Fix summarization including empty descriptions (microsoft#1124)

* Fix summarization including empty descriptions * Update
KylinMountain · Sep 12, 2024 · fa2c3d0 · fa2c3d0
1 parent 88a071d
commit fa2c3d0
Show file tree

Hide file tree

Showing 2 changed files with 10 additions and 2 deletions.
diff --git a/.semversioner/next-release/patch-20240911214150592475.json b/.semversioner/next-release/patch-20240911214150592475.json
@@ -0,0 +1,4 @@
+{
+  "type": "patch",
+  "description": "Fix summarization including empty descriptions"
+}
diff --git a/graphrag/index/graph/extractors/summarize/description_summary_extractor.py b/graphrag/index/graph/extractors/summarize/description_summary_extractor.py
@@ -67,7 +67,7 @@ async def __call__(
         result = ""
         if len(descriptions) == 0:
             result = ""
-        if len(descriptions) == 1:
+        elif len(descriptions) == 1:
             result = descriptions[0]
         else:
             result = await self._summarize_descriptions(items, descriptions)
@@ -87,7 +87,11 @@ async def _summarize_descriptions(
         if not isinstance(descriptions, list):
             descriptions = [descriptions]
 
-            # Iterate over descriptions, adding all until the max input tokens is reached
+        # Sort description lists
+        if len(descriptions) > 1:
+            descriptions = sorted(descriptions)
+
+        # Iterate over descriptions, adding all until the max input tokens is reached
         usable_tokens = self._max_input_tokens - num_tokens_from_string(
             self._summarization_prompt
         )