The community member is using pagedcsvreader and chromadb to query CSV data, but is finding that the top 10 most similar documents are not relevant to the query, even when specifying important keywords. The comments suggest that using top_k similarity may not be the best approach for most CSV data, and that text2sql or a similar tool may be more appropriate, especially if the CSV contains textual data like QA pairs. The community members also discuss challenges with setting up a chatbot using CSV or SQL data, including issues with handling history and memory, and mention a tutorial on using a QueryPipelineAgent. However, there is no explicitly marked answer in the comments.
Hi all , iam used pagedcsvreader along with chromadb and when iam querying with similarity_top_k as 10 but none of the documents are relevant to query. Although iam directly specifying important keywords in query. What can I do to improve rag with CSV data
Tried doing text2sql by loading data into sqlite, but it's not able to answer based on previous question/ previous output . Also even for a hello it's trying to execute SQL query . Is there any tutorial to create chatbot with CSV , SQL data where history is also taken care or we have memory buffer
