👍 For feature.

There might also be issues with the _bbox_patch and it's associated helper functions.

This should address additional comments raised in #345. I'm not sure if this TextCollection has any significant performance benefits (since that's not my driving interest), but #347 may be relevant.

I'll fix the pep8 failures and add some tests once the design is vetted.

I am also 👍 on not being worried about performance with this.

The ability to easily do broadcasting of interesting properties over artists is one of the places that bokeh is well ahead of us. I wonder if in addition to the Collection subclasses (which really are about performance optimizations as all of the broadcast logic lives in their draw methods) it would be worth providing a top-level generic broadcasting function, something like:

def broadcast_magic(ax, Art_cls, **kwargs):
    static_kwargs = dict()
    broadcast_names = []
    max_count = -1
    cycle_values = []
    for k, v in kwargs.items:
        if is_iterable(v) and not is_string(v):
           broadcast_name.append(k)
           cycle_values = cycle(v)
           if len(v) > max_count:
               max_count = len(v)
        else:
            static_kwargs[k] = v
    ret = []
    for count, b_vals in enumerate(zip(cycle_values)):
        if not count <max_count:
            break
        local_kwargs = dict(static_kwargs)
        local_kwargs.update({k: v for k, v in zip(broadcast_names, b_vals)})
        art = Art_cls(**local_kwargs)
        ax.add_artist(art)
        ret.append(art)
    return ret

If you only need TextCollection to group properties, then can't we do that with ASS? As far as I see that should do the same thing but in a much more general way.

@OceanWolf I'm lost on your acronym, and strangely, googling it isn't helping. :)

Artist Style Sheets. It is one of the MEPs.

On Mon, Jun 29, 2015 at 2:32 PM, Ryan May notifications@github.com wrote:

@OceanWolf https://github.com/OceanWolf I'm lost on your acronym, and
strangely, googling it isn't helping. :)

—
Reply to this email directly or view it on GitHub
#4063 (comment)
.

It's unclear to me if MEP26 is relevant. I want to easily be able to put a bunch of text on a plot, all with the same style (maybe changing color, but that's all). This facilitates plotting many text values (or even just strings) in data coordinates, often offset from the center location.

This can certainly be accomplished now directly with a for loop; this seems inefficient and verbose. TextCollection would essentially let Text take arrays. As far as I can tell, ASS would simply be a better way to do the for loop method.

This should inherit from Collection so that it may be color-mappable and such.

No objections to this being re-opened, but closing because its been 3 years, and I'm not entirely sure what the advantage is over just having a for-loop to add the text (and creating a list of handles if you want to store them for future reference). I imagine most folks who want this are getting by using this method.

It's relevant to #5665 (moving ticks to use LineCollections instead of individual lines would (perhaps) logically move label texts to TextCollections as well). But I'm fine leaving this closed for now, we can always revisit it later.