feat: add SQL examples in prompt

Author: ulleo

backend/apps/data_training/curd/data_training.py

@@ -198,7 +198,7 @@ def save_embeddings(session: Session, ids: List[int]):
 ( 1 - (embedding <=> :embedding_array) ) AS similarity
 FROM data_training AS child
 ) TEMP
-WHERE similarity > {settings.EMBEDDING_SIMILARITY} and oid = :oid and datasource = :datasource
+WHERE similarity > {settings.EMBEDDING_DATA_TRAINING_SIMILARITY} and oid = :oid and datasource = :datasource
 ORDER BY similarity DESC
 LIMIT {settings.EMBEDDING_DATA_TRAINING_TOP_COUNT}
 """

backend/apps/terminology/curd/terminology.py

@@ -300,7 +300,7 @@ def save_embeddings(session: Session, ids: List[int]):
 ( 1 - (embedding <=> :embedding_array) ) AS similarity
 FROM terminology AS child
 ) TEMP
-WHERE similarity > {settings.EMBEDDING_SIMILARITY} and oid = :oid
+WHERE similarity > {settings.EMBEDDING_TERMINOLOGY_SIMILARITY} and oid = :oid
 ORDER BY similarity DESC
 LIMIT {settings.EMBEDDING_TERMINOLOGY_TOP_COUNT}
 """

backend/common/core/config.py

@@ -89,7 +89,9 @@ def SQLALCHEMY_DATABASE_URI(self) -> PostgresDsn | str:
     LOCAL_MODEL_PATH: str = '/opt/sqlbot/models'
     DEFAULT_EMBEDDING_MODEL: str = 'shibing624/text2vec-base-chinese'
     EMBEDDING_ENABLED: bool = True
-    EMBEDDING_SIMILARITY: float = 0.4
+    EMBEDDING_DEFAULT_SIMILARITY: float = 0.4
+    EMBEDDING_TERMINOLOGY_SIMILARITY: float = EMBEDDING_DEFAULT_SIMILARITY
+    EMBEDDING_DATA_TRAINING_SIMILARITY: float = EMBEDDING_DEFAULT_SIMILARITY
     EMBEDDING_DEFAULT_TOP_COUNT: int = 5
     EMBEDDING_TERMINOLOGY_TOP_COUNT: int = EMBEDDING_DEFAULT_TOP_COUNT
     EMBEDDING_DATA_TRAINING_TOP_COUNT: int = EMBEDDING_DEFAULT_TOP_COUNT