Spaces:
Runtime error
Runtime error
Create app.py
Browse files
app.py
ADDED
@@ -0,0 +1,57 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
+
import gradio as gr
|
2 |
+
from pptx import Presentation
|
3 |
+
from pptx.util import Inches
|
4 |
+
import pdfplumber
|
5 |
+
from transformers import pipeline
|
6 |
+
import tempfile
|
7 |
+
|
8 |
+
# Initialize the summarization model
|
9 |
+
summarizer = pipeline("summarization", model="facebook/bart-large-cnn")
|
10 |
+
|
11 |
+
# Function to process PDF and create PowerPoint
|
12 |
+
def pdf_to_pptx(pdf_file):
|
13 |
+
# Step 1: Extract text from the PDF
|
14 |
+
pdf_text = ""
|
15 |
+
with pdfplumber.open(pdf_file) as pdf:
|
16 |
+
for page in pdf.pages:
|
17 |
+
pdf_text += page.extract_text()
|
18 |
+
|
19 |
+
# Step 2: Summarize text
|
20 |
+
def chunk_text(text, max_len=1024):
|
21 |
+
return [text[i:i+max_len] for i in range(0, len(text), max_len)]
|
22 |
+
|
23 |
+
chunked_text = chunk_text(pdf_text)
|
24 |
+
summarized_text = " ".join([summarizer(chunk)[0]['summary_text'] for chunk in chunked_text])
|
25 |
+
|
26 |
+
# Step 3: Create PowerPoint presentation
|
27 |
+
presentation = Presentation()
|
28 |
+
slide_content = summarized_text.split(". ")
|
29 |
+
|
30 |
+
for i, content in enumerate(slide_content):
|
31 |
+
slide = presentation.slides.add_slide(presentation.slide_layouts[1]) # Title and Content layout
|
32 |
+
title = slide.shapes.title
|
33 |
+
body = slide.placeholders[1]
|
34 |
+
title.text = f"Slide {i+1}" # Slide title
|
35 |
+
body.text = content # Slide content
|
36 |
+
|
37 |
+
# Step 4: Save presentation to a temporary file and return path
|
38 |
+
temp_pptx = tempfile.NamedTemporaryFile(delete=False, suffix=".pptx")
|
39 |
+
presentation.save(temp_pptx.name)
|
40 |
+
return temp_pptx.name
|
41 |
+
|
42 |
+
# Gradio app interface
|
43 |
+
def generate_pptx(pdf_file):
|
44 |
+
pptx_file_path = pdf_to_pptx(pdf_file.name)
|
45 |
+
return pptx_file_path
|
46 |
+
|
47 |
+
# Define Gradio interface
|
48 |
+
interface = gr.Interface(
|
49 |
+
fn=generate_pptx,
|
50 |
+
inputs=gr.File(label="Upload PDF File"),
|
51 |
+
outputs=gr.File(label="Download PowerPoint Presentation"),
|
52 |
+
title="PDF to PowerPoint Converter",
|
53 |
+
description="Upload a PDF file to convert it into a professional PowerPoint presentation."
|
54 |
+
)
|
55 |
+
|
56 |
+
# Launch the app
|
57 |
+
interface.launch()
|