DOC: Explain the technical background of autoscaling#31673
Conversation
github-actions
Bot
added
the
Documentation: user guide
| # | ||
| # As a final step, the view limits can optionally be expanded outward to the | ||
| # nearest "nice" tick position, so that the axis edges coincide with tick | ||
| # marks. This is disabled by default, but can be turned on with the |
There was a problem hiding this comment.
You may want to add a comment that this is why tick Locators are involved in autoscaling at all.
There was a problem hiding this comment.
Involvement of tick locators is necessarily implied if you want to find "nice" tick positions.
Why is it meaningful to explicitly mention this? Is there any specific user-facing API or restriction that would make it valuable to highlight this? I would have thought this doesn't play any role unless you want to dig into the details of limit rounding.
There was a problem hiding this comment.
It's not strictly necessary, but I remember being confused by why locators mattered at all. (Another thing that makes this confusing is that round_numbers is not the default anymore, and probably not turned on very often.)
Anyways, either way is fine here, it was just a suggestion.
timhoffm
force-pushed
the
doc-autoscale-background
branch
from
d270950 to
8485193
Compare
timhoffm
force-pushed
the
doc-autoscale-background
branch
from
8485193 to
dd96766
Compare
anntzer
left a comment
anntzer
left a comment
There was a problem hiding this comment.
Still the minor suggestion about the role of locators in round numbers, but I don't mind either way, it's already a clear improvement.
…673-on-v3.11.x Backport PR #31673 on branch v3.11.x (DOC: Explain the technical background of autoscaling)
4 participants
This adds an in-depth explanation on the different components relevant to autoscaling and how they interact.
This should help users with advanced autoscaling needs to set the relevant parts appropriately. It should also
serve maintainers as an overview how this currently works to give context when thinking on how to improve
the autoscaling mechanism.