absolutize.py

#!/usr/bin/env python3

import argparse
import fileinput
import re
import sys
from pathlib import Path
from urllib.parse import urlparse, urlunparse

parser = argparse.ArgumentParser(description='Make relative links in markup file absolute with given base-url.')
parser.add_argument('base_url', type=str, help='the base-url used to make relative links absolute to')
parser.add_argument('md_file', type=str, nargs='+', help='path to markup file')
args = parser.parse_args()

base_url = urlparse(args.base_url)

allowed_schemes = ['http', 'https']
if base_url.scheme not in allowed_schemes:
    sys.exit('Please use a base_url the following kind: ' + ', '.join(allowed_schemes))

base_url_path = Path(base_url.path)

link_pattern = re.compile(r'\[.*?\]\((.+?)\)')
nop_path = Path('.')

report = []

for md_file in args.md_file:
    with fileinput.FileInput(md_file, inplace=True) as file:
        for line in file:
            def replace(match):
                try:
                    matchstr = match.group(0)
                    link = match.group(1)
                    url = urlparse(link)
                    if not url.hostname and not url.scheme and not url.fragment:
                        rel_path = Path(url.path)
                        if not rel_path.is_absolute() and rel_path != nop_path and not rel_path.exists():
                            scheme, netloc, _, params, query, fragment = base_url
                            path = str((base_url_path / rel_path).resolve(strict=False))
                            fixed_link = urlunparse((scheme, netloc, path, params, query, fragment))
                            report.append((md_file, f'{matchstr} ==> {fixed_link}'))
                            return matchstr.replace(link, fixed_link)
                except Exception as ex:
                    report.append((md_file, ex))

                return matchstr


            fixed_line = link_pattern.sub(replace, line.rstrip('\n'))
            print(fixed_line)

for md_file, info in report:
    sink = sys.stderr if isinstance(info, Exception) else sys.stdout
    print(f'{md_file}: {info}', file=sink)