Build Agentic workflows, RAG pipelines, with rich AI model and tool support on one collaborative workspace. Deploy on cloud, VPC, or self-hosted, so teams move from prototype to production without rebuilding the stack.
Python agentagentic-aiagentic-frameworkagentic-workflowaiautomationclaudedeepseekgenaigptllmlow-codemcpnextjsno-codeopenaipythonskillsworkflow
fix(vdb): add document_id keyword mapping and fix VectorType in elasticsearch-ja adapter (#40400)
Rafli Surya Wijaya385e8662bbfc
diff --git a/api/providers/vdb/vdb-elasticsearch/src/dify_vdb_elasticsearch/elasticsearch_ja_vector.py b/api/providers/vdb/vdb-elasticsearch/src/dify_vdb_elasticsearch/elasticsearch_ja_vector.pyindex cff39830f88..49da9db987a 100644--- a/api/providers/vdb/vdb-elasticsearch/src/dify_vdb_elasticsearch/elasticsearch_ja_vector.py+++ b/api/providers/vdb/vdb-elasticsearch/src/dify_vdb_elasticsearch/elasticsearch_ja_vector.py@@ -72,7 +72,10 @@ class ElasticSearchJaVector(ElasticSearchVector):Field.METADATA_KEY: {"type": "object","properties": {- "doc_id": {"type": "keyword"} # Map doc_id to keyword type+ "doc_id": {"type": "keyword"}, # Map doc_id to keyword type+ "document_id": {+ "type": "keyword"+ }, # Map document_id to keyword type; required for document-scoped retrieval},},}@@ -91,7 +94,7 @@ class ElasticSearchJaVectorFactory(ElasticSearchVectorFactory):else:dataset_id = dataset.idcollection_name = Dataset.gen_collection_name_by_id(dataset_id)- dataset.index_struct = json.dumps(self.gen_index_struct_dict(VectorType.ELASTICSEARCH, collection_name))+ dataset.index_struct = json.dumps(self.gen_index_struct_dict(VectorType.ELASTICSEARCH_JA, collection_name))config = current_app.configreturn ElasticSearchJaVector(