Closed tomtseng closed 1 year ago
Name | Link |
---|---|
Latest commit | 367d448e14f7a2ca8306bb5040032aaf4a901f1f |
Latest deploy log | https://app.netlify.com/sites/goattack/deploys/63ed7ecc6ec36d000889eb07 |
Deploy Preview | https://deploy-preview-68--goattack.netlify.app/ |
Preview on mobile | Toggle QR Code...Use your smartphone camera to open QR code link. |
To edit notification comments on pull requests, go to your Netlify site settings.
Good catch. I remember having to do something to get those games formatted on one line too, originals look like below, which outright crashes prepare_data.
I think Nora prepared the files with these games for me back when I didn't have NAS access. So might be from different source/process than most. If we encounter this more we might want to adjust prepare_data so it processes them directly, but hopefully only one-time thing so the copy and sed solution seems reasonable. And sounds good on the "robustness to space" bit.
prepare_data.py
processes the comments in SGFs to make the adversary/victim predictions more human-readable. The games in the "Positions analyzed with varying visits" encounter a few issues with this comment processing:C[...]W[...]
instead ofW[...]C[...]
.prepare_data.py
did not process such comments.sed -i "s/ *\(C\[[^]\]*\]\) *\(W\[[a-z]\{,2\}]\) */\2\1/g" *.sgf
on them to replaceC[...]W[...]
withW[...]C[...]
.B[...] C[...]
(with a space in between) instead ofB[...] C[...]
.prepare_data.py
did not process these comments.prepare_data.py
to accept spaces betweenB[...]
andC[...]