Similar to https://github.com/pandas-dev/pandas/issues/27522.
Currently we document pandas.io
as:
In addition, public functions in
pandas.io
andpandas.tseries
submodules are mentioned in the documentation.
Some submodules of pandas.io
have a clear delineation of whats public vs private, e.g. pandas.io.excel
where all submodules are prefixed with an underscore. Others, such as pandas.io.formats
, are not.
One possible way to clean this up is to do add pandas.api.io
and make various objects public by exposing them there.
Pinging people who helped introduce pandas.api.typing
: cc @simonjayhawkins @Dr-Irv @mroeschke @jorisvandenbossche @topper-123 @bashtage
Comment From: Dr-Irv
I think we are inconsistent on what we document from pandas.io
between the user guide and reference guide. I support the idea of moving them all to a pandas.api.io
package. But this should also be done by looking at the doc references and making sure that whatever is mentioned in the user guide has a corresponding reference manual page.
Comment From: rhshadrach
I think we are inconsistent on what we document from
pandas.io
between the user guide and reference guide.
Can you provide details here? The way I read the reference guide is "anything that appears in our documentation is public, otherwise it's private".
Comment From: Dr-Irv
Can you provide details here? The way I read the reference guide is "anything that appears in our documentation is public, otherwise it's private".
Here is one example. There may be more. In https://pandas.pydata.org/pandas-docs/stable/user_guide/io.html#sql-queries, we say "The pandas.io.sql module provides a collection of query wrappers ...", but pandas.io.sql
is not in the reference manual.