Script for fixing PDF image references in exported logseq graph.

loqseq_pdf_fix.py

#!/usr/bin/env python3

from re import sub, Match
from pathlib import Path
from typer import Typer
from functools import partial
from itertools import chain
from typing import List
from sh import cp


main = Typer(add_completion=False)

PATTERN = '\(\([0-9a-z\-]+\)\)'
# matches things like ((62a6a8e3-1b2c-42de-aaab-0b495459b61f))


def fixer(assets : List[Path], match : Match):

    # Removes double parenthesis:
    ref_id = match.group(0)[2:-2]
    # Takes first matching filename:
    asset = next(chain(filter(lambda a: ref_id in a.name, assets), [None]))
    
    if asset:
        # Copy asset to public assets folder:
        cp(asset, Path('./assets/'))

        new_path = Path('../assets') / asset.name

    # If no asset was found, we return the same string:
    token = f'![{asset.name}]({new_path})' if asset else match.group(0)

    print(f'* {ref_id = }\n* {asset = }\n* {token = }\n')
    return token


@main.command()
def _main(assets_dir : Path):

    try:
        html = open('./index.html').read()
    except FileNotFoundError:
        print('You must run this in the exported folder.')
        raise SystemExit(1)

    assets = list(assets_dir.glob('*.png'))

    fixed_html = sub(PATTERN, partial(fixer, assets), html)
    open('./index.html', 'w').write(fixed_html)


'main' in __name__ and main()