Spaces:
Running
Running
import streamlit as st | |
import tempfile | |
import os | |
from pdf2markdown4llm import PDF2Markdown4LLM | |
st.set_page_config(page_title="PDF to Markdown Converter", layout="wide") | |
st.title("PDF to Markdown Converter") | |
st.write("Convert your PDF files to Markdown format") | |
def progress_callback(progress): | |
"""Callback function to handle progress updates""" | |
progress_bar.progress(progress.percentage / 100) | |
status_text.text(f"Phase: {progress.phase.value}, Page {progress.current_page}/{progress.total_pages}\n" | |
f"Progress: {progress.percentage:.1f}%, Message: {progress.message}") | |
def format_markdown_for_preview(markdown_text): | |
"""Format markdown text for proper preview display""" | |
# Ensure newlines are preserved by adding two spaces at the end of each line | |
lines = markdown_text.split('\n') | |
formatted_lines = [line + ' ' if line.strip() else line for line in lines] | |
return '\n'.join(formatted_lines) | |
# File upload | |
uploaded_file = st.file_uploader("Select a PDF file", type=['pdf']) | |
if uploaded_file is not None: | |
# Configuration options | |
with st.expander("Conversion Settings"): | |
remove_headers = st.checkbox("Remove Headers", value=False) | |
skip_empty_tables = st.checkbox("Skip Empty Tables", value=True) | |
table_header = st.text_input("Table Header Format", value="### Table") | |
if st.button("Start Conversion"): | |
# Initialize progress bar and status text | |
progress_bar = st.progress(0) | |
status_text = st.empty() | |
try: | |
# Create temporary file | |
with tempfile.NamedTemporaryFile(delete=False, suffix='.pdf') as tmp_file: | |
tmp_file.write(uploaded_file.getvalue()) | |
tmp_file_path = tmp_file.name | |
# Initialize converter | |
converter = PDF2Markdown4LLM( | |
remove_headers=remove_headers, | |
skip_empty_tables=skip_empty_tables, | |
table_header=table_header, | |
progress_callback=progress_callback | |
) | |
# Perform conversion | |
markdown_content = converter.convert(tmp_file_path) | |
# Clean up temporary file | |
os.unlink(tmp_file_path) | |
# Display results | |
st.success("Conversion completed successfully!") | |
# Raw markdown download | |
st.download_button( | |
label="Download Markdown File", | |
data=markdown_content, | |
file_name="converted.md", | |
mime="text/markdown" | |
) | |
# Preview with proper formatting | |
st.subheader("Preview") | |
# Create tabs for different preview modes | |
preview_tab, raw_tab = st.tabs(["Formatted Preview", "Raw Markdown"]) | |
with preview_tab: | |
formatted_content = format_markdown_for_preview(markdown_content) | |
st.markdown(formatted_content) | |
with raw_tab: | |
st.code(markdown_content, language="markdown") | |
except Exception as e: | |
st.error(f"An error occurred: {str(e)}") | |
if 'progress_bar' in locals(): | |
progress_bar.empty() | |
if 'status_text' in locals(): | |
status_text.empty() |