MathML tweaks by mmatera · Pull Request #1661 · Mathics3/mathics-core

mmatera · 2026-01-29T02:45:33Z

This PR makes some tweaks on how mathml output is render.

adds line breaks following the WMA usage.
Handle more carefully InterpretationBox to work with InputForm and OutputForm.
Remove OutputForm[s_String] format rule.
update format-tests-WMA.yaml with more tests compared against WS.
Update format-tests.yaml accordingly.

rocky · 2026-01-29T12:18:29Z

mathics/format/render/mathml.py

-    "\u2146",
+    "\u301a",  # [[
+    "\u301b",  # ]]
+    "\u00d7",  # \[Times]


Why can't this be picked up from character tables?

Good question. I have to check if we are not doing already that.

rocky · 2026-01-29T12:23:14Z

mathics/format/render/mathml.py

-    "\u2062",
-    "\u222b",
-    "\u2146",
+    "\u301a",  # [[


Recently, RightDoubleBracket and LeftDoubleBracket were tagged as operators. I suspect that these can now be picked up from the character tables.

rocky · 2026-01-29T12:25:26Z

mathics/format/render/mathml.py

    _options.update(options)
    options = _options
-    return "<mfrac>%s %s</mfrac>" % (
+    return "<mfrac>\n%s\n %s\n</mfrac>" % (


Is there a typographic difference here, or is this simply to make human comprehension easier when looking at the text?

It is just human readability, and helps to compare with WMA output, which split the output in lines. Also, WMA uses indentation for the different levels, which we are not supporting right now. For example

In[1]:= a^b/c//MathMLForm Out[1]//MathMLForm= <math> <mfrac> <msup> <mi>a</mi> <mi>b</mi> </msup> <mi>c</mi> </mfrac> </math>

while in Mathics3

In[1]:= a^b/c//MathMLForm Out[1]//MathMLForm= <mfrac> <msup> <mi>a</mi> <mi>b</mi> </msup> <mi>c</mi> </mfrac>

Having the output split in lines, we can add the indentation later, or remove it from the WMA output for doing the comparison.

It is just human readability, and helps to compare with WMA output, which split the output in lines. Also, WMA uses indentation for the different levels, which we are not supporting right now. For example

In[1]:= a^b/c//MathMLForm Out[1]//MathMLForm= <math> <mfrac> <msup> <mi>a</mi> <mi>b</mi> </msup> <mi>c</mi> </mfrac> </math>

while in Mathics3

In[1]:= a^b/c//MathMLForm Out[1]//MathMLForm= <mfrac> <msup> <mi>a</mi> <mi>b</mi> </msup> <mi>c</mi> </mfrac>

Ok. Got it. I was not opposed to it, but just wondered about the intention. (BTW, adding a comment to this effect will help others from wondering or thinking about this.)

The question is where to leave the comment. Maybe in the head of the file?

I did realize that indentation is something that can be done on the fly. The question is if we want to do it. I had my doubts about adding the line breaks, because -differently from WMA graphics interface- we use mathml form to render output. So, adding complexity and length to the output makes harder to the browser show the output.

On the other hand, adding the line breaks makes easier the comparison between Mathics3 and WMA results. So, I proposed an intermediate format.

It's okay to discuss approaches before coding and PR's.

OK, so, thoughts?

BoxExpression should have an integer "nesting_level" attribute, and that value times the string value of an indent is what prefaces the tags after a "\n".

INDENT_SPACES = " " # or "\t" or whatever (f"\n{self.nesting_level * INDENT_SPACES}).join(...)

OK, but the first thing is that we want to indent the mathml code. A way to achieve that is to use an argument for the render functions tracking the indentation level of the context. If we agree that it a good idea, I can implement it.

If it is set in a BoxExpresssion object, MathML rendering can pick up this attribute using its self parameter.

…utput

rocky · 2026-01-29T17:19:24Z

mathics/format/render/mathml.py

-            for line in text.split("\n"):
-                outtext += render("<mtext>%s</mtext>", line)
-            return outtext
+            return "".join(


This is where getting the proper indent level would be done.

…oxExpressions.

…utput

…oxExpressions.

Also: remove "_" from non-private classes. Go over render docstrings.

Initialize: * BoxExpression.boxes * GraphicsBox.boxwidth * GraphicsBox.boxheight * GrapnicsBox.boxes Make sure to convert sympy.Float to Python float

mathics/format/render/mathml.py

Co-authored-by: R. Bernstein <rocky@users.noreply.github.com>

rocky · 2026-02-07T14:52:54Z

@mmatera What the test expects is ⅆ (U2146), but what the named character encoded is 𝑑 (U0001D451).

Notes in the YAML say that the latter is what we want.

So do we adjust the test expectation or change to use 2146, which I think we decided looks worse?

mathics/format/render/mathml.py

This code does not follow the comment about how this turns into invisible space. When this was \u2146, that might have been true, but we use for DifferentialD the value U0001D451

mmatera · 2026-02-07T19:05:23Z

Superseed by #1665

mmatera · 2026-02-07T21:43:54Z

@mmatera What the test expects is ⅆ (U2146),

I would keep this one (ⅆ). Let's adjust it in another round.

but what the named character encoded is 𝑑 (U0001D451).

Notes in the YAML say that the latter is what we want.

