chore: update mirrored site snapshots and action capture outputs

scripts/build_deep_action_report.py

@@ -0,0 +1,131 @@
+#!/usr/bin/env python3
+from __future__ import annotations
+
+import json
+from collections import Counter, defaultdict
+from pathlib import Path
+
+
+ROOT = Path(__file__).resolve().parents[1]
+MANIFEST_PATH = ROOT / "deep-action-manifest-hc-pos.sqygj.cn.json"
+MIRROR_MANIFEST_PATH = ROOT / "mirror-manifest-hc-pos.sqygj.cn.json"
+REPORT_PATH = ROOT / "deep-action-report.md"
+SUMMARY_PATH = ROOT / "deep-action-summary.json"
+GAPS_PATH = ROOT / "deep-action-gaps.txt"
+
+
+def classify(label: str) -> str:
+    pairs = [
+        ("统计", "统计"),
+        ("详情", "详情"),
+        ("查看", "查看"),
+        ("编辑", "编辑"),
+        ("新增", "新增"),
+        ("新建", "新建"),
+        ("创建", "创建"),
+        ("添加", "添加"),
+        ("记录", "记录"),
+        ("工单", "工单"),
+        ("复制", "复制"),
+        ("绑定", "绑定"),
+        ("确认", "确认"),
+    ]
+    for needle, category in pairs:
+        if needle in label:
+            return category
+    return "其他"
+
+
+def route_key(url: str) -> str:
+    return url.split("#/", 1)[1] if "#/" in url else url
+
+
+def main() -> int:
+    manifest = json.loads(MANIFEST_PATH.read_text(encoding="utf-8"))
+    items = manifest["items"]
+    mirror_manifest = json.loads(MIRROR_MANIFEST_PATH.read_text(encoding="utf-8"))
+
+    by_route: dict[str, list[dict]] = defaultdict(list)
+    for item in items:
+        by_route[item["source_url"]].append(item)
+
+    route_rows = []
+    category_counter: Counter[str] = Counter()
+    for source_url, route_items in sorted(by_route.items(), key=lambda kv: route_key(kv[0])):
+        labels = [item["label"] for item in route_items]
+        categories = Counter(classify(label) for label in labels)
+        category_counter.update(categories)
+        route_rows.append(
+            {
+                "source_url": source_url,
+                "route": route_key(source_url),
+                "count": len(route_items),
+                "categories": dict(sorted(categories.items())),
+                "labels": labels,
+                "html_paths": [item["html_path"] for item in route_items],
+            }
+        )
+
+    summary = {
+        "captured_deep_states": manifest["captured_deep_states"],
+        "route_count": len(route_rows),
+        "category_counts": dict(sorted(category_counter.items())),
+        "routes": route_rows,
+    }
+    SUMMARY_PATH.write_text(json.dumps(summary, ensure_ascii=False, indent=2), encoding="utf-8")
+
+    deep_urls = {item["source_url"] for item in items}
+    gap_urls = sorted(
+        page["source_url"]
+        for page in mirror_manifest.get("pages", [])
+        if page.get("source_url") not in deep_urls
+    )
+    GAPS_PATH.write_text("\n".join(gap_urls) + ("\n" if gap_urls else ""), encoding="utf-8")
+
+    lines = [
+        "# Deep Action Report",
+        "",
+        f"- Captured deep states: `{manifest['captured_deep_states']}`",
+        f"- Source routes: `{len(route_rows)}`",
+        f"- Uncaptured mirrored routes: `{len(gap_urls)}`",
+        f"- Asset failures: `{len(manifest.get('asset_failures', []))}`",
+        f"- Remaining failures: `{len(manifest.get('failures', []))}`",
+        "",
+        "## Category Counts",
+        "",
+    ]
+    for category, count in sorted(category_counter.items()):
+        lines.append(f"- {category}: `{count}`")
+
+    lines.extend(["", "## Routes", ""])
+    for row in route_rows:
+        lines.append(f"### `{row['route']}`")
+        lines.append("")
+        lines.append(f"- Source URL: `{row['source_url']}`")
+        lines.append(f"- Captured states: `{row['count']}`")
+        lines.append(f"- Categories: `{json.dumps(row['categories'], ensure_ascii=False)}`")
+        lines.append("- Labels:")
+        for label in row["labels"]:
+            lines.append(f"  - {label}")
+        lines.append("")
+
+    REPORT_PATH.write_text("\n".join(lines) + "\n", encoding="utf-8")
+    print(
+        json.dumps(
+            {
+                "report": str(REPORT_PATH),
+                "summary": str(SUMMARY_PATH),
+                "gaps": str(GAPS_PATH),
+                "captured": manifest["captured_deep_states"],
+                "routes": len(route_rows),
+                "uncaptured_mirror_routes": len(gap_urls),
+            },
+            ensure_ascii=False,
+            indent=2,
+        )
+    )
+    return 0
+
+
+if __name__ == "__main__":
+    raise SystemExit(main())