Greetings everyone! I am new to OpenAI API. So basically I am looking to build a chatbot that underneath is powered by 2 Excel sheets contained inside a single Excel file. One of 1600 rows x 30 cols and the other of about 20,000 rows x 20 cols. I have used AssistantsAPI via the code_intepretor and it works well but too expensive as the GPT-3.5-turbo doesn’t seem to answer only GPT-4-turbo. 2 main issues:
-
I created a new assistant but the assistant can’t seem to read the Excel files and it says that it is missing when it’s not.
-
I am looking for alternatives to cut the costs. I have tried PandasAI - some weird “code_runner” bug. I am unsure whether to explore vector data databases as the results I want it to produce involve heavy EDA/ Filtering. Is there a way to filter them via vector databases? Also please suggest your top choices for a beginner.
Thank you for any suggestions in advance