Update the ConfigReader to allow for empty chunk-by arrays (microsoft…

…#742)

.semversioner/next-release/patch-20240726200425411495.json

@@ -0,0 +1,4 @@
+{
+  "type": "patch",
+  "description": "update config-reader to allow for empty chunk-by arrays"
+}

graphrag/config/create_graphrag_config.py

@@ -382,12 +382,14 @@ def hydrate_parallelization_params(
                 base_dir=reader.str(Fragment.base_dir) or defs.STORAGE_BASE_DIR,
             )
         with reader.envvar_prefix(Section.chunk), reader.use(values.get("chunks")):
+            group_by_columns = reader.list("group_by_columns", "BY_COLUMNS")
+            if group_by_columns is None:
+                group_by_columns = defs.CHUNK_GROUP_BY_COLUMNS
+
             chunks_model = ChunkingConfig(
                 size=reader.int("size") or defs.CHUNK_SIZE,
                 overlap=reader.int("overlap") or defs.CHUNK_OVERLAP,
-                group_by_columns=reader.list("group_by_columns", "BY_COLUMNS")
-                or defs.CHUNK_GROUP_BY_COLUMNS,
-                encoding_model=reader.str(Fragment.encoding_model),
+                group_by_columns=group_by_columns,
             )
         with (
             reader.envvar_prefix(Section.snapshot),

graphrag/config/environment_reader.py

@@ -149,6 +149,7 @@ def list(
 
         if result is None:
             result = self.str(key, env_key)
-        if result:
-            return [s.strip() for s in result.split(",")]
+        if result is not None:
+            result = [s.strip() for s in result.split(",")]
+            return [s for s in result if s]
         return default_value