Improve clarification for piece representation in infer_pieces.py

[nayan2167]

• Modified __PREDS_DICT in infer_pieces.py to use piece names as keys and piece Ids as values (e.g., {'B': 0, 'K': 1, 'N': 2, ...}).
• Updated the __max_piece() function in infer_pieces.py to access tops using the piece names instead of numeric indices (e.g., tops[0][0]["B"]).
• Updated the is_empty_square(square_probs) function in infer_pieces.py to utilize the modified __PREDS_DICT values.

These changes were made to improve the clarity and accuracy of piece representation within the infer_pieces.py script. The __PREDS_DICT dictionary now associates piece names with their respective Ids, allowing for more intuitive referencing. The __max_piece() function and is_empty_square() function has been adjusted accordingly to accommodate the updated piece representation, ensuring correct functionality.

[nayan2167]

Hi @davidmallasen please review the changes I made, I mistakenly created a new PR(#26) for this commit pardon me for that, I am closing it. also #25 is the same as this issue I would like to fix it, one more question should I create a fresh PR for that or can I commit to this branch?

[davidmallasen]

I added the review comments. No problem on opening and closing the other PR :)
Issue #25 now sould be related to this now as #24 was merged. If you would like to fix it, please go ahead and add those commits to this PR. I will add the trigger to fix issue #25 also when this PR is merged.

[nayan2167]

Hi @davidmallasen
__max_piece, __determine_promoted_piece, __determine_most_probable_white_piece, __determine_most_probable_black_piece, infer_chess_pieces these are the functions where I found the magic numbers and updated them with values corresponding to __PREDS_DICT if needs further changes let me know.

Here is the sample output

Promoted piece (__determine_promoted_piece): Q
----------------------------------------
Most probable White piece (__determine_most_probable_white_piece) : P
----------------------------------------
Most probable Blacj piece (__determine_most_probable_black_piece): n
----------------------------------------
Warning: the selected model is not accurate enough to predict a balanced board configuration
        Please consider providing the previous FEN, selecting a different model, or performing
        transfer learning on that model
Infer chess piece (infer_chess_pieces): ['N', 'q', 'R', 'p', '_', 'B', 'r', 'r', 'n', 'p', 'P', 'Q', 'b', 'R', 'N', 'N', 'r', 'N', 'p', 'q', 'P', 'P', 'P', 'B', 'n', 'n', 'k', 'P', '_', '_', 'P', 'B', 'n', 'r', 'n', 'N', 'P', 'p', 'Q', 'R', 'r', 'b', 'B', 'B', 'q', 'q', 'p', 'q', 'Q', 'r', 'p', 'B', 'n', 'p', 'p', 'b', 'K', 'B', 'b', 'N', 'q', 'p', 'R', '_']

[davidmallasen]

@nayan2167 Could you also add this for the kings (numbers 1 and 8) in

LiveChess2FEN/lc2fen/infer_pieces.py

Line 495 in 89591d9

white_king = max(pieces_probs_sort, key=lambda prob: prob[0][1])

and

LiveChess2FEN/lc2fen/infer_pieces.py

Line 497 in 89591d9

pieces_probs_sort, key=lambda prob: prob[0][8], reverse=True

?

[nayan2167]

@nayan2167 Could you also add this for the kings (numbers 1 and 8) in

LiveChess2FEN/lc2fen/infer_pieces.py

Line 495 in 89591d9

white_king = max(pieces_probs_sort, key=lambda prob: prob[0][1])

and

LiveChess2FEN/lc2fen/infer_pieces.py

Line 497 in 89591d9

pieces_probs_sort, key=lