WeKnora deployment source with the reviewed wiki identity, summary and recovery fixes

scripts/verify_category_etl.py

@@ -0,0 +1,89 @@
+#!/usr/bin/env python3
+"""Verify category migration through authenticated, read-only acceptance APIs."""
+import argparse
+import datetime
+import json
+import urllib.error
+import urllib.parse
+import urllib.request
+from pathlib import Path
+
+from weknora_category_etl import TENANT, save_json
+
+
+BASE = "http://127.0.0.1:18080"
+
+
+def main():
+    parser = argparse.ArgumentParser(description=__doc__)
+    parser.add_argument("--manifest", type=Path, required=True)
+    parser.add_argument("--credentials", type=Path, required=True)
+    parser.add_argument("--out", type=Path, required=True)
+    args = parser.parse_args()
+    manifest = json.loads(args.manifest.read_text())
+    credentials = json.loads(args.credentials.read_text())
+    opener = urllib.request.build_opener(urllib.request.ProxyHandler({}))
+
+    def request(path, token=None, payload=None):
+        headers = {"Content-Type": "application/json"}
+        if token:
+            headers["Authorization"] = "Bearer " + token
+        req = urllib.request.Request(BASE + path, data=json.dumps(payload).encode() if payload else None, headers=headers)
+        try:
+            with opener.open(req, timeout=20) as response:
+                return response.status, json.load(response)
+        except urllib.error.HTTPError as error:
+            return error.code, None
+
+    status, auth = request("/api/v1/auth/login", payload={k: credentials[k] for k in ("email", "password")})
+    if status != 200 or auth.get("active_tenant", {}).get("id") != TENANT:
+        raise RuntimeError("Verification login failed or outside migration tenant")
+    token = auth["token"]
+    report = {"checked_at": datetime.datetime.now(datetime.timezone.utc).isoformat(), "tenant_id": TENANT, "checks": []}
+
+    def check(name, passed, **counts):
+        report["checks"].append({"check": name, "passed": bool(passed), **counts})
+
+    for kb in manifest["knowledge_bases"]:
+        files = [f for f in manifest["files"] if f["knowledge_base_id"] == kb]
+        tags = [t for t in manifest["tags"] if t["knowledge_base_id"] == kb]
+        status, response = request(f"/api/v1/knowledge-bases/{kb}/knowledge/folders", token)
+        tree = response.get("data", {}) if response else {}
+        nodes = {}
+
+        def walk(items):
+            for node in items:
+                nodes[node["path"]] = node
+                walk(node.get("children") or [])
+
+        walk(tree.get("folders") or [])
+        counts_match = all(t["folder_path"] in nodes and nodes[t["folder_path"]]["total_count"] == sum(
+            f["folder_path"] == t["folder_path"] or f["folder_path"].startswith(t["folder_path"] + "/") for f in files) for t in tags)
+        check("folder_tree", status == 200 and tree.get("total_document_count") == len(files) and counts_match,
+              knowledge_base_id=kb, folders=len(nodes), documents=tree.get("total_document_count"))
+        status, response = request(f"/api/v1/knowledge-bases/{kb}/tags?page=1&page_size=100", token)
+        data = response.get("data", {}) if response else {}
+        items = data.get("data", [])
+        check("tag_list", status == 200 and {t["id"] for t in items} == {t["id"] for t in tags},
+              knowledge_base_id=kb, tags=len(items))
+        for tag in tags:
+            params = urllib.parse.urlencode({"page": 1, "page_size": 100, "tag_ids": tag["id"]})
+            status, response = request(f"/api/v1/knowledge-bases/{kb}/knowledge?{params}", token)
+            data = response.get("data", []) if response else []
+            if isinstance(data, dict):
+                data = data.get("items", [])
+            expected = {r["knowledge_id"] for r in manifest["relations"] if r["tag_id"] == tag["id"]}
+            actual = {f["id"] for f in data}
+            check("tag_document_scope", status == 200 and actual == expected,
+                  legacy_category_id=tag["legacy_id"], expected=len(expected), returned=len(actual))
+        status, _ = request(f"/api/v1/knowledge-bases/{kb}/knowledge/folders")
+        check("anonymous_denied", status in (401, 403), knowledge_base_id=kb)
+    report["passed"] = all(c["passed"] for c in report["checks"])
+    save_json(args.out, report)
+    print(json.dumps({"passed": report["passed"], "checks": len(report["checks"]),
+                      "failed": [c for c in report["checks"] if not c["passed"]], "out": str(args.out)}))
+    return 0 if report["passed"] else 1
+
+
+if __name__ == "__main__":
+    raise SystemExit(main())