BOKEH_STRICT does not raise validation errors

Hello,

I am plotting a rectangular grid where the axis labels are factors. If there are duplicated factors along any axis, as expected, an error should be raised. When I run my code I see:

ERROR:bokeh.core.validation.check:E-1019 (DUPLICATE_FACTORS): FactorRange must specify a unique list of categorical factors for an axis: duplicate factors found: 'a'

If I set BOKEH_STRICT=1 I don’t see any difference. I was expecting an exception being raised.

In the validation section of the documentation I found the check_integrity function which I hoped would raise the DUPLICATE_FACTORS error, but it does not. Is this a bug or expected behaviour?

How can I detect this issue before rendering to html? How can I catch the DUPLICATE_FACTORS validation error?

I’ve tested this on Bokeh 2.2.3 and 2.1.1

Below is my code:

import pandas as pd
import numpy as np

from bokeh.plotting import Figure
import bokeh.palettes as palettes
from bokeh.models import LinearColorMapper
from bokeh.transform import transform
from bokeh.resources import CDN
from bokeh.embed import file_html


index = ["a", "a"]
columns = ["b"]
N = len(index)
M = len(columns)
data = pd.DataFrame(np.random.rand(N * M).reshape(N, M), index=index, columns=columns)

ii, jj = np.meshgrid(data.columns, data.index, indexing="ij")
ii = ii.flatten()
jj = jj.flatten()
cc = [data.loc[j, i] for i, j in zip(ii, jj)]
plot_data = dict(x=ii, y=jj, colors=cc)

fig = Figure(
    x_axis_location="above",
    x_range=list(map(str, data.columns)),
    y_range=list(map(str, reversed(data.index))),
)
fig.rect(
    "x",
    "y",
    1,
    1,
    source=plot_data,
    color=transform("colors", LinearColorMapper(palette=palettes.magma(7))),
    line_color=None,
)

file_html(fig, CDN, "boom")

[bryevdv]

@RandyRcw that’s great, let me know if there is anything I can do to help you get started, or questions I can answer

[pablo-esteban]

I’ll do it. I just need to setup the dev environment…