Better support for reactive data sources

[jnhyeon]

Thank you for sharing this wonderful and easy-to-use package.

I'm using query_chat on a large dataset.
Since the dataset is large, I'd like to filter it by date range and then input the dataframe into QueryChat().

Please add a feature that can solve this.

Thank you.

[cpsievert]

Hi @jnhyeon, this seems like a good use case for #191. With that, you could add filters on top of .df() instead of filtering the data that goes into QueryChat().

This does bring up a great larger point though that the current API isn't very supportive of reactive data sources (e.g., data upload, etc). I think this would actually make sense to address as a part of multi-table support, where we'll likely add a .add_table(), and as a result, QueryChat() could be constructed without a data source / table.

[jnhyeon]

Thank you for your reply.
I will test it once the new version is released.