Create app.py

app.py

@@ -0,0 +1,71 @@
+import gradio as gr
+import fitz  # PyMuPDF
+from transformers import pipeline
+
+# Load models
+qa_pipeline = pipeline("question-answering", model="bert-large-uncased-whole-word-masking-finetuned-squad")
+summarizer = pipeline("summarization", model="t5-small", tokenizer="t5-small")  # Or use 'facebook/bart-large-cnn'
+
+# Extract text from uploaded PDF
+def extract_text(pdf_file):
+    text = ""
+    with fitz.open(pdf_file.name) as doc:
+        for page in doc:
+            text += page.get_text()
+    return text
+
+# QA pipeline function
+def qa_from_pdf_upload(pdf_file, question):
+    if not pdf_file:
+        return "❌ Please upload a PDF.", "", "", ""
+
+    context = extract_text(pdf_file)
+    if not context.strip():
+        return "❌ Could not extract text from the PDF.", "", "", ""
+
+    result = qa_pipeline(question=question, context=context)
+    return result["answer"], round(result["score"] * 100, 2), result["start"], result["end"]
+
+# Summarization function
+def summarize_pdf(pdf_file):
+    if not pdf_file:
+        return "❌ Please upload a PDF."
+
+    text = extract_text(pdf_file)
+    if not text.strip():
+        return "❌ Could not extract text from the PDF."
+
+    # Truncate text if too long for model
+    max_input_length = 1024
+    text = text[:max_input_length]
+
+    summary = summarizer(text, max_length=150, min_length=40, do_sample=False)[0]["summary_text"]
+    return summary
+
+# Gradio UI with Tabs
+with gr.Blocks(title="📘 Morningstar PDF Analyzer") as demo:
+    gr.Markdown("## 📘 Morningstar Fund PDF Analyzer\nUpload a PDF fund report and either ask questions or get a summary.")
+
+    with gr.Tabs():
+        with gr.TabItem("🔍 Q&A from PDF"):
+            pdf_input_qa = gr.File(label="📥 Upload Fund PDF")
+            question_input = gr.Textbox(label="❓ Your Question", placeholder="e.g., Who is the fund manager?")
+            answer_output = gr.Textbox(label="✅ Answer")
+            score_output = gr.Textbox(label="📊 Confidence Score (%)")
+            start_output = gr.Textbox(label="Start Index")
+            end_output = gr.Textbox(label="End Index")
+            qa_button = gr.Button("🧠 Get Answer")
+            qa_button.click(
+                fn=qa_from_pdf_upload,
+                inputs=[pdf_input_qa, question_input],
+                outputs=[answer_output, score_output, start_output, end_output]
+            )
+
+        with gr.TabItem("📝 Summary"):
+            pdf_input_sum = gr.File(label="📥 Upload Fund PDF")
+            summary_output = gr.Textbox(label="📝 Summary", lines=10)
+            sum_button = gr.Button("📄 Generate Summary")
+            sum_button.click(fn=summarize_pdf, inputs=[pdf_input_sum], outputs=[summary_output])
+
+# Launch the interface
+demo.launch()