A community member is having trouble executing a code snippet from the "chain_of_table" notebook in the LlamaHub repository, encountering an "AttributeError: 'NoneType' object has no attribute 'group'" when running a query. Other community members suggest that the issue may be related to the LLM responding in a way that breaks the regex, and recommend exploring custom pipelines or developing a custom Text2SQL agent to handle more complicated queries. The community members also discuss the possibility of fine-tuning an LLM on the user's own SQL database to improve the performance of text-to-SQL conversions.
I am getting the "AttributeError: 'NoneType' object has no attribute 'group'" when running "response = query_engine.query("Who won best Director in the 1972 Academy Awards?")"
Thanks for the reply. Do you have any idea to deal with text2sql when the queries are so complicated and needs a more powerful tool than the standard text2sql query engine of LlamaIndex?
I have tried this, but still unable to execute a bit complicated queries. Can I develop my own agent for Text2SQL to work with Open-source LLMs? Also, do you think finetuning an LLM on my own SQL database can be helpful?