Skip to content

Make Text._get_layout simpler to follow. #12951

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 1 commit into from
Jan 13, 2019

Conversation

anntzer
Copy link
Contributor

@anntzer anntzer commented Dec 9, 2018

  • Instead of constructing and accessing multidimensional arrays whs and
    horizLayout by (opaque) index, construct separate lists with more
    meaningful names.
  • Define xmin/xmax/ymin/ymax/width/height closer to their place of use.
  • Make computation of offsetx and offsety more symmetric between
    rotation_mode == "anchor" and == "default".

PR Summary

PR Checklist

  • Has Pytest style unit tests
  • Code is Flake 8 compliant
  • New features are documented, with examples if plot related
  • Documentation is sphinx and numpydoc compliant
  • Added an entry to doc/users/next_whats_new/ if major new feature (follow instructions in README.rst there)
  • Documented in doc/api/api_changes.rst if API changed in a backward-incompatible way

Copy link
Member

@timhoffm timhoffm left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Also would benefit from renaming some few-character variables, e.g. d -> descent

baseline = (h - d) - thisy

if i == 0:
Copy link
Member

@timhoffm timhoffm Dec 9, 2018

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Suggested change
if i == 0:
if thisy == 0: # first line

gets rid of the need for enumerate().

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

It's not immediately obvious that thisy cannot be zero for the second line too even if the first line is empty (it's true, but only thanks to the h = max(h, lp_h)/d = max(d, lp_d) lines) so I think leaving it as it is is clearer.

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Just by common sense, the vertical position thisy must increase also for empty lines. So it should only be zero before handling the first line.

But ok, if you want to leave it.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I don't think common sense applies to this function (or to a lot of matplotlib, sadly). (e.g. if you look at the way center_baseline is defined a bit below, it looks weirdly different between anchor and non-anchor rotations...)

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Funnily that got fixed by #13029 :)

horizLayout[i] = thisx, thisy, w, h
thisy -= max(min_dy, (h - d) * self._linespacing)

xs.append(thisx) # == 0.
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Alterantively

xs = [0.] * len(ys)

outside of the loop.

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I think the symmetry of handling between xs and ys is nicer.

Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

IMO xs = [0.] * len(ys) is much clearer about xs than secretly constructing a list of zeros in the loop. The loop should only need to bother about the quantities that actually depend on the lines. But not going to argue about this.

@anntzer
Copy link
Contributor Author

anntzer commented Dec 9, 2018

The single-character variables are "inside the loop", the longer ones are the one that outlive the loop.

@tacaswell tacaswell added this to the v3.1 milestone Dec 9, 2018
@timhoffm
Copy link
Member

timhoffm commented Dec 9, 2018

The single-character variables are "inside the loop", the longer ones are the one that outlive the loop.

Which still doesn't make the single-character variables readable.

Copy link
Member

@timhoffm timhoffm left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Accepting as an incremental improvement, even though I see room for further improvement.

@anntzer
Copy link
Contributor Author

anntzer commented Dec 9, 2018

The single-letter variables are defined as

w, h, d = renderer.get_text_width_height_descent(...)

so their meaning is pretty clear IMO (in the context).

@timhoffm
Copy link
Member

timhoffm commented Dec 9, 2018

That's fine in a narrow scope like a list comprehension or a few-lines for. Single letter variables should just not live for more than 20 lines. And in particular not multiple different ones at the same time.

@anntzer
Copy link
Contributor Author

anntzer commented Jan 2, 2019

Additionally factored out offsets computation, following #13029.

@anntzer anntzer mentioned this pull request Jan 12, 2019
6 tasks
- Instead of constructing and accessing multidimensional arrays whs and
  horizLayout by (opaque) index, construct separate lists with more
  meaningful names.
- Define xmin/xmax/ymin/ymax/width/height closer to their place of use.
- Make computation of offsetx and offsety more symmetric between
  rotation_mode == "anchor" and == "default".
@tacaswell tacaswell merged commit 4cdd071 into matplotlib:master Jan 13, 2019
@anntzer anntzer deleted the get_layout branch January 13, 2019 21:31
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

Successfully merging this pull request may close these issues.

3 participants