Python script to read RPM tags.md file and print KSY enum definition of the tags

rpm_read_tags_md.py

import re
from collections import OrderedDict

tag_regex = r"\* (\w+) \((\d+)\)"
heading_regex = r"^## (.+)"
sections = {}

# https://github.com/rpm-software-management/rpm/raw/911448f2/doc/manual/tags.md
with open('tags.md', 'r', encoding='utf-8') as f:
	contents = f.read()

matches = list(re.finditer(heading_regex, contents, re.MULTILINE))
for i, m in enumerate(matches):
	(heading,) = m.groups()
	body_start = m.end()
	body_end = matches[i + 1].start() if i + 1 < len(matches) else None
	sections[heading] = contents[body_start:body_end]

tags = {}

for heading, body in sections.items():
	if heading == 'Aliases':
		continue
	is_obsolete = bool(re.match(r"deprecated|obsolete", heading, re.IGNORECASE))
	matches = re.finditer(tag_regex, body)
	for m in matches:
		(tag_name, tag_value) = m.groups()
		tags[int(tag_value)] = (tag_name, is_obsolete)

tags_ordered = OrderedDict(sorted(tags.items()))
enum_str = ''.join([(f'    # {k}: {v.lower()}\n' if is_obsolete else f"""    {k}:
      id: {v.lower()}
      -orig-id: RPMTAG_{v.upper()}
""") for k, (v, is_obsolete) in tags_ordered.items()])
print(enum_str)