[fix] plot_data_collect on paretoresults

[alxlyj]

Project Robyn

Summary

We were getting some errors mapping plot_data_collect accurately to our Python data classes.

The original implementation was flattening the R plot data structure, causing loss of hierarchy and making it difficult to access plot-specific data. This was evident from the different data structure we observed:

R Original Structure (Hierarchical):

plotDataCollect[[model_id]] <- list(
    plot1data = list(
        plotMediaShareLoopBar = data,
        plotMediaShareLoopLine = data,
        ySecScale = value
    ),
    plot2data = list(...),
    plot3data = list(...),
    # ... and so on
)

Python Before and After:

### Before:
converted_data = {
    'plotMediaShareLoopBar': pd.DataFrame(...),
    'plotMediaShareLoopLine': pd.DataFrame(...)
}

After:

converted_data = {
    'model_id': {
        'plot1data': {
            'plotMediaShareLoopBar': pd.DataFrame(...),
            'plotMediaShareLoopLine': pd.DataFrame(...),
            'ySecScale': value
        },
        'plot2data': {
            'plotWaterfallLoop': pd.DataFrame(...)
        }
        # ... other plot types
    }
    # ... other models
}

In this PR, we:

1. Modified _convert_plot_data function to maintain R's hierarchical structure
2. Added proper type handling for different data components
3. Preserved the plot type categories (plot1data through plot7data)

See data_mapper.ipynb for usage examples.

Test Plan