#!/usr/bin/env python3
"""Verify offline bundles and their exact, unabridged authored-source inventory."""
from __future__ import annotations

import hashlib
import json
import re
from html.parser import HTMLParser
from pathlib import Path

from build_labs import ROOT, build_page, source_paths


class ConfigParser(HTMLParser):
    def __init__(self) -> None:
        super().__init__(convert_charrefs=False)
        self.collect = False
        self.chunks: list[str] = []
        self.count = 0

    def handle_starttag(self, tag: str, attrs: list[tuple[str, str | None]]) -> None:
        if tag == 'script' and dict(attrs).get('id') == 'lab-config':
            self.collect = True
            self.count += 1

    def handle_endtag(self, tag: str) -> None:
        if tag == 'script':
            self.collect = False

    def handle_data(self, data: str) -> None:
        if self.collect:
            self.chunks.append(data)


def verify_lab(lab: Path, page: str | None = None) -> set[str]:
    data = json.loads((lab / 'lab.json').read_text())
    page = (lab / 'index.html').read_text() if page is None else page
    parser = ConfigParser()
    parser.feed(page)
    assert parser.count == 1, 'expected one inert source/config manifest'
    config = json.loads(''.join(parser.chunks))
    assert config['spec'] == (lab / 'component.md').read_bytes().decode('utf-8')
    expected_paths = [path.relative_to(ROOT.parent).as_posix() for path in source_paths(lab)]
    assert [s['path'] for s in config['sources']] == expected_paths, 'omitted or reordered source'
    for source, path in zip(config['sources'], source_paths(lab), strict=True):
        raw = path.read_bytes()
        assert source['content'].encode('utf-8') == raw, f'abridged source: {path}'
        assert source['sha256'] == hashlib.sha256(raw).hexdigest(), f'source digest drift: {path}'
    assert data['diagram_kind'] == 'illustrative-lesson-flow'
    project = json.loads((ROOT.parent / 'literate.project.json').read_text())
    assert config['harness']['project'] == project['project_id']
    assert config['harness']['selected_flavors'] == []
    assert config['harness']['documentation_root'] is True
    assert config['harness']['component_root'] is False
    assert 'Surrounding inputs & harness' in page
    assert config['nodes'] == data['nodes'] and config['edges'] == data['edges']
    ids = [n['id'] for n in data['nodes']]
    assert len(set(ids)) == len(ids)
    assert all(a in ids and b in ids for a, b in data['edges'])
    assert all('Flavor' not in n['label'] and n['label'] != 'Component' for n in data['nodes'])
    assert 'Lesson flow (illustrative)' in page
    assert 'No resolved Component DAG is supplied.' in page
    assert 'Readable lab specification' in page and 'Exact bundled source' in page
    # An empty data-URL favicon prevents the browser's implicit network request.
    # Every other linked resource remains forbidden for these standalone labs.
    resources = page.replace('<link rel="icon" href="data:,">', '')
    assert '<script src=' not in resources and '<link ' not in resources
    assert not re.search(r'<(?:script|link|img|iframe)[^>]+(?:src|href)=[\"\']https?://', page)
    assert not re.search(r'\{(?:title|css|runtime|main|config)\}', page.split('<script>')[0])
    assert page == build_page(lab), 'stale or modified bundled executable/style/wrapper'
    return set(data['media'])


def main() -> None:
    labs = sorted(ROOT.glob('[0-9][0-9]-*'))
    assert len(labs) == 6
    media: set[str] = set()
    for expected, lab in enumerate(labs):
        assert int(json.loads((lab / 'lab.json').read_text())['level']) == expected
        media.update(verify_lab(lab))
    required = {'math', 'physics', 'audio', 'music', '3d', 'video', 'html'}
    assert required <= media, f'missing media coverage: {sorted(required - media)}'
    print(f'verified {len(labs)} offline lab bundles, illustrative diagrams and exact source inventories')


if __name__ == '__main__':
    main()