ODT to custom md

gistfile1.py

import zipfile, subprocess, re, sys
from pathlib import Path

if len(sys.argv) < 2:
    print("Usage: python convert.py <input.odt>")
    sys.exit(1)

input_path = sys.argv[1]

if not Path(input_path).exists():
    print(f"Error: file not found: {input_path}")
    sys.exit(1)

if Path(input_path).suffix.lower() != '.odt':
    print(f"Error: expected .odt file, got: {Path(input_path).suffix}")
    sys.exit(1)

patched = '/tmp/input_patched.odt'

print(f"Reading {input_path}...")
with zipfile.ZipFile(input_path) as z:
    content = z.read('content.xml').decode()

empty_paras = len(re.findall(r'<text:p[^>]*/>', content))
print(f" -> Found {empty_paras} empty paragraph(s), patching...")
content = re.sub(
    r'<text:p[^>]*/>', 
    '<text:p text:style-name="Text_20_Body">EMPTY_PARA_PLACEHOLDER</text:p>',
    content
)

print(f"Writing patched ODT to {patched}...")
with zipfile.ZipFile(patched, 'w') as zout:
    with zipfile.ZipFile(input_path) as zin:
        for item in zin.infolist():
            if item.filename == 'content.xml':
                zout.writestr(item, content)
            else:
                zout.writestr(item, zin.read(item.filename))

print(f"Running pandoc on {patched}...")
print(f" -> pandoc {patched} -t markdown-smart --wrap=none")
result = subprocess.run(
    ['pandoc', patched, '-t', 'markdown-smart', '--wrap=none'],
    capture_output=True, text=True
)

if result.returncode != 0:
    print(f"Error: pandoc failed:\n{result.stderr}")
    sys.exit(1)

print(" -> removing patched markers")
output = result.stdout.replace('EMPTY_PARA_PLACEHOLDER\n', '')


if 'EOFEOFEOF' in output:
    output = output[:output.index('EOFEOFEOF')]
    print(" -> Truncated output at eof marker")

print("Outputting")

output_dir = Path('md-export')
print(f" -> Creating output folder {output_dir.resolve()}")
output_dir.mkdir(exist_ok=True)
output_path = output_dir / Path(input_path).with_suffix('.md').name

print(f" -> Writing output to {output_path}...")
with open(output_path, 'w') as f:
    f.write(output)

print(f"Done. {len(output.splitlines())} lines written to {output_path}")