feat(crawler): add pixelrag visual evidence lane

scripts/ops/report_pixelrag_crawler_integration.py

@@ -0,0 +1,122 @@
+#!/usr/bin/env python3
+"""Report PixelRAG-style crawler integration readiness."""
+
+from __future__ import annotations
+
+import argparse
+import json
+import sys
+from pathlib import Path
+
+
+ROOT = Path(__file__).resolve().parents[2]
+if str(ROOT) not in sys.path:
+    sys.path.insert(0, str(ROOT))
+
+from services.pixelrag_crawler_integration_service import (  # noqa: E402
+    build_pixelrag_crawler_integration_assessment,
+    build_pixelrag_visual_evidence_manifest,
+)
+
+
+CAPABILITY_FLAGS = {
+    "no_playwright": ("playwright_artifact_pipeline", False),
+    "ollama_multimodal_ready": ("ollama_multimodal_embedding_ready", True),
+    "pgvector_visual_ready": ("pgvector_visual_index_ready", True),
+    "faiss_allowed": ("faiss_allowed_in_production", True),
+}
+
+
+def _size_arg(value: str) -> dict[str, int | str]:
+    try:
+        name, size = value.split("=", 1)
+    except ValueError:
+        name, size = "", value
+    try:
+        width_text, height_text = size.lower().split("x", 1)
+        width = int(width_text)
+        height = int(height_text)
+    except (ValueError, TypeError):
+        raise argparse.ArgumentTypeError("size must be name=WIDTHxHEIGHT or WIDTHxHEIGHT")
+    payload: dict[str, int | str] = {"width": width, "height": height}
+    if name:
+        payload["name"] = name
+    return payload
+
+
+def main() -> int:
+    parser = argparse.ArgumentParser(
+        description="輸出 PixelRAG 導入現有爬蟲解決方案的機器可讀評估。"
+    )
+    parser.add_argument(
+        "--platform",
+        action="append",
+        dest="platforms",
+        help="限制目標平台，可重複指定；預設 momo + pchome。",
+    )
+    parser.add_argument(
+        "--capability",
+        action="append",
+        choices=sorted(CAPABILITY_FLAGS),
+        default=[],
+        help="覆寫 readiness 假設，用於模擬下一階段條件。",
+    )
+    parser.add_argument("--manifest-url", help="輸出單一 URL 的視覺證據 manifest。")
+    parser.add_argument("--crawler", default="crawler_diagnostics", help="manifest 來源 crawler 名稱。")
+    parser.add_argument(
+        "--trigger-reason",
+        default="parser_empty_or_low_confidence",
+        help="啟動視覺 fallback 的原因。",
+    )
+    parser.add_argument(
+        "--evidence-intent",
+        default="recover_visual_offer_evidence",
+        help="視覺證據用途。",
+    )
+    parser.add_argument(
+        "--viewport",
+        type=_size_arg,
+        help="viewport，例如 desktop-1440=1440x950。",
+    )
+    parser.add_argument(
+        "--page-size",
+        type=_size_arg,
+        help="預估頁面尺寸，例如 page=1440x1900。",
+    )
+    parser.add_argument(
+        "--tile-size",
+        type=_size_arg,
+        help="tile 尺寸，例如 tile=512x512。",
+    )
+    args = parser.parse_args()
+
+    if args.manifest_url:
+        platform = (args.platforms or ["momo"])[0]
+        payload = build_pixelrag_visual_evidence_manifest(
+            url=args.manifest_url,
+            platform=platform,
+            crawler=args.crawler,
+            trigger_reason=args.trigger_reason,
+            evidence_intent=args.evidence_intent,
+            viewport=args.viewport,
+            page_size=args.page_size,
+            tile_size=args.tile_size,
+        )
+        print(json.dumps(payload, ensure_ascii=False, indent=2, sort_keys=True))
+        return 0 if payload.get("success") else 1
+
+    capabilities = {
+        key: value
+        for flag in args.capability
+        for key, value in [CAPABILITY_FLAGS[flag]]
+    }
+    payload = build_pixelrag_crawler_integration_assessment(
+        capabilities=capabilities or None,
+        target_platforms=tuple(args.platforms) if args.platforms else None,
+    )
+    print(json.dumps(payload, ensure_ascii=False, indent=2, sort_keys=True))
+    return 0 if payload.get("success") else 1
+
+
+if __name__ == "__main__":
+    raise SystemExit(main())