Spaces:
Running
Running
Delete app.py
Browse files
app.py
DELETED
@@ -1,88 +0,0 @@
|
|
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 PyPDF2
|
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 |
-
|
18 |
-
# Ensure API key is available
|
19 |
-
if not GROQ_API_KEY:
|
20 |
-
raise ValueError("GROQ_API_KEY is not set. Add it in Hugging Face Secrets.")
|
21 |
-
|
22 |
-
# Initialize LLM (Mistral-8x7B)
|
23 |
-
llm = ChatGroq(model_name="mixtral-8x7b-32768")
|
24 |
-
|
25 |
-
def extract_text_from_pdf(pdf_file):
|
26 |
-
"""Extract text from a PDF file."""
|
27 |
-
text = ""
|
28 |
-
reader = PyPDF2.PdfReader(pdf_file)
|
29 |
-
for page in reader.pages:
|
30 |
-
page_text = page.extract_text()
|
31 |
-
if page_text:
|
32 |
-
text += page_text + "\n"
|
33 |
-
return text
|
34 |
-
|
35 |
-
def summarize_text(text, length='Medium', style='Concise Paragraph'):
|
36 |
-
"""Summarize the text with CoT and adjustable format."""
|
37 |
-
|
38 |
-
# Adjust summary length
|
39 |
-
length_map = {
|
40 |
-
'Short': 'Summarize in 3-4 lines.',
|
41 |
-
'Medium': 'Summarize in 6-8 lines.',
|
42 |
-
'Long': 'Provide a detailed summary in multiple paragraphs.'
|
43 |
-
}
|
44 |
-
|
45 |
-
# Adjust summary style
|
46 |
-
style_map = {
|
47 |
-
'Bulleted List': 'Format the summary as a list of key points.',
|
48 |
-
'Key Takeaways': 'Extract the most important insights as key takeaways.',
|
49 |
-
'Concise Paragraph': 'Write the summary as a structured paragraph.'
|
50 |
-
}
|
51 |
-
|
52 |
-
prompt = f"""
|
53 |
-
Step 1: Identify the main topics covered in the document.
|
54 |
-
Step 2: Extract key facts, arguments, and conclusions.
|
55 |
-
Step 3: Generate a structured summary based on extracted information.
|
56 |
-
{length_map[length]} {style_map[style]}
|
57 |
-
|
58 |
-
Document:
|
59 |
-
{text[:10000]}
|
60 |
-
"""
|
61 |
-
|
62 |
-
response = llm.predict(prompt)
|
63 |
-
return response
|
64 |
-
|
65 |
-
def process_pdf(file, length, style):
|
66 |
-
"""Extract text and summarize PDF using Mistral-8x7B with customization."""
|
67 |
-
if file is None:
|
68 |
-
return "No file uploaded."
|
69 |
-
|
70 |
-
text = extract_text_from_pdf(file)
|
71 |
-
summary = summarize_text(text, length, style)
|
72 |
-
return summary
|
73 |
-
|
74 |
-
# Create Gradio Interface
|
75 |
-
interface = gr.Interface(
|
76 |
-
fn=process_pdf,
|
77 |
-
inputs=[
|
78 |
-
gr.File(label="Upload a PDF"),
|
79 |
-
gr.Radio(["Short", "Medium", "Long"], value="Medium", label="Summary Length"),
|
80 |
-
gr.Radio(["Bulleted List", "Key Takeaways", "Concise Paragraph"], value="Concise Paragraph", label="Summary Style")
|
81 |
-
],
|
82 |
-
outputs="text",
|
83 |
-
title="📄 AI-Powered PDF Summarizer",
|
84 |
-
description="Upload a PDF file and customize the summary format and length."
|
85 |
-
)
|
86 |
-
|
87 |
-
# Run the app
|
88 |
-
interface.launch()
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|