boldsymbol support for mathtext #25661

devRD · 2023-04-11T06:28:46Z

PR Summary

Fixes #25643
Fixes #1366
Depends on PR #25359
Add boldsymbol command

PR Checklist

Documentation and Tests

Has pytest style unit tests (and pytest passes)
Documentation is sphinx and numpydoc compliant (the docs should build without error).
New plotting related features are documented with examples.

Release Notes

New features are marked with a .. versionadded:: directive in the docstring and documented in doc/users/next_whats_new/
API changes are marked with a .. versionchanged:: directive in the docstring and documented in doc/api/next_api_changes/
Release notes conform with instructions in next_whats_new/README.rst or next_api_changes/README.rst

lib/matplotlib/_mathtext.py

oscargus

Looks good! One may consider updating the what's new-note in #25359 in some way instead of having a separate one, but for now this is as good as it can be.

And maybe add something to the documentation, but will be easier once #25359 is in.

oscargus · 2023-04-11T12:26:03Z

doc/users/next_whats_new/boldsym_mathtext.rst

+To change font to bold and italic enclose the text in a font command as
+shown:
+
+.. code-block::


I do not fully get this warning, but I guess it must be fixed one way or the other. Maybe better to add a figure in the other PR, like plt.text(...) and then add the alternative version in that figure here?

Currently, I am getting these results from the current implementation. Should I add the mathbfit to the other PR, and the boldsymbol to the current one?

I think it is just missing a blank line.

.. code-block:: r'$\boldsymbol{a+2+\alpha}$'

(note also that the closing quote is missing anyway)

anntzer · 2023-04-20T19:27:36Z

lib/matplotlib/_mathtext.py

+                hlist.append(c)
+            else:
+                hlist.append(c)
+        self.pop_state()


Didn't really check, but does that work with nesting? (i.e.\boldsymbol{\mathnormal{foo}} -- the inner one should win)

I'm not actually sure what the semantics of boldsymbol are in TeX, but shouldn't it just be an alias for mathbfit? (and perhaps the test could be based on check_figures_equal too)

The check_figures_equal test in the PR currently gives this output:

ksunden · 2023-04-25T19:37:54Z

The PR linked above has been merged, is this still waiting on anything?

devRD · 2023-04-26T04:21:37Z

Hi @ksunden, This PR is not waiting, I opened it for review. Thanks!

lib/matplotlib/tests/test_mathtext.py

anntzer · 2023-05-01T16:01:12Z

lib/matplotlib/tests/test_mathtext.py

+    fig_test.text(0.1, 0.1, r"$\boldsymbol{abc0123\alpha}$")
+    fig_test.text(0.1, 0.2, r"$\boldsymbol{\mathrm{abc0123\alpha}}$")
+
+    fig_ref.text(0.1, 0.1, r"$\boldsymbol{abc0123\alpha}$")


I think this should use \mathbfit, no?

anntzer · 2023-05-02T07:15:51Z

I'm sorry for doing this review in a piecemeal fashion, but I think that we still don't know exactly what are the semantics of boldsymbol in tex and should thus figure this out exactly first.
In particular, looking at this again, it seems like \boldsymbol{\mathrm{foo}} prints foo in bold roman (try with text.latex.preamble = "\usepackage{amsmath}" and usetex=True, which suggests to me that boldsymbol is in fact not synonym with mathbfit at all, but perhaps simply toggles the boldness of the font without touching italicness or anything else?

devRD · 2023-05-02T19:48:15Z

After exploring a bit on boldsymbol, I agree that it slightly differs from mathbfit. From my limited understanding, I suppose the format differs in accordance with various symbols.
For e.g.:
Digits from 0-9 are upright bold
Alphabets A-Za-z are bold+italic
Greek symbols like \alpha and \omega are bold+italic but...
symbols \Gamma and \Omega are just upright bold.

I generated this image for \boldsymbol, \boldsymbol{\mathrm{foo}} and \mathbfit (in order) using text.preamble and text.usetex=True

(some symbols weren't rendered properly, I'm guessing due to LaTeX package differences)

This is the link that I used for reference. http://tug.ctan.org/tex-archive/macros/amstex/doc/amsguide.pdf

QuLogic · 2023-06-12T17:15:23Z

lib/matplotlib/_mathtext.py

+                           .lower(),
+                           range(ord('\N{GREEK SMALL LETTER ALPHA}'),
+                                 ord('\N{GREEK SMALL LETTER OMEGA}') + 1)))
+        allLatin = [*lw, *up]


This is equal to string.ascii_letters.

