Spaces:
Build error
Build error
Update app.py
Browse files
app.py
CHANGED
|
@@ -8,27 +8,35 @@ summarizer = pipeline("summarization", model="facebook/bart-large-cnn")
|
|
| 8 |
def pdf_to_text(pdf_file):
|
| 9 |
"""Extract text from a PDF file."""
|
| 10 |
text = ""
|
| 11 |
-
|
| 12 |
-
|
| 13 |
-
|
| 14 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 15 |
return text
|
| 16 |
|
| 17 |
def summarize_pdf(pdf_file):
|
| 18 |
"""Summarize the content of a PDF file."""
|
| 19 |
text = pdf_to_text(pdf_file)
|
| 20 |
-
if len(text) == 0:
|
| 21 |
return "No text found in the PDF."
|
| 22 |
-
|
| 23 |
-
#
|
| 24 |
-
|
| 25 |
-
|
|
|
|
|
|
|
|
|
|
| 26 |
|
| 27 |
# Create a Gradio interface
|
| 28 |
interface = gr.Interface(
|
| 29 |
fn=summarize_pdf,
|
| 30 |
inputs=gr.File(label="Upload a PDF file"),
|
| 31 |
-
outputs=gr.Textbox(label="Summary"),
|
| 32 |
title="PDF Summarizer",
|
| 33 |
description="Upload a PDF file to receive a summary."
|
| 34 |
)
|
|
|
|
| 8 |
def pdf_to_text(pdf_file):
|
| 9 |
"""Extract text from a PDF file."""
|
| 10 |
text = ""
|
| 11 |
+
try:
|
| 12 |
+
with open(pdf_file, 'rb') as file:
|
| 13 |
+
reader = PyPDF2.PdfReader(file)
|
| 14 |
+
for page in reader.pages:
|
| 15 |
+
page_text = page.extract_text()
|
| 16 |
+
if page_text: # Only add non-empty pages
|
| 17 |
+
text += page_text + "\n"
|
| 18 |
+
except Exception as e:
|
| 19 |
+
return f"Error reading PDF: {str(e)}"
|
| 20 |
return text
|
| 21 |
|
| 22 |
def summarize_pdf(pdf_file):
|
| 23 |
"""Summarize the content of a PDF file."""
|
| 24 |
text = pdf_to_text(pdf_file)
|
| 25 |
+
if len(text.strip()) == 0:
|
| 26 |
return "No text found in the PDF."
|
| 27 |
+
|
| 28 |
+
# Attempt to summarize the text
|
| 29 |
+
try:
|
| 30 |
+
summary = summarizer(text, max_length=130, min_length=30, do_sample=False)
|
| 31 |
+
return summary[0]['summary_text']
|
| 32 |
+
except Exception as e:
|
| 33 |
+
return f"Error summarizing text: {str(e)}"
|
| 34 |
|
| 35 |
# Create a Gradio interface
|
| 36 |
interface = gr.Interface(
|
| 37 |
fn=summarize_pdf,
|
| 38 |
inputs=gr.File(label="Upload a PDF file"),
|
| 39 |
+
outputs=gr.Textbox(label="Summary", lines=10),
|
| 40 |
title="PDF Summarizer",
|
| 41 |
description="Upload a PDF file to receive a summary."
|
| 42 |
)
|