#!/usr/bin/env python3
"""汇总 lighthouse 五页报告：分数表 + 未达标项的文件级嫌疑清单。

输入：目录内 page-*.json（lighthouse 全量输出）
输出：/tmp/e2e-shots/lighthouse-report.json（全量合并）+ stdout 摘要表
"""
import json
import pathlib
import sys

# 只挑与"可执行优化"相关的审计做嫌疑展开；其余仅计分
SUSPECT_AUDITS = (
    'renderblocking-resources', 'unused-javascript', 'unused-css-rules',
    'modern-image-formats', 'uses-optimized-images', 'total-byte-weight',
    'layout-shift-elements', 'layout-shifts', 'largest-contentful-paint-element',
    'color-contrast', 'font-display', 'mainthread-work-breakdown', 'bootup-time',
    'dom-size', 'offscreen-images', 'uses-responsive-images', 'unminified-css',
    'unminified-javascript', 'efficient-animated-content', 'duplicated-javascript',
    'legacy-javascript', 'prioritize-lcp-image', 'long-tasks',
)
CATEGORY_LABEL = {'performance': 'Perf', 'accessibility': 'A11y', 'best-practices': 'BP', 'seo': 'SEO'}
METRICS = ('first-contentful-paint', 'largest-contentful-paint', 'total-blocking-time',
           'cumulative-layout-shift', 'speed-index', 'interactive')


def fmt_ms(v):
    return f"{v / 1000:.1f}s" if isinstance(v, (int, float)) else str(v)


def main(outdir: pathlib.Path):
    pages = {}
    rows = []
    for f in sorted(outdir.glob('page-*.json')):
        data = json.loads(f.read_text())
        url = data.get('finalDisplayedUrl') or data.get('finalUrl') or f.stem
        path = '/' + url.split('/', 3)[-1].replace('127.0.0.1:18743', '') if url.count('/') >= 3 else url
        cats = {k: round(v['score'] * 100) for k, v in data['categories'].items() if v.get('score') is not None}
        metrics = {m: fmt_ms(data['audits'][m].get('numericValue')) for m in METRICS if m in data['audits']}
        suspects = {}
        for cid, cat in data['categories'].items():
            if cats.get(cid, 100) >= 100:
                continue
            for ref in cat.get('auditRefs', []):
                aid = ref['id']
                a = data['audits'].get(aid, {})
                if a.get('score') is not None and a['score'] < 0.9 and aid in SUSPECT_AUDITS:
                    items = (a.get('details', {}) or {}).get('items', [])
                    hint = []
                    for it in items[:4]:
                        for key in ('url', 'node.selector', 'node.path', 'source.url'):
                            v = it
                            for seg in key.split('.'):
                                v = v.get(seg) if isinstance(v, dict) else None
                            if v:
                                hint.append(str(v).split('?')[0][:110])
                                break
                        else:
                            hint.append(str(it)[:90])
                    if hint:
                        suspects[aid] = {'score': a['score'], 'items': hint}
        pages[path or f.stem] = {'scores': cats, 'metrics': metrics, 'suspects': suspects}
        rows.append((path, cats))

    labels = [CATEGORY_LABEL[c] for c in ('performance', 'accessibility', 'best-practices', 'seo')]
    print(f"{'页面':<28}" + ''.join(f"{l:>6}" for l in labels))
    for path, cats in rows:
        cells = ''.join(f"{str(cats.get(c, "-"))[:3]:>6}" for c in ('performance', 'accessibility', 'best-practices', 'seo'))
        mark = ''.join(' ✗' if cats.get(c, 100) < 90 else ' ✓' for c in ('performance', 'accessibility', 'best-practices', 'seo'))
        print(f"{path:<28}{cells}   {mark}")

    dest = pathlib.Path('/tmp/e2e-shots/lighthouse-report.json')
    dest.write_text(json.dumps(pages, ensure_ascii=False, indent=1))
    print(f"\n全量合并报告: {dest}（{dest.stat().st_size} B）")


if __name__ == '__main__':
    main(pathlib.Path(sys.argv[1]))