QuLogic · 2023-06-12T21:04:53Z

lib/matplotlib/_mathtext.py

+        for c in name:
+            if isinstance(c, Char):
+                c.font = "bf"
+                if str(c)[1] in allLatin or str(c)[2:-1] in smGreek:


str(c)[1:-1] is equal to c.c for Char, I think?

Suggested change

if str(c)[1] in allLatin or str(c)[2:-1] in smGreek:

if c.c in string.ascii_letters or c.c[1:] in lower_greek:

QuLogic · 2023-06-12T21:26:39Z

lib/matplotlib/_mathtext.py

+        smGreek = list(map(lambda a: unicodedata.name(chr(a))
+                           .split()[-1]
+                           .lower(),
+                           range(ord('\N{GREEK SMALL LETTER ALPHA}'),
+                                 ord('\N{GREEK SMALL LETTER OMEGA}') + 1)))


Should rename for PEP8, and maybe also prefer list comprehensions

Suggested change

smGreek = list(map(lambda a: unicodedata.name(chr(a))

.split()[-1]

.lower(),

range(ord('\N{GREEK SMALL LETTER ALPHA}'),

ord('\N{GREEK SMALL LETTER OMEGA}') + 1)))

lower_greek = [

unicodedata.name(chr(a)).split()[-1].lower()

for a in range(ord('\N{GREEK SMALL LETTER ALPHA}'),

ord('\N{GREEK SMALL LETTER OMEGA}') + 1))]

QuLogic · 2023-06-19T17:00:50Z

doc/users/next_whats_new/boldsym_mathtext.rst

+Boldsymbol mathtext command ``\boldsymbol``
+-----------------------------------------------------


Should match text length to underline.

oscargus

Trying the image test code out on normal LaTeX, it seems like the single + should not be bold?

devRD · 2023-06-23T11:31:02Z

I checked the + using an online LaTeX document and with text.usetex=True. It looks like a bold + when I compare it with \mathrm{+}. I got the following figure using text.usetex=True

oscargus · 2023-06-23T12:02:37Z

OK! I guess my point really was that the + in \boldsymbol{\Gamma + \Omega} looks different compared to \boldsymbol{+}, but it seems like I draw the wrong conclusion about the reason.

Is it really different or is there just some pixel snapping issue in the test image that makes it look that way?

devRD · 2023-06-24T14:33:34Z

Good catch!
Looking further into this, it is different. The \Gamma + \Omega seems to parse the + in this context as an Hlist and not an instance of Char so the state.font is not applying.

oscargus · 2023-06-26T08:54:44Z

lib/matplotlib/_mathtext.py

+            if isinstance(c, Hlist):
+                k = c.children[1]
+                if isinstance(k, Char):
+                    k.font = "bf"
+                    k._update_metrics()
+            if isinstance(c, Char):
+                c.font = "bf"
+                if c.c in latin_alphabets or c.c[1:] in small_greek:
+                    c.font = "bfit"
+                    c._update_metrics()
+                c._update_metrics()
+                hlist.append(c)
+            else:
+                hlist.append(c)


Suggested change

if isinstance(c, Hlist):

k = c.children[1]

if isinstance(k, Char):

k.font = "bf"

k._update_metrics()

if isinstance(c, Char):

c.font = "bf"

if c.c in latin_alphabets or c.c[1:] in small_greek:

c.font = "bfit"

c._update_metrics()

c._update_metrics()

hlist.append(c)

else:

hlist.append(c)

if isinstance(c, Hlist):

k = c.children[1]

if isinstance(k, Char):

k.font = "bf"

k._update_metrics()

elif isinstance(c, Char):

c.font = "bf"

if c.c in latin_alphabets or c.c[1:] in small_greek:

c.font = "bfit"

c._update_metrics()

c._update_metrics()

hlist.append(c)

oscargus · 2023-06-26T08:56:40Z

lib/matplotlib/_mathtext.py

+        state = self.get_state()
+        small_greek = [unicodedata.name(chr(i)).split()[-1].lower() for i in
+                           range(ord('\N{GREEK SMALL LETTER ALPHA}'),
+                                 ord('\N{GREEK SMALL LETTER OMEGA}') + 1)]


Maybe one should move this to be a member of the parser class so that it is not recreated every time?

