question on a new line..."
+        ),
+    ],
+    outputs=gr.Textbox(
+        label="🧩 Answers",
+        lines=25,
+        placeholder="Model responses will appear here..."
+    ),
+    title="🤖 Machine Learning QA System",
+    description=(
+        "This Gradio app uses a fine-tuned RoBERTa model to answer questions from a given context.\n\n"
+        "👉 Enter a long paragraph in the context box.\n"
+        "👉 Write multiple questions (one per line).\n"
+        "👉 Get all answers organized line by line!"
+    ),
+    theme="soft",  # optional theme
+    allow_flagging="never"
+)
+
+# Launch app
+if __name__ == "__main__":
+    interface.launch()