Conversation
|
I'm a bit surprised by the test failure on python 2.7 (why would a hovertemplate be different with python 2.7 and python 3?) |
|
ahhh thank you :-). I'll change the test then. Python 3 users won't have the problem since dicts are insertion-ordered, and py2 users can always use an OrderedDict instead. |
|
So this approach looks pretty good. I don't love the string manipulations, although they seem pretty safe! One wrinkle: we support two different kinds of formatting... d3-format for numbers and d3-time-format for dates, and the delimiter between the attribute name and the formatting string is different ( |
|
Sure, let's add the delimiter. As for string manipulations, maybe there's a cleaner way to do it, this was a first pass to see if it worked. |
|
@nicolaskruchten I think this one is ready for review. |
|
Can we also accept something like |
I could maybe be convinced, but I liked the idea that the value is a formatting information, except when it's a tuple, in which case it's (formatting, data). It's also the logics which I have used in the code, converting everything to tuples internally. |
|
Right, but I feel like if we've already got a |
…ly.py into hover-format-or-skip
|
Not quite done the review, but it looks great so far! We need to merge this in before #2336, as I think I'll have an easier time rebasing :) |
|
So when we iterate through an "array attrable" and it's a dict, it sort of "just works" :) nice! |
short-circuit instead of bypass flag
Yes when you iterate through a dict it iterates through its keys. Simple and elegant ;-). |
|
OK, this is awesome :) 💃 |
|
Thanks for the 💃 let me just add the changelog entry and I'll merge. |



Closes #1774
Still need to add doc (and probably more tests), but early feedback is welcome.