(And considering that it is only used for testing membership, make it a set?)

oscargus · 2023-06-26T08:57:33Z

Nice that you got it to work! Just some minor suggestions for clarity and performance.

oscargus added status: waiting for other PR topic: text/mathtext labels Apr 11, 2023

oscargus reviewed Apr 11, 2023

View reviewed changes

lib/matplotlib/_mathtext.py Outdated Show resolved Hide resolved

oscargus approved these changes Apr 11, 2023

View reviewed changes

oscargus reviewed Apr 11, 2023

View reviewed changes

anntzer reviewed Apr 20, 2023

View reviewed changes

devRD force-pushed the boldsym branch 3 times, most recently from 7371e42 to 1447b6d Compare April 21, 2023 16:11

devRD marked this pull request as ready for review April 26, 2023 04:20

QuLogic removed the status: waiting for other PR label Apr 26, 2023

devRD requested review from anntzer and ksunden May 1, 2023 11:28

oscargus reviewed May 1, 2023

View reviewed changes

lib/matplotlib/tests/test_mathtext.py Outdated Show resolved Hide resolved

anntzer reviewed May 1, 2023

View reviewed changes

devRD requested review from oscargus and anntzer May 10, 2023 04:30

ksunden mentioned this pull request Jun 7, 2023

[ENH]: Support for \boldsymbol #25643

Closed

github-actions bot added the status: needs rebase label Jun 10, 2023

devRD force-pushed the boldsym branch from 120a9f1 to 5b27064 Compare June 12, 2023 13:40

github-actions bot removed the status: needs rebase label Jun 12, 2023

QuLogic reviewed Jun 12, 2023

View reviewed changes

devRD force-pushed the boldsym branch 2 times, most recently from 6e7850e to 3ac9e61 Compare June 13, 2023 09:10

QuLogic approved these changes Jun 19, 2023

View reviewed changes

devRD force-pushed the boldsym branch from 3ac9e61 to e56ce6c Compare June 19, 2023 17:43

oscargus reviewed Jun 22, 2023

View reviewed changes

devRD force-pushed the boldsym branch 4 times, most recently from dd805ef to 10a1ad7 Compare June 26, 2023 08:44

oscargus reviewed Jun 26, 2023

View reviewed changes

devRD force-pushed the boldsym branch 2 times, most recently from 2cfd449 to 05a3261 Compare June 26, 2023 15:42

github-actions bot added the status: needs rebase label Jun 29, 2023

Add boldsymbol functionality in mathtext

1de8ffb

devRD force-pushed the boldsym branch from 05a3261 to 1de8ffb Compare July 4, 2023 09:46

github-actions bot removed the status: needs rebase label Jul 4, 2023

ksunden merged commit fa68f46 into matplotlib:main Jul 11, 2023

ksunden added this to the v3.8.0 milestone Jul 11, 2023

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

boldsymbol support for mathtext #25661

boldsymbol support for mathtext #25661

devRD commented Apr 11, 2023 •

edited

Loading

oscargus left a comment •

edited

Loading

oscargus Apr 11, 2023

devRD Apr 12, 2023

ksunden Apr 20, 2023

anntzer Apr 20, 2023 •

edited

Loading

devRD May 1, 2023

ksunden commented Apr 25, 2023

devRD commented Apr 26, 2023

anntzer May 1, 2023

anntzer commented May 2, 2023 •

edited

Loading

devRD commented May 2, 2023

QuLogic Jun 12, 2023

QuLogic Jun 12, 2023 •

edited

Loading

QuLogic Jun 12, 2023

QuLogic Jun 19, 2023

oscargus left a comment

devRD commented Jun 23, 2023

oscargus commented Jun 23, 2023

devRD commented Jun 24, 2023 •

edited

Loading

oscargus Jun 26, 2023

oscargus Jun 26, 2023

oscargus commented Jun 26, 2023

	if str(c)[1] in allLatin or str(c)[2:-1] in smGreek:
	if c.c in string.ascii_letters or c.c[1:] in lower_greek:

		Boldsymbol mathtext command ``\boldsymbol``
		-----------------------------------------------------

boldsymbol support for mathtext #25661

boldsymbol support for mathtext #25661

Conversation

devRD commented Apr 11, 2023 • edited Loading

PR Summary

PR Checklist

oscargus left a comment • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

anntzer Apr 20, 2023 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

ksunden commented Apr 25, 2023

devRD commented Apr 26, 2023

Choose a reason for hiding this comment

anntzer commented May 2, 2023 • edited Loading

devRD commented May 2, 2023

Choose a reason for hiding this comment

QuLogic Jun 12, 2023 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

oscargus left a comment

Choose a reason for hiding this comment

devRD commented Jun 23, 2023

oscargus commented Jun 23, 2023

devRD commented Jun 24, 2023 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

oscargus commented Jun 26, 2023

devRD commented Apr 11, 2023 •

edited

Loading

oscargus left a comment •

edited

Loading

anntzer Apr 20, 2023 •

edited

Loading

anntzer commented May 2, 2023 •

edited

Loading

QuLogic Jun 12, 2023 •

edited

Loading

devRD commented Jun 24, 2023 •

edited

Loading