Spaces:
Running
Running
Create app.py
Browse files
app.py
ADDED
|
@@ -0,0 +1,69 @@
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 1 |
+
import streamlit as st
|
| 2 |
+
import tempfile
|
| 3 |
+
import os
|
| 4 |
+
from pdf2markdown4llm import PDF2Markdown4LLM
|
| 5 |
+
|
| 6 |
+
st.set_page_config(page_title="PDF to Markdown Converter", layout="wide")
|
| 7 |
+
|
| 8 |
+
st.title("PDF to Markdown Converter")
|
| 9 |
+
st.write("Convert your PDF files to Markdown format")
|
| 10 |
+
|
| 11 |
+
def progress_callback(progress):
|
| 12 |
+
"""Callback function to handle progress updates"""
|
| 13 |
+
progress_bar.progress(progress.percentage / 100)
|
| 14 |
+
status_text.text(f"Phase: {progress.phase.value}, Page {progress.current_page}/{progress.total_pages}\n"
|
| 15 |
+
f"Progress: {progress.percentage:.1f}%, Message: {progress.message}")
|
| 16 |
+
|
| 17 |
+
# File upload
|
| 18 |
+
uploaded_file = st.file_uploader("Select a PDF file", type=['pdf'])
|
| 19 |
+
|
| 20 |
+
if uploaded_file is not None:
|
| 21 |
+
# Configuration options
|
| 22 |
+
with st.expander("Conversion Settings"):
|
| 23 |
+
remove_headers = st.checkbox("Remove Headers", value=False)
|
| 24 |
+
skip_empty_tables = st.checkbox("Skip Empty Tables", value=True)
|
| 25 |
+
table_header = st.text_input("Table Header Format", value="### Table")
|
| 26 |
+
|
| 27 |
+
if st.button("Start Conversion"):
|
| 28 |
+
# Initialize progress bar and status text
|
| 29 |
+
progress_bar = st.progress(0)
|
| 30 |
+
status_text = st.empty()
|
| 31 |
+
|
| 32 |
+
try:
|
| 33 |
+
# Create temporary file
|
| 34 |
+
with tempfile.NamedTemporaryFile(delete=False, suffix='.pdf') as tmp_file:
|
| 35 |
+
tmp_file.write(uploaded_file.getvalue())
|
| 36 |
+
tmp_file_path = tmp_file.name
|
| 37 |
+
|
| 38 |
+
# Initialize converter
|
| 39 |
+
converter = PDF2Markdown4LLM(
|
| 40 |
+
remove_headers=remove_headers,
|
| 41 |
+
skip_empty_tables=skip_empty_tables,
|
| 42 |
+
table_header=table_header,
|
| 43 |
+
progress_callback=progress_callback
|
| 44 |
+
)
|
| 45 |
+
|
| 46 |
+
# Perform conversion
|
| 47 |
+
markdown_content = converter.convert(tmp_file_path)
|
| 48 |
+
|
| 49 |
+
# Clean up temporary file
|
| 50 |
+
os.unlink(tmp_file_path)
|
| 51 |
+
|
| 52 |
+
# Display results
|
| 53 |
+
st.success("Conversion completed successfully!")
|
| 54 |
+
st.download_button(
|
| 55 |
+
label="Download Markdown File",
|
| 56 |
+
data=markdown_content,
|
| 57 |
+
file_name="converted.md",
|
| 58 |
+
mime="text/markdown"
|
| 59 |
+
)
|
| 60 |
+
|
| 61 |
+
st.subheader("Preview")
|
| 62 |
+
st.markdown(markdown_content)
|
| 63 |
+
|
| 64 |
+
except Exception as e:
|
| 65 |
+
st.error(f"An error occurred: {str(e)}")
|
| 66 |
+
if 'progress_bar' in locals():
|
| 67 |
+
progress_bar.empty()
|
| 68 |
+
if 'status_text' in locals():
|
| 69 |
+
status_text.empty()
|