Skip to content

RAGFlow is an agent-native database: a Retrieval-Augmented Generation (RAG) engine that grew into context infrastructure.

fix(api): add missing comma delimiter between question and extracted keywords (#19661)

ming1523c5a932ffb042
diff --git a/api/apps/restful_apis/bot_api.py b/api/apps/restful_apis/bot_api.py
index 3da6ae6a2a..9703b58c57 100644
--- a/api/apps/restful_apis/bot_api.py
+++ b/api/apps/restful_apis/bot_api.py
@@ -448,7 +448,7 @@ async def retrieval_test_embedded(tenant_id=None):
if req.get("keyword", False):
default_chat_model = await thread_pool_exec(get_tenant_default_model_by_type, kb.tenant_id, LLMType.CHAT)
chat_mdl = LLMBundle(kb.tenant_id, default_chat_model)
- _question += await keyword_extraction(chat_mdl, _question)
+ _question += "," + await keyword_extraction(chat_mdl, _question)
labels = label_question(_question, [kb])
ranks = await settings.retriever.retrieval(
diff --git a/api/apps/restful_apis/chunk_api.py b/api/apps/restful_apis/chunk_api.py
index 68983a8b29..8bb23d5306 100644
--- a/api/apps/restful_apis/chunk_api.py
+++ b/api/apps/restful_apis/chunk_api.py
@@ -521,7 +521,7 @@ async def retrieval_test(tenant_id, dataset_id=None):
question = await cross_languages(kb.tenant_id, None, question, langs)
if req.get("keyword", False):
chat_model_config = get_tenant_default_model_by_type(kb.tenant_id, LLMType.CHAT)
- question += await keyword_extraction(LLMBundle(kb.tenant_id, chat_model_config), question)
+ question += "," + await keyword_extraction(LLMBundle(kb.tenant_id, chat_model_config), question)
ranks = await settings.retriever.retrieval(
question,
diff --git a/api/apps/services/dataset_api_service.py b/api/apps/services/dataset_api_service.py
index 16eb7ffd3e..08cfcfe124 100644
--- a/api/apps/services/dataset_api_service.py
+++ b/api/apps/services/dataset_api_service.py
@@ -1154,7 +1154,7 @@ async def search(dataset_id: str, tenant_id: str, req: dict):
if search_config.get("keyword", req.get("keyword", False)):
default_chat_model_config = get_tenant_default_model_by_type(kb.tenant_id, LLMType.CHAT)
chat_mdl = LLMBundle(kb.tenant_id, default_chat_model_config)
- _question += await keyword_extraction(chat_mdl, _question)
+ _question += "," + await keyword_extraction(chat_mdl, _question)
labels = label_question(_question, [kb])
ranks = await settings.retriever.retrieval(
@@ -1560,7 +1560,7 @@ async def search_datasets(tenant_id: str, req: dict):
if search_config.get("keyword", req.get("keyword", False)):
default_chat_model_config = get_tenant_default_model_by_type(kb.tenant_id, LLMType.CHAT)
chat_mdl = LLMBundle(kb.tenant_id, default_chat_model_config)
- _question += await keyword_extraction(chat_mdl, _question)
+ _question += "," + await keyword_extraction(chat_mdl, _question)
labels = label_question(_question, kbs)
ranks = await settings.retriever.retrieval(