DH-4314 Add more logs for debugging

dataherald/api/fastapi.py

@@ -71,6 +71,7 @@ def scan_db(self, db_alias: str, table_name: str | None = None) -> bool:
     @override
     def answer_question(self, question: str, db_alias: str) -> NLQueryResponse:
         """Takes in an English question and answers it based on content from the registered databases"""
+        logger.info(f"Answer question: {question}")
         cache = self.system.instance(SmartCache)
         sql_generation = self.system.instance(SQLGenerator)
         evaluator = self.system.instance(Evaluator)
@@ -95,6 +96,8 @@ def answer_question(self, question: str, db_alias: str) -> NLQueryResponse:
             generated_answer = sql_generation.generate_response(
                 user_question, database_connection, context
             )
+
+            logger.info("Starts evaluator...")
             if evaluator.is_acceptable_response(
                 user_question, generated_answer, database_connection
             ):

dataherald/context_store/default.py

@@ -20,8 +20,7 @@ def __init__(self, system: System):
     def retrieve_context_for_question(
         self, nl_question: NLQuery, number_of_samples: int = 3
     ) -> List[dict] | None:
-        logger.info(f"getting context for {nl_question.question}")
-
+        logger.info(f"Getting context for {nl_question.question}")
         closest_questions = self.vector_store.query(
             query_texts=[nl_question.question],
             db_alias=nl_question.db_alias,

dataherald/smart_cache/in_memory.py

@@ -20,7 +20,7 @@ def __init__(self, settings: Settings):
 
     @override
     def add(self, key: str, value: NLQueryResponse) -> dict[str, Any]:
-        logger.info(f"Adding to cache: {value.dict()}")
+        logger.info(f"Adding to cache: {key}")
         self.cache[key] = value
         return {key: value}
 

dataherald/sql_database/base.py

@@ -1,4 +1,5 @@
 """SQL wrapper around SQLDatabase in langchain."""
+import logging
 from typing import Any, List
 
 from langchain.sql_database import SQLDatabase as LangchainSQLDatabase
@@ -9,6 +10,8 @@
 from dataherald.sql_database.models.types import DatabaseConnection
 from dataherald.utils.encrypt import FernetEncrypt
 
+logger = logging.getLogger(__name__)
+
 
 class SQLDatabase(LangchainSQLDatabase):
     """SQL Database.
@@ -43,6 +46,7 @@ def from_uri(
 
     @classmethod
     def get_sql_engine(cls, database_info: DatabaseConnection) -> "SQLDatabase":
+        logger.info(f"Connecting db: {database_info.alias}")
         fernet_encrypt = FernetEncrypt()
         if database_info.use_ssh:
             return cls.from_uri_ssh(database_info)

dataherald/sql_generator/langchain_sqlchain.py

@@ -46,18 +46,18 @@ def generate_response(
         context: List[dict] = None,
     ) -> NLQueryResponse:
         start_time = time.time()
+        self.database = SQLDatabase.get_sql_engine(database_connection)
         logger.info(
             f"Generating SQL response to question: {str(user_question.dict())} with passed context {context}"
         )
-        self.database = SQLDatabase.get_sql_engine(database_connection)
         if context is not None:
             samples_prompt_string = "The following are some similar previous questions and their correct SQL queries from these databases: \
             \n"
             for sample in context:
                 samples_prompt_string += (
                     f"Question: {sample['nl_question']} \nSQL: {sample['sql_query']} \n"
                 )
-        if context is not None:
+
             prompt = PROMPT_WITH_CONTEXT.format(
                 user_question=user_question.question, context=samples_prompt_string
             )