annotations: Crash on invalid move annotations

[junghon3709]

I have the following annofen file:

v1 r1bakab1r/4c4/4c4/p1[n<a+1-2>]3p1p/9/3CC1P2/P1P1P3P/2N3N2/4A4/2BAK1B1R w - - 0 12

This renders successfully.

However, should I change the numbers in the arrows to:

v1 r1bakab1r/4c4/4c4/p1[n<a+1-3>]3p1p/9/3CC1P2/P1P1P3P/2N3N2/4A4/2BAK1B1R w - - 0 12,

an error message pops up:

Traceback (most recent call last):
  File "`PATH`", line 8, in <module>
    sys.exit(main())
  File "`PATH`/__main__.py", line 260, in main
    run(options)
  File "`PATH`/__main__.py", line 60, in run
    atoms_to_put = list(iterate_annofen_tokens(content))
  File "/home/~/Documents/Projects/Xiangqi-Analyser/venv/lib/python3.8/site-packages/xiangqi_setup/file_formats/annofen.py", line 91, in iterate_annofen_tokens
    raise ValueError(f'Malformed annoFEN token: {character!r}')
ValueError: Malformed annoFEN token: '['

[hartwork]

Hi @junghon3709 nice to meet you!

I confirm the issue, here's what I did to reproduce based on your data on the terminal:

Click here to expand for details

# cat issue-113.py 
from xiangqi_setup.file_formats.annofen import iterate_annofen_tokens, _escaped_annotation_names

with open('issue-113.annofen') as f:
   list(iterate_annofen_tokens(f.read()))

# cat issue-113.annofen 
v1 r1bakab1r/4c4/4c4/p1[n<a+1-3>]3p1p/9/3CC1P2/P1P1P3P/2N3N2/4A4/2BAK1B1R w - - 0 12

# python3.10 issue-113.py 
Traceback (most recent call last):
  File "[..]/xiangqi-setup/issue-113.py", line 4, in <module>
    list(iterate_annofen_tokens(f.read()))
  File "[..]/xiangqi-setup/xiangqi_setup/file_formats/annofen.py", line 92, in iterate_annofen_tokens
    raise ValueError(f'Malformed annoFEN token: {character!r}')
ValueError: Malformed annoFEN token: '['

What is happening here is that "+1-3" is not a valid move to the eyes of xiangqi-setup and the way the parser works makes it end up with this admittedly unhelpful-to-humans error. For each supported move there is one SVG file, e.g. see this directory. If a move is legal to any of the pieces, it is supported, e.g. "a+1-2" is one out of eight for the knights/horses. The full list is this:

+0+1 +0-1 +0+2 +0-2 +0+3 +0-3 +0+4 +0-4 +0+5 +0-5 +0+6 +0-6 +0+7 +0-7 +0+8 +0-8 +0+9 +0-9 +1+0 -1+0 +1+1 +1-1 -1+1 -1-1 +1+2 +1-2 -1+2 -1-2 +2+0 -2+0 +2+1 +2-1 -2+1 -2-1 +2+2 +2-2 -2+2 -2-2 +3+0 -3+0 +4+0 -4+0 +5+0 -5+0 +6+0 -6+0 +7+0 -7+0 +8+0 -8+0

I have some hope that "+1-3" was in part a typo and that legal moves are enough for your use case? Then I would only need to improve the error message. I have some idea how to do that. What do you think, and what was your intention when trying "+1-3"?

Best, Sebastian

[hartwork]

@junghon3709 what do you think?

[hartwork]

@junghon3709 what do you think?

[junghon3709]

I see now, I guess it makes sense to change the error message

[hartwork]

With #123 merged, the output would say…

xiangqi_setup.annotations.InvalidAnnotationCode: Invalid annotation code 'a+1-3'

…now.

[hartwork]

Version 2.2.0 with #123 merged released to GitHub and PyPI just now.

[junghon3709]

Thank you for the change!