Spaces:
Running
Running
Upload app.py
Browse files
app.py
ADDED
|
@@ -0,0 +1,75 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
# -*- coding: utf-8 -*-
|
| 2 |
+
"""app.ipynb
|
| 3 |
+
|
| 4 |
+
Automatically generated by Colab.
|
| 5 |
+
|
| 6 |
+
Original file is located at
|
| 7 |
+
https://colab.research.google.com/drive/1XblbxoRxB4XOHixjGij789FPD9KjKdhi
|
| 8 |
+
"""
|
| 9 |
+
|
| 10 |
+
import os
|
| 11 |
+
import pdfplumber
|
| 12 |
+
import gradio as gr
|
| 13 |
+
from langchain_groq.chat_models import ChatGroq
|
| 14 |
+
|
| 15 |
+
# Set Groq API key securely
|
| 16 |
+
GROQ_API_KEY = os.getenv("GROQ_API_KEY") # Fetch from environment variables
|
| 17 |
+
if not GROQ_API_KEY:
|
| 18 |
+
raise ValueError("GROQ_API_KEY is not set. Add it in Hugging Face Secrets.")
|
| 19 |
+
|
| 20 |
+
# Initialize LLM (Mistral-8x7B)
|
| 21 |
+
llm = ChatGroq(model_name="mixtral-8x7b-32768")
|
| 22 |
+
|
| 23 |
+
def extract_text_from_pdf(pdf_file):
|
| 24 |
+
"""Extract text from a PDF file using pdfplumber."""
|
| 25 |
+
text = ""
|
| 26 |
+
with pdfplumber.open(pdf_file) as pdf:
|
| 27 |
+
for page in pdf.pages:
|
| 28 |
+
text += page.extract_text() or "" # Extract text safely
|
| 29 |
+
return text.strip()
|
| 30 |
+
|
| 31 |
+
def summarize_text(text, length, style):
|
| 32 |
+
"""Summarize text with adjustable length and style."""
|
| 33 |
+
prompt = (
|
| 34 |
+
f"""
|
| 35 |
+
Read the following document and summarize it in {style.lower()} format.
|
| 36 |
+
Keep the summary {length.lower()}.
|
| 37 |
+
Follow this step-by-step approach:
|
| 38 |
+
1. Identify key sections.
|
| 39 |
+
2. Extract essential points.
|
| 40 |
+
3. Remove unnecessary details.
|
| 41 |
+
4. Ensure factual accuracy without adding extra information.
|
| 42 |
+
|
| 43 |
+
Document:
|
| 44 |
+
{text[:10000]} # Limit input to 10,000 characters for efficiency
|
| 45 |
+
"""
|
| 46 |
+
)
|
| 47 |
+
response = llm.predict(prompt)
|
| 48 |
+
return response.strip()
|
| 49 |
+
|
| 50 |
+
def process_pdf(file, length, style):
|
| 51 |
+
"""Extract text and summarize PDF using Mistral-8x7B with customization."""
|
| 52 |
+
if not file:
|
| 53 |
+
return "No file uploaded."
|
| 54 |
+
|
| 55 |
+
text = extract_text_from_pdf(file.name) # Extract text
|
| 56 |
+
if not text:
|
| 57 |
+
return "Could not extract text from the PDF. Try another file."
|
| 58 |
+
|
| 59 |
+
return summarize_text(text, length, style)
|
| 60 |
+
|
| 61 |
+
# Create Gradio Interface
|
| 62 |
+
interface = gr.Interface(
|
| 63 |
+
fn=process_pdf,
|
| 64 |
+
inputs=[
|
| 65 |
+
gr.File(label="Upload a PDF"),
|
| 66 |
+
gr.Radio(["Short", "Medium", "Long"], label="Summary Length", value="Medium"),
|
| 67 |
+
gr.Radio(["Bullets", "Key Takeaways", "Concise Paragraph"], label="Summary Style", value="Key Takeaways"),
|
| 68 |
+
],
|
| 69 |
+
outputs="text",
|
| 70 |
+
title="📄 Advanced PDF Summarizer",
|
| 71 |
+
description="Upload a PDF file and get a structured summary with customization options."
|
| 72 |
+
)
|
| 73 |
+
|
| 74 |
+
# Run the app
|
| 75 |
+
interface.launch()
|