UTIL: PGN Translator already written?

[Dann Corbit]

The latest SSDF games come with a weird mish-mash of nomenclature in them. They start off with English PGN, but then switch off to a jumble of Swedish (I assume) piece notation but it's an odd format that I don't remember seeing before. For example, OOO is obviously long castle, but not in any standards based format that I am familiar with.

Is there (already written, I am not asking anyone to write this for me) a utility that will convert stuff like "Sf6xe4 Le2f3 d6d5" into standard PGN?

Part of a sample game:

Code: Select all

[Event "SSDF"]
[Site "TONY-DATOR"]
[Date "2018.05.20"]
[Round "1"]
[White "Stockfish 8 x64 Q6600"]
[Black "Stockfish 9 x64 1800X"]
[Result "1/2-1/2"]
[ECO "B47"]
[PlyCount "229"]
[EventDate "2018.??.??"]
[TimeControl "40/7200:40/7200:40/7200"]

1. e4 c5 2. Nf3 e6 3. d4 cxd4 4. Nxd4 Nc6 5. Nc3 d6 6. Be3 Bd7 7. g4 {[%eval
48,30] (g2g4 a7a6 h2h4 h7h6 f2f4 Ta8c8 Sd4xc6 Tc8xc6 Dd1d3 Sg8f6 Lf1e2 Tc6xc3
b2xc3 Sf6xe4 Le2f3 d6d5 Lf3xe4 d5xe4 Dd3d4 Ld7c6 Dd4xd8+ Ke8xd8 c3c4 Kd8c7
Ta1b1 Lf8e7 Le3b6+ Kc7c8 Ke1e2 Lc6a4 Ke2d2 e4e3+ Kd2d3 e3e2 Kd3xe2 La4xc2) 388}
a6 8. h4 {[%eval 49,31] (h2h4 Sg8e7 a2a3 Sc6xd4 Dd1xd4 Se7c6 Dd4d1 b7b5 g4g5
h7h6 g5xh6 g7xh6 Th1g1 Dd8xh4 Dd1e2 Ta8c8 OOO Sc6e5 f2f4 Se5c4 Kc1b1 h6h5
Le3d4 e6e5 f4xe5 Ld7g4 Tg1xg4 Dh4xg4 e5xd6 Dg4xe2 Lf1xe2 Th8h6 e4e5 Th6e6
Sc3d5 Sc4xe5 Sd5c7+ Ke8d7 Sc7xe6) 70} Nge7 9. Nb3 {[%eval 40,31] (Sd4b3 b7b5
a2a3 Se7c8 Dd1e2 Sc8b6 OOO Sb6c4 Le3g5 f7f6 Lg5f4 Lf8e7 Lf4g3 OO f2f4 Sc6a5
h4h5 h7h6 Kc1b1 Ta8b8 Lg3e1 Sa5xb3 c2xb3 Sc4a5 b3b4 Sa5c4 De2f3 Dd8c7 Lf1e2
a6a5 Le1f2 a5xb4 a3xb4 Tf8c8) 332} b5 10. a3 {[%eval 26,31] (a2a3 Sc6a5 Sb3d2
Ta8b8 g4g5 h7h5 g5xh6/ep g7xh6 Sc3a2 Se7g6 Sa2b4 Th8g8 Sb4xa6 Tb8c8 c2c3
Sg6xh4 Dd1h5 Sh4g2+ Lf1xg2 Tg8xg2 OOO Dd8f6 Kc1b1 Sa5c4 Sd2xc4 b5xc4 Sa6b4
Df6g6 Dh5xg6 Tg2xg6 Td1g1 Tg6xg1+ Th1xg1 Ld7a4 Tg1g8 Ke8d7) 246} Nc8 11. Nd4 {
[%eval 48,28] (Sb3d4 Lf8e7 g4g5 OO f2f4 Sc8b6 Dd1d2 Sc6a5 b2b3 d6d5 h4h5 d5xe4
h5h6 g7g6 Sc3xe4 Sb6d5 Sd4e2 Sd5xe3 Dd2xe3 Dd8c7 Lf1g2 Ta8c8 Se2c3 f7f6 OO
Ld7c6 Ta1d1 f6f5 Se4f6+ Le7xf6 De3xe6+ Dc7f7 De6xf6) 111} h6 12. Be2 {[%eval
58,28] (Lf1e2 Ta8b8 Dd1d2 Dd8c7 Sd4xc6 Dc7xc6 b2b3 Lf8e7 Le3d4 a6a5 b3b4 e6e5
Ld4e3 Le7d8 Ta1d1 Sc8e7 OO Tb8b7 f2f4 a5xb4 a3xb4 Se7g6 f4f5 Sg6f4 Le3xf4
e5xf4 Sc3d5 Ld8xh4 Dd2xf4 Lh4g5 Df4f2 Dc6xc2 f5f6) 137} Rb8 13. Qd2 {[%eval 37,
33] (Dd1d2 Sc6xd4 Le3xd4 a6a5 f2f4 Sc8e7 b2b3 b5b4 a3xb4 a5xb4 Sc3a4 Se7c6
Ld4e3 e6e5 f4f5 Lf8e7 OOO Sc6d4 Le3xd4 e5xd4 g4g5 Dd8a5 Dd2xd4 h6xg5 Dd4xg7
Da5e5 Dg7xe5 d6xe5 h4xg5 Th8xh1 Td1xh1 Le7xg5+ Kc1b1 Ke8e7 Th1h7 Ld7c6 Le2c4
Tb8f8 Sa4c5 Ke7d6) 187} N8e7 14. f4 {[%eval 56,31] (f2f4 Sc6xd4 Le3xd4 a6a5

[Dann Corbit]

On second glance, I may already be OK with my existing parser.
It appears that the unusual notation is contained in the comments.

[hgm]

Looks like German (but it could be that Swedish is the same).

WinBoard has an option -pieceNickNames that can be used to define a set of alternative IDs during SAN or FEN parsing. E.g. specifying -pieceNickNames ".SLTD..sltd." would make it recognize S as Knight, L as Bishop, etc. It would still recognize the usual piece IDs, as long these are not eclipsed by any of the nick-names.

I use this in Xiangqi, where the official WXF standard is Horse = H and Elephant = E, but where one often encounters games that use Elephant = B and Horse = N.