Enable automatic word wrap based on desired maximum column width

[magelisk]

This address as submitted issue #90 that allows automatically word wrapping long text string. From an API standpoint, this is automatically enabled by simply specifying the maximum column width for a given column. I selected max column width instead of just column width to remove any confusion where a user might think that specifying colwidth might be setting a guaranteed size, which is not the behavior here (though I could see it as a possible future update?)

The "magic" here is a customized version of the textwrap.TextWrapper class to not use just len for strings and to also preserve ANSI color codes around wrapped lines. This is a non-trivial bit of code (even though heavily copy/pasted) with possible regression concerns, so I put it into a new test file to help tack this closely. Hopefully this fits within this projects idioms.

To address the 'tests that need to be covered' from the issue discussion:

• colwidths parameter can be specified as a number (the same for all column)
  • Done
• colwidths parameter can be shorter than the number of column (don't wrap text in the remaining columns)
  • Done
• the feature works when all columns are text
  • Done
• the feature works when some columns are numeric
  • Done - I think. I think it behaves as sensibly as it can (ignores numbers), and it also works with disable_numparse. Let me know if you had something else in mind,
• the feature works when some columns contain wide characters (Japanese/Chinese/Korean)
  • Done
• the feature works when some columns contain invisible control characters (ASCII escape codes, color, hyperlinks)
  • Done
• the feature works reasonably well or is disabled in formats which cannot support it (HTML? LaTeX?)
  • This has to be explicitly enabled by setting maxcolwidth, so I believe this meets the ability to ensure it does not impact these formats
• when bidirectional script support is implemented first, this feature should work correctly when there's mixed left-to-right and right-to-left text in a cell (English + Arabic/Hebrew/...)
  • N/A

Here's a small sample program I used to quickly demonstrate all the big capability updates

from tabulate import tabulate
all_data = [
    [('1', 'John Smith', '\033[31mThis is a rather long description that might look better if it is wrapped a bit\033[0m')],
    [('1', 'John Smith', '\033[31m약간 감싸면 더 잘 보일 수있는 다소 긴 설명입니다 설명입니다 설명입니다 설명입니다 설명\033[0m')],
    [('1', 'John Smith', '\033[31m\033[43mThis is a rather long description that might look better if it is wrapped a bit\033[0m')],
    [('1', 'John Smith', 'This is a rather \033[31mlong description\033[0m that might look better \033[93mif it is wrapped\033[0m a bit')]
]
for data in all_data:
    headers=("Issue Id", "Author", "Description")
    print(tabulate(data,  headers, tablefmt="grid", maxcolwidths=[None, None, 30]))

[astanin]

The pull request looks good, nothing that catches my eye. I'm merging it. Thank you, Matt.

[tejasjadhav]

Any idea which version will this change get published in?

[astanin]

@tejasjadhav Next version (0.8.10). There is not much else to release yet. You may install the git version if you want.

[tejasjadhav]

@astanin understood 👍 I'll use the Git version till then.

[csp197]

@astanin - I know its already been asked, but any chance that this PR can get pushed out soon?

[nickelser]

getting an error using the git version & maxcolwidths with datetime objects:

    print(tabulate.tabulate(result.table, headers="keys", tablefmt="psql", maxcolwidths=80))
  File "<...>/lib/python3.9/site-packages/tabulate.py", line 1592, in tabulate
    list_of_lists = _wrap_text_to_colwidths(
  File "<...>/python3.9/site-packages/tabulate.py", line 1233, in _wrap_text_to_colwidths
    wrapped = wrapper.wrap(cell)
  File "/usr/local/Cellar/python@3.9/3.9.1_8/Frameworks/Python.framework/Versions/3.9/lib/python3.9/textwrap.py", line 351, in wrap
    chunks = self._split_chunks(text)
  File "/usr/local/Cellar/python@3.9/3.9.1_8/Frameworks/Python.framework/Versions/3.9/lib/python3.9/textwrap.py", line 337, in _split_chunks
    text = self._munge_whitespace(text)
  File "/usr/local/Cellar/python@3.9/3.9.1_8/Frameworks/Python.framework/Versions/3.9/lib/python3.9/textwrap.py", line 154, in _munge_whitespace
    text = text.expandtabs(self.tabsize)
AttributeError: 'datetime.timedelta' object has no attribute 'expandtabs'

i think text needs a str() or similar.

[Barathwaja]

Hi @astanin ,
Is there an ETA when this feature will be released?

Thanks.

[AtomicNess123]

Has this been released?

[krpatter-intc]

Tried this out from source, and I noticed the header does not honor the maxcolwidths...is there any thoughts to perhaps supporting that capability as well?