tommbendall / tomplot Goto Github PK

https://github.com/tommbendall/tomplot/blob/main/tomplot/data_extraction.py#L458

panel_ids should be a list: [panel_ids]

Sometimes the highest value bin on the colorbar, above the reds, is set to blue again. This is quite confusing, as the very lowest and highest bins have the same color.

I don't know if this is a big issue, as you don't often want to plot a completely constant field! However, I noticed this when trying to verify that my initial field was indeed a constant value.

They are currently not tested

On machines without latex, we have to comment out set_tomplot_style(). It would be better to allow an exception in this situation.

The data extraction and reshaping routines don't have unit-tests -- and these have suffered from some bugs which would have been caught by having unit-tests

It would be nice to have the functionality to plot a diagnostic/prognostic along just one axis.

For example, a plot of temperature against height at a given horizontal point

The reshape function doesn't order the data points properly and will mix up different heights and not reshape correctly.
The original code

# for lev_idx in range(num_levels):
    #     data_slice = slice(lev_idx, num_hori_points*num_levels+lev_idx, num_levels)
    #     field_data[:, lev_idx] = data['field'].values[data_slice]
    #     coords_X[:, lev_idx] = data['X'].values[data_slice]
    #     coords_Y[:, lev_idx] = data['Y'].values[data_slice]
    #     coords_Z[:, lev_idx] = data['Z'].values[data_slice]
    #     for idx, _ in enumerate(other_arrays):
    #         new_other_arrays[idx][:, lev_idx] = data[f'other_{idx}'].values[data_slice]

The Hack fix

for hori_point in range(num_hori_points):
        hori_index = hori_point*num_levels
        this_X, this_Y = data['X'].values[hori_index], data['Y'].values[hori_index]
        data_column = data[(np.isclose(data['X'], this_X)) &
                           (np.isclose(data['Y'], this_Y))]
        data_column = data_column.sort_values(by=['Z'])
        field_data[hori_point,:] = data_column['field'].values[:]
        coords_X[hori_point,:] = data_column['X'].values[:]
        coords_Y[hori_point,:] = data_column['Y'].values[:]
        coords_Z[hori_point,:] = data_column['Z'].values[:]
        for idx, _ in enumerate(other_arrays):
            new_other_arrays[idx][hori_point,:] = data_column[f'other_{idx}'].values[:]

There needs to be an example showing how to use the convergence plot

I'm not sure that the vertical slice data is ordered by horizontal coordinate, which means it can't be used with the "contour" method. This should be straightforward to do.

It would be nice to generate automatic paddings for colorbar and axes labels