So do we adjust the test expectation or change to use 2146, which I think we decided looks worse?

rocky · 2026-02-07T22:04:18Z

@mmatera What the test expects is ⅆ (U2146),

I would keep this one (ⅆ). Let's adjust it in another round.

We don't have a named character for this. How does this get into the output then?

mmatera · 2026-02-07T23:36:24Z

@mmatera What the test expects is ⅆ (U2146),

I would keep this one (ⅆ). Let's adjust it in another round.

We don't have a named character for this. How does this get into the output then?

Maybe '[DifferentialD]' must be associated to 'ⅆ' in Mathics-Scanner.

rocky · 2026-02-07T23:43:22Z

@mmatera What the test expects is ⅆ (U2146),

I would keep this one (ⅆ). Let's adjust it in another round.

We don't have a named character for this. How does this get into the output then?

As I wrote before:

Notes in the YAML say that the latter is what we want.

mmatera · 2026-02-07T23:52:56Z

In that case, OK, let's update the renders to use the U0001D451 character.

rocky · 2026-02-08T00:05:02Z

In that case, OK, let's update the renders to use the U0001D451 character.

That's already done automatically. That's why we got the test failure.

Improve processing of MathMLForm for Input and Output Form

95d09f2

mmatera force-pushed the mathml_tweaks branch from 672183d to 95d09f2 Compare January 29, 2026 09:51

remove trailing print

842fc58

rocky reviewed Jan 29, 2026

View reviewed changes

mmatera added 2 commits January 29, 2026 10:58

remove space from extra_operators, and add a comment

466bc6e

add a comment about mathml multiline output. Give a more consistent o…

8f0f52b

…utput

rocky reviewed Jan 29, 2026

View reviewed changes

mmatera added 2 commits January 30, 2026 17:59

Merge branch 'master' into mathml_tweaks

9e816d2

Merge branch 'master' into mathml_tweaks

81689de

This was referenced Feb 1, 2026

Mathml tweaks indent #1665

Merged

MathML rendering refactor #1671

Closed

mmatera added 8 commits February 2, 2026 10:26

improve Expression.sameQ to deal with BoxExpressions. Add tests for B…

a5e3104

…oxExpressions.

add a test involving FractionBox

270f82d

Merge branch 'improve_Expression_sameQ' into mathml_tweaks

cb3ff7a

Improve processing of MathMLForm for Input and Output Form

70f84ed

remove trailing print

5b89c3f

remove space from extra_operators, and add a comment

e88108d

add a comment about mathml multiline output. Give a more consistent o…

5295f48

…utput

improve Expression.sameQ to deal with BoxExpressions. Add tests for B…

1831e6b

…oxExpressions.

rocky force-pushed the mathml_tweaks branch from 798fc80 to 1831e6b Compare February 3, 2026 12:16

rocky and others added 6 commits February 3, 2026 09:54

Add Box type annotations to render routines.

6c529bd

Also: remove "_" from non-private classes. Go over render docstrings.

Fix type errors found by mypy

3c53f98

Initialize: * BoxExpression.boxes * GraphicsBox.boxwidth * GraphicsBox.boxheight * GrapnicsBox.boxes Make sure to convert sympy.Float to Python float

More mypy/type work.

942752a

Merge branch 'master' into mathml_tweaks

6d25078

merge

bddeb7a

merge

e57789a

mmatera force-pushed the mathml_tweaks branch from 5851ea3 to e57789a Compare February 5, 2026 10:50

merge

17413ed

mmatera and others added 6 commits February 6, 2026 11:12

fix

1e50e58

merge

262b628

merge

76bee53

fix mypy

5488a13

ruff

684e474

Merge branch 'master' into mathml_tweaks

7dbc388

rocky reviewed Feb 7, 2026

View reviewed changes

mathics/format/render/mathml.py Outdated Show resolved Hide resolved

rocky reviewed Feb 7, 2026

View reviewed changes

mathics/format/render/mathml.py Outdated Show resolved Hide resolved

mmatera and others added 3 commits February 7, 2026 10:07

Update mathics/format/render/mathml.py

351eee1

Co-authored-by: R. Bernstein <rocky@users.noreply.github.com>

Update mathics/format/render/mathml.py

18abb93

Co-authored-by: R. Bernstein <rocky@users.noreply.github.com>

fix syntactic typo

8e17b74

rocky reviewed Feb 7, 2026

View reviewed changes

mathics/format/render/mathml.py Outdated Show resolved Hide resolved

Apply suggestion from @rocky

42c6226

This code does not follow the comment about how this turns into invisible space. When this was \u2146, that might have been true, but we use for DifferentialD the value U0001D451

mmatera closed this Feb 7, 2026

mmatera deleted the mathml_tweaks branch February 7, 2026 19:09

Uh oh!

Conversation

mmatera commented Jan 29, 2026

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

rocky Jan 29, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

rocky commented Feb 7, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Uh oh!

mmatera commented Feb 7, 2026

Uh oh!

mmatera commented Feb 7, 2026

Uh oh!

rocky commented Feb 7, 2026

Uh oh!

mmatera commented Feb 7, 2026

Uh oh!

rocky commented Feb 7, 2026

Uh oh!

mmatera commented Feb 7, 2026

Uh oh!

rocky commented Feb 8, 2026 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

rocky Jan 29, 2026 •

edited

Loading

rocky commented Feb 7, 2026 •

edited

Loading

rocky commented Feb 8, 2026 •

edited

Loading