Spaces:
Sleeping
Sleeping
File size: 2,931 Bytes
7dec78f |
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 |
import json
import gradio as gr
from pdfminer.high_level import extract_pages, extract_text
from pdfminer.layout import LTTextBoxHorizontal, LTFigure, LTImage
import mistletoe # for Markdown table generation (optional)
def parse_pdf(pdf_file, output_format):
with open(pdf_file, 'rb') as file:
pages = extract_pages(file)
text = ""
tables = [] # Placeholder for extracted table data
images = [] # Placeholder for extracted image data
for page in pages:
for element in page:
if isinstance(element, LTTextBoxHorizontal):
text += element.get_text()
elif isinstance(element, LTFigure):
# Extract image data (e.g., save as image, convert to base64)
images.append(element)
elif isinstance(element, LTImage):
# Extract image data (e.g., save as image, convert to base64)
images.append(element)
# Implement table extraction logic (e.g., using heuristics or advanced techniques)
# You can use libraries like Camelot for complex tables
# ...
# Convert extracted data to desired format and populate download_data
if output_format == "JSON":
json_data = {
"text": text,
"tables": tables, # Replace with actual table data
"images": images # Replace with actual image data (e.g., base64)
}
download_data = json.dumps(json_data).encode("utf-8") # Encode JSON for download
elif output_format == "Markdown":
# Implement table conversion using mistletoe or other Markdown libraries
markdown_tables = mistletoe.markdown(convert_table=True)(tables) # Example using mistletoe
markdown_text = f"# Extracted Text\n\n{text}\n\n{markdown_tables}\n\n# Images\n"
# Implement image conversion (e.g., relative paths or base64 encoding)
# ...
download_data = markdown_text.encode("utf-8")
elif output_format == "HTML":
# Implement table conversion using HTML table tags
html_tables = "<table>" # Start of HTML table (replace with actual table structure)
# ... (Implement table data conversion to HTML)
html_tables += "</table>"
html_text = f"<p>{text}</p>\n\n<h2>Tables</h2>\n{html_tables}\n\n<h2>Images</h2>\n"
# Implement image conversion using `<img>` tag
# ...
download_data = html_text.encode("utf-8")
return text, download_data
iface = gr.Interface(
fn=parse_pdf,
inputs=["file", gr.Dropdown(["JSON", "Markdown", "HTML"])],
outputs=[
gr.Text(label="Output Text"),
gr.File(label="Download Output")
],
title="PDF Parser",
description="Parse a PDF and choose the output format."
)
if __name__ == "__main__":
iface.launch(share=True) # Set share=True to create a public link |