color_bins_layer breaks argument should take in iterables

I tried passing a numpy array for the breaks argument of a color_bins_layer helper:

from cartoframes.viz.helpers import color_bins_layer

breaks = np.arange(
    0,
    np.log(counties_w_counts.properties_per_household.max()),
    np.log(counties_w_counts.properties_per_household.max()) / 7,
)

Map(
    [
        color_bins_layer(
            counties_w_counts[["geometry", "properties_per_household"]],
            "properties_per_household",
            stroke_color="transparent",
            palette="sunset",
            breaks=breaks,
        ),
        Layer("SELECT * FROM mamataakella.state_bounds", "color: opacity(black, 0.5)"),
    ],
    show_info=True,
    viewport={"zoom": 3.15, "lat": 39.051997, "lng": -99.333419},
)

And received the following error:

---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
<ipython-input-308-147ad3fa75fa> in <module>
     14             stroke_color="transparent",
     15             palette="sunset",
---> 16             breaks=breaks,
     17         ),
     18         Layer("SELECT * FROM mamataakella.state_bounds", "color: opacity(black, 0.5)"),

~/.local/share/virtualenvs/property_data-_sgn6fy1/lib/python3.7/site-packages/cartoframes/viz/helpers/color_bins_layer.py in color_bins_layer(source, value, title, method, bins, breaks, palette, size, opacity, stroke_color, stroke_width, description, footer, legend, popup, widget, animate)
     47         raise ValueError('Available methods are: "quantiles", "equal", "stdev".')
     48 
---> 49     func = 'buckets' if breaks else {
     50         'quantiles': 'globalQuantiles',
     51         'equal': 'globalEqIntervals',

ValueError: The truth value of an array with more than one element is ambiguous. Use a.any() or a.all()

My expectation is that breaks can be an ordered iterable (list, tuple, numpy array). It works fine with list and tuple, but the error above leads me to think that it’s testing for whether it’s a list or tuple (and not a string). Since numpy has such a common use in peoples workflows for calculating custom arrays, we should support it too. A the very least we need to improve the error message that it should be a list/tuple or one of the named strings.

[andy-esch]

Thanks for the response, @makella! Yeah, looks like she has the same issue of needing to convert it to a python list first.

[andy-esch]

I used it because I wanted clear, evenly spaced values for the percentages. I could’ve tried method='equal' and bins=10 I guess but it felt easier to generate the bins I know I want ahead of time.

Just tried the way you suggest:

More or less the same except for the < and > for the lower/upper values.