Fix id baseline (#2036)

* Fix all human_readable_id columns to start at 0

* Semver

.semversioner/next-release/patch-20250827005334747623.json

@@ -0,0 +1,4 @@
+{
+  "type": "patch",
+  "description": "Fix all human_readable_id fields to be 0-based."
+}

graphrag/index/workflows/create_final_documents.py

@@ -69,7 +69,7 @@ def create_final_documents(
     ).reset_index(drop=True)
 
     rejoined["id"] = rejoined["id"].astype(str)
-    rejoined["human_readable_id"] = rejoined.index + 1
+    rejoined["human_readable_id"] = rejoined.index
 
     if "metadata" not in rejoined.columns:
         rejoined["metadata"] = pd.Series(dtype="object")

graphrag/index/workflows/create_final_text_units.py

@@ -60,7 +60,7 @@ def create_final_text_units(
 ) -> pd.DataFrame:
     """All the steps to transform the text units."""
     selected = text_units.loc[:, ["id", "text", "document_ids", "n_tokens"]]
-    selected["human_readable_id"] = selected.index + 1
+    selected["human_readable_id"] = selected.index
 
     entity_join = _entities(final_entities)
     relationship_join = _relationships(final_relationships)

graphrag/index/workflows/extract_covariates.py

@@ -88,6 +88,6 @@ async def extract_covariates(
     )
     text_units.drop(columns=["text_unit_id"], inplace=True)  # don't pollute the global
     covariates["id"] = covariates["covariate_type"].apply(lambda _x: str(uuid4()))
-    covariates["human_readable_id"] = covariates.index + 1
+    covariates["human_readable_id"] = covariates.index
 
     return covariates.loc[:, COVARIATES_FINAL_COLUMNS]

tests/verbs/test_extract_covariates.py

@@ -58,8 +58,8 @@ async def test_extract_covariates():
     assert_series_equal(actual["text_unit_id"], input["id"], check_names=False)
 
     # make sure the human ids are incrementing
-    assert actual["human_readable_id"][0] == 1
-    assert actual["human_readable_id"][1] == 2
+    assert actual["human_readable_id"][0] == 0
+    assert actual["human_readable_id"][1] == 1
 
     # check that the mock data is parsed and inserted into the correct columns
     assert actual["covariate_type"][0] == "claim"

tests/verbs/test_finalize_graph.py

@@ -30,9 +30,6 @@ async def test_finalize_graph():
         "relationships", context.output_storage
     )
 
-    assert len(nodes_actual) == 291
-    assert len(edges_actual) == 452
-
     # x and y will be zero with the default configuration, because we do not embed/umap
     assert nodes_actual["x"].sum() == 0
     assert nodes_actual["y"].sum() == 0
@@ -58,9 +55,6 @@ async def test_finalize_graph_umap():
         "relationships", context.output_storage
     )
 
-    assert len(nodes_actual) == 291
-    assert len(edges_actual) == 452
-
     # x and y should have some value other than zero due to umap
     assert nodes_actual["x"].sum() != 0
     assert nodes_actual["y"].sum() != 0

tests/verbs/test_prune_graph.py

@@ -28,4 +28,4 @@ async def test_prune_graph():
 
     nodes_actual = await load_table_from_storage("entities", context.output_storage)
 
-    assert len(nodes_actual) == 21
+    assert len(nodes_actual) == 20