lightrag-memgraph 0.3.2__tar.gz → 0.3.3__tar.gz

{lightrag_memgraph-0.3.2 → lightrag_memgraph-0.3.3}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.5
 Name: lightrag-memgraph
-Version: 0.3.2
+Version: 0.3.3
 Summary: LightRAG integration with Memgraph
 Project-URL: Homepage, https://github.com/memgraph/ai-toolkit
 Project-URL: Source, https://github.com/memgraph/ai-toolkit/tree/main/integrations/lightrag-memgraph

{lightrag_memgraph-0.3.2 → lightrag_memgraph-0.3.3}/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "lightrag-memgraph"
-version = "0.3.2"
+version = "0.3.3"
 description = "LightRAG integration with Memgraph"
 readme = "README.md"
 requires-python = ">=3.10"

{lightrag_memgraph-0.3.2 → lightrag_memgraph-0.3.3}/src/lightrag_memgraph/docstatus_impl.py

@@ -240,6 +240,7 @@ class MemgraphDocStatusStorage(MemgraphCrudMixin, DocStatusStorage):
 
         where = "" if status_values is None else "WHERE n.status IN $statuses"
         sort_expr = "n.id" if sort_field == "id" else f"n.{sort_field}"
+        order_by = f"{sort_expr} {order}" if sort_field == "id" else f"n.{sort_field} {order}, n.id {order}"
         params: dict[str, Any] = {"skip": skip, "limit": page_size}
         filter_params: dict[str, Any] = {}
         if status_values is not None:
@@ -265,7 +266,7 @@ class MemgraphDocStatusStorage(MemgraphCrudMixin, DocStatusStorage):
                         f"""
                         MATCH (n:`{self._label}`) {where}
                         RETURN n.id AS id, n.data AS data
-                        ORDER BY {sort_expr} {order}
+                        ORDER BY {order_by}
                         SKIP $skip LIMIT $limit
                         """
                     ),

{lightrag_memgraph-0.3.2 → lightrag_memgraph-0.3.3}/tests/test_integration_memgraph.py

@@ -368,6 +368,42 @@ async def test_doc_status_roundtrip(shared_driver, workspace):
         await store.drop()
 
 
+@pytest.mark.parametrize("sort_direction", ["asc", "desc"])
+async def test_doc_status_pagination_breaks_ties_by_id(shared_driver, workspace, sort_direction):
+    store = MemgraphDocStatusStorage(
+        namespace="doc_status",
+        workspace=workspace,
+        global_config=_global_config(),
+        embedding_func=_fake_embedding_func(),
+    )
+    await store.initialize()
+    try:
+        doc_count = 25  # page_size is clamped to >= 10, so this spans three pages
+        doc_ids = [f"doc-{i:02d}" for i in range(doc_count)]
+        # Written out of id order, so storage order can't pass for the tiebreaker.
+        scrambled = [doc_ids[(i * 7) % doc_count] for i in range(doc_count)]
+        # _doc() gives every document the same updated_at: the whole set is one tie.
+        await store.upsert({doc_id: _doc(DocStatus.PROCESSED, doc_id, f"{doc_id}.txt") for doc_id in scrambled})
+
+        async def page_through() -> list[str]:
+            seen: list[str] = []
+            for page_number in (1, 2, 3):
+                page, total = await store.get_docs_paginated(
+                    page=page_number, page_size=10, sort_direction=sort_direction
+                )
+                assert total == doc_count
+                seen.extend(doc_id for doc_id, _ in page)
+            return seen
+
+        first_pass = await page_through()
+        assert len(first_pass) == len(set(first_pass)), "a document appeared on more than one page"
+        assert set(first_pass) == set(doc_ids), "a document was missing from every page"
+        assert first_pass == sorted(doc_ids, reverse=sort_direction == "desc")
+        assert await page_through() == first_pass
+    finally:
+        await store.drop()
+
+
 # --- vector storage ----------------------------------------------------------