Preconfig conditional format

[aarondr77]

Description

Preconfigure conditional formatting in a Streamlit / Enterprise app.

Testing

See the tests + use the sample app I created in streamlit/v1/apps

new_dfs, code = spreadsheet(
    df, 
    column_definitions=[
        [
            {
                'columns': ['A', 'B'],
                'conditional_formats': [{
                    'filters': [{'condition': 'greater_than_or_equal', 'value': 5}], 
                    'font_color': '#c30010', 
                    'background_color': '#ffcbd1' 
                }] 
            },
            {
                'columns': ['A'],
                'conditional_formats': [{
                    'filters': [{'condition': 'less', 'value': 2}], 
                    'font_color': '#f30010', 
                    'background_color': '#ddcbd1' 
                }] 
            }
        ],
        [
            # Add code here to style the second dataframe
        ]
    ]
)

Documentation

Note if any new documentation needs to addressed or reviewed.

[vercel]

The latest updates on your projects. Learn more about Vercel for Git ↗︎

Name	Status	Preview	Comments	Updated (UTC)
monorepo	✅ Ready (Inspect)	Visit Preview	💬 Add feedback	Mar 13, 2024 1:40pm

[aarondr77]

Should this be a preprocessing Step?

Because the preprocessing steps are particularly ordered (they rely on each other's previous results to create dataframes for the passed args and finally return the df_names and dataframe args) it doesn't make sense to treat this as the same sort of preprocessing step as it causes additional complexity that is not obvious by just reading the step_managers.

Not really sure what the advantage of it being a preprocessing step vs just something that happens in the setup of the steps manager. The second approach is cleaner cause there is infrastructure overhead like passing the args, df_names around, and setting the execution data...

Proposal

1. Just make a function get_dataframe_formats_from_column_definititions(dfs: List[pd.DataFrame], column_definitions)
2. Call this function after the preprocessing steps in the step performer.

This will give us more type safety, (we can actually pass the dfs as a list of dataframes to do additional validation of the column definitions), and make the order of operations obvious in the steps manager.

In addition, it's even kinda of weird to have this as a step performer because its just setting the df_formats step performer and not actually generating any code / comments, etc.

I don't think we gain anything by explicitly making this a preprocessing step, right?

[naterush]

Code is really clean and looks nice. LGTM once you fix up the minor review here!

[naterush]

1. Merge in dev to get tests passing
2. Let's not forget to write docs