#!/usr/bin/env python3
"""Validate the authoritative Gitea wiki without generating or publishing files."""
from html import unescape
from pathlib import Path
import re
import sys
from urllib.parse import unquote, urlsplit

PROJECT = Path(__file__).resolve().parents[2]
WIKI = PROJECT / 'fds-os.wiki'
SOURCE_URL = 'http://gitea.home.arpa/felis/fds-os/src/branch/main/'
WIKI_URL = 'http://gitea.home.arpa/felis/fds-os/wiki/'
FENCE = re.compile(r'^ {0,3}(`{3,}|~{3,})(.*)$')
LINK = re.compile(r'!?\[[^\]]*\]\((<[^>\n]+>|[^\s)]+)(?:\s+["\'][^\n]*?["\'])?\)')
REFERENCE = re.compile(r'^ {0,3}\[[^\]\n]+\]:[ \t]*(<[^>\n]+>|\S+)', re.M)


def prose(text):
    lines, fence = [], None
    for line in text.splitlines(keepends=True):
        found = FENCE.match(line)
        if fence:
            if found and found[1][0] == fence[0] and len(found[1]) >= len(fence) and not found[2].strip():
                fence = None
            continue
        if found:
            fence = found[1]
        elif not line.startswith(('    ', '\t')):
            lines.append(line)
    return ''.join(lines)


def anchors(text):
    result, counts = set(), {}
    for heading in re.findall(r'^#{1,6}\s+(.+?)\s*#*$', prose(text), re.M):
        heading = re.sub(r'\[([^\]]+)\]\([^)]*\)', r'\1', heading)
        heading = re.sub(r'<[^>]*>', '', heading)
        slug = re.sub(r'[^\w\- ]', '', unescape(heading).lower()).replace(' ', '-')
        count = counts.get(slug, 0)
        counts[slug] = count + 1
        result.add(slug + (f'-{count}' if count else ''))
    return result


def links(text):
    text = re.sub(r'(`+)(?!`)[^\n]*?(?<!`)\1(?!`)', '', prose(text))
    return [m[1].strip('<>') for pattern in (LINK, REFERENCE) for m in pattern.finditer(text)]


def main():
    if not (WIKI / 'Home.md').is_file():
        sys.exit('ERROR: missing wiki; run git submodule update --init fds-os.wiki')
    pages = {p.stem: p for p in WIKI.glob('*.md')}
    errors = []
    require = lambda condition, message: errors.append(message) if not condition else None
    require(len({name.casefold() for name in pages}) == len(pages), 'Page names collide ignoring case')
    for name in ('Home', 'Overview', 'Page-Index', '_Sidebar', '_Footer'):
        require(name in pages, f'Missing required page: {name}.md')
    for name, page in pages.items():
        require(bool(re.fullmatch(r'(?:[A-Z][A-Za-z0-9]*(?:-[A-Za-z0-9]+)*|_Sidebar|_Footer)', name)),
                f'Invalid page name: {page.name}')
        if not name.startswith('_'):
            require(page.read_text().startswith('---\ninclude_toc: true\n---\n'),
                    f'{page.name}: missing table-of-contents frontmatter')
    page_anchors = {name: anchors(page.read_text()) for name, page in pages.items()}
    index_links = set(links(pages['Page-Index'].read_text())) if 'Page-Index' in pages else set()
    for name in pages.keys() - {'Page-Index', '_Sidebar', '_Footer'}:
        require(name in index_links, f'{name}.md is missing from Page-Index.md')
    checked = 0
    for page in [*pages.values(), PROJECT / 'README.md']:
        for target in links(page.read_text()):
            url = urlsplit(target)
            if target.startswith(SOURCE_URL):
                relative = unquote(url.path.removeprefix(urlsplit(SOURCE_URL).path))
                path = (PROJECT / relative).resolve()
                require(path.is_relative_to(PROJECT) and path.is_file(),
                        f'{page.name}: missing source file: {target}')
                checked += 1
                continue
            if target.startswith(WIKI_URL):
                url = urlsplit(target[len(WIKI_URL):])
            elif url.scheme or url.netloc:
                continue
            name = unquote(url.path)
            if name:
                require('/' not in name and name != '..', f'{page.name}: non-flat wiki link: {target}')
                if Path(name).suffix:
                    require(Path(name).suffix != '.md', f'{page.name}: wiki page links must omit .md: {target}')
                    require((WIKI / name).is_file(), f'{page.name}: missing asset: {target}')
                    checked += 1
                    continue
            else:
                name = page.stem
            require(name in pages, f'{page.name}: missing wiki page: {target}')
            if url.fragment:
                require(unquote(url.fragment) in page_anchors.get(name, set()),
                        f'{page.name}: missing section: {target}')
            checked += 1
    if errors:
        sys.exit('\n'.join('ERROR: ' + message for message in errors))
    print(f'PASS: {len(pages)} wiki pages; {checked} page, section, asset and source links')


if __name__ == '__main__':
    main()