Enable :widths: option in table directives.

[jhcole]

Closes #793

The output from myst-docutils-html5 differs from the results I'm getting from jupyter-book build and I'm still tracking down where the difference are coming from.

There are 4 possible states for the width option of a table directive (not set, "auto", "grid", or a list of integers).

The results from myst-docutils-html5 are as follows:

Option	Previous behavior	New behavior
Not set (default)	no class or colspecs	no class or colspecs
auto	no class or colspecs	no class or colspecs
grid	no class or colspecs	Add colgroup with column widths set equal to each other.
a list of integers	no class or colspecs	Add colgroup with column widths set in proportion to the integers given.

The results from jupyter-book build are as follows:

Option	Previous behavior	New behavior
Not set (default)	Add the class "colwidths-auto" to the table and not set any column widths.	Set equal proportion widths to each column. This is a regression I hope to fix.
auto	Add the class "colwidths-auto" to the table and not set any column widths.	Add the class "colwidths-auto" to the table and not set any column widths (no change in behavior).
grid	Add the classes "colwidths-auto" and "colwidths-given" to the table, and not set any column widths.	Add the class "colwidths-given" to the table, and set equal widths to each column. The documentation claims this option will set widths based on column contents, but it just sets them all equal.
a list of integers	Add the classes "colwidths-auto" and "colwidths-given" to the table, and not set any column widths.	Add the class "colwidths-given" to the table and set columns widths in proportion to the integers given.

[codecov]

Codecov Report

Patch coverage has no change and project coverage change: -0.01 ⚠️

Comparison is base (350c633) 90.26% compared to head (7cf70b4) 90.26%.

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #795      +/-   ##
==========================================
- Coverage   90.26%   90.26%   -0.01%     
==========================================
  Files          23       23              
  Lines        2970     2969       -1     
==========================================
- Hits         2681     2680       -1     
  Misses        289      289              

Flag	Coverage Δ
pytests	90.26% <ø> (-0.01%)	⬇️

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
myst_parser/mdit_to_docutils/base.py	93.51% <ø> (-0.01%)	⬇️

☔ View full report in Codecov by Sentry.
📢 Do you have feedback about the report comment? Let us know in this issue.

[chrisjsewell]

The output from myst-docutils-html5 differs from the results I'm getting from jupyter-book build and I'm still tracking down where the difference are coming from.

To note, this shouldn't have anything to do with jupyter-book, since jupyter-book is just a thin wrapper on top of sphinx+myst-parser, that does not change its behaviour
If there is a difference, it would either be between docutils vs. sphinx behaviour, or a change in behaviour between versions of docutils/sphinx.

This is the code for the table directive: https://github.com/live-clones/docutils/blob/6de53a0de5415174d58e775110d89e13dd76fc83/docutils/docutils/parsers/rst/directives/tables.py#L134

This is where it used in the HTML conversion: https://github.com/live-clones/docutils/blob/6de53a0de5415174d58e775110d89e13dd76fc83/docutils/docutils/writers/_html_base.py#L686

[goroderickgo]

Hi, I saw the linked issue and was wondering if there was a roadmap to have this PR get merged into a release? Thanks!

[jhcole]

Hi, I saw the linked issue and was wondering if there was a roadmap to have this PR get merged into a release? Thanks!

Sorry, I haven't had time to finish this. It will likely be a few months before I have time to focus on it again. You're welcome to take it over.

[berlin2123]

It works In my sphinx, If one just only change the code in /usr/local/lib/python3.<version>/dist-packages/myst_parser/mdit_to_docutils/base.py as this PR.

i.e., delete or comment the line 1352:

        #table["classes"] += ["colwidths-auto"]

[dbitouze]

@chrisjsewell Too bad this issue isn't fixed :)

It may be useful to note that the :widths: option is ignored with the {table} directive, but not with the {list-table} directive, as can be verified with the following code, for example in the Live Preview:

# With `{list-table}`

:::{list-table}
:widths: 30 30 30
:header-rows: 1

*   - Header 1
    - Header 2
    - Header 3
*   - Cell 1-1
    - Cell 1-2
    - Cell 1-3
*   - Cell 2-1
    - Cell 2-2
    - Cell 2-3 long long long long long long long
:::

# With `{table}`

:::{table} 
:width: 100
:widths: 30 30 30

| Header 1 | Header 2 | Header 3                                    |
|----------|----------|---------------------------------------------|
| Cell 1-1 | Cell 1-2 | Cell 1-3                                    |
| Cell 2-1 | Cell 2-2 | Cell 2-3 long long long long long long long |
:::

[jduf]

Any update on this issue?