fix: harden qdrant payload and paging

harness/tht/adapters/vector/qdrant.py

@@ -318,15 +318,32 @@ class QdrantVectorStore:
         return result
 
     def _scroll(self, must: list[dict]) -> list[dict]:
-        response = self._call(
-            "POST",
-            f"/collections/{self._collection}/points/scroll",
-            {"with_payload": True, "limit": 10000, "filter": {"must": must}},
-        )
-        points = response.get("result", {}).get("points")
-        if not isinstance(points, list):
-            raise VectorStoreError("Qdrant returned malformed scroll response")
-        return points
+        points: list[dict] = []
+        offset = None
+        seen_offsets = set()
+        while True:
+            response = self._call(
+                "POST",
+                f"/collections/{self._collection}/points/scroll",
+                {
+                    "with_payload": True,
+                    "limit": 10000,
+                    "filter": {"must": must},
+                    "offset": offset,
+                },
+            )
+            result = response.get("result", {})
+            page = result.get("points")
+            if not isinstance(page, list):
+                raise VectorStoreError("Qdrant returned malformed scroll response")
+            points.extend(page)
+            next_page_offset = result.get("next_page_offset")
+            if next_page_offset is None:
+                return points
+            if next_page_offset in seen_offsets:
+                raise VectorStoreError("Qdrant returned malformed scroll response")
+            seen_offsets.add(next_page_offset)
+            offset = next_page_offset
 
     def _hit_from_point(self, point: dict) -> VectorHit:
         payload = point.get("payload")

harness/tht/vectorstore/records.py

@@ -30,6 +30,7 @@ def qdrant_semantic_kind(kind: str) -> str:
 def qdrant_payload(record: VectorRecord, *, content_hash: str, workspace_id: str) -> dict:
     semantic_kind = qdrant_semantic_kind(record.kind)
     return {
+        **record.metadata,
         "workspace_id": workspace_id,
         "kind": semantic_kind,
         "record_kind": record.kind,
@@ -38,7 +39,6 @@ def qdrant_payload(record: VectorRecord, *, content_hash: str, workspace_id: str
         "title": record.title,
         "content": record.content,
         "content_hash": content_hash,
-        **record.metadata,
     }