Fix too large payload error on function apps by reducing rate of processing (#162)

Author: BenConstable9

deploy_ai_search_indexes/src/deploy_ai_search_indexes/ai_search.py

@@ -212,7 +212,7 @@ def get_mark_up_cleaner_skill(self, chunk_by_page: False) -> WebApiSkill:
             batch_size = 2
             degree_of_parallelism = 2
         else:
-            batch_size = 16
+            batch_size = 4
             degree_of_parallelism = 16
 
         if chunk_by_page:
@@ -273,10 +273,10 @@ def get_mark_up_cleaner_skill(self, chunk_by_page: False) -> WebApiSkill:
 
     def get_semantic_chunker_skill(
         self,
-        num_surrounding_sentences: int = 1,
+        num_surrounding_sentences: int = 2,
         similarity_threshold: float = 0.8,
-        max_chunk_tokens: int = 200,
-        min_chunk_tokens: int = 50,
+        max_chunk_tokens: int = 500,
+        min_chunk_tokens: int = 150,
     ) -> SplitSkill:
         """Get the skill for text split.
 
@@ -296,7 +296,7 @@ def get_semantic_chunker_skill(
             batch_size = 2
             degree_of_parallelism = 2
         else:
-            batch_size = 16
+            batch_size = 8
             degree_of_parallelism = 16
 
         semantic_text_chunker_skill_inputs = [

deploy_ai_search_indexes/src/deploy_ai_search_indexes/image_processing.py

@@ -301,10 +301,10 @@ def get_indexer(self) -> SearchIndexer:
         # Only place on schedule if it is not a test deployment
         if self.test:
             schedule = None
-            batch_size = 4
+            batch_size = 1
         else:
             schedule = {"interval": "PT15M"}
-            batch_size = 16
+            batch_size = 2
 
         if self.environment.use_private_endpoint:
             execution_environment = IndexerExecutionEnvironment.PRIVATE

image_processing/src/image_processing/semantic_text_chunker.py

@@ -12,10 +12,10 @@
 class SemanticTextChunker:
     def __init__(
         self,
-        num_surrounding_sentences: int = 1,
+        num_surrounding_sentences: int = 2,
         similarity_threshold: float = 0.8,
-        max_chunk_tokens: int = 200,
-        min_chunk_tokens: int = 50,
+        max_chunk_tokens: int = 500,
+        min_chunk_tokens: int = 150,
     ):
         self.num_surrounding_sentences = num_surrounding_sentences
         self.similarity_threshold = similarity_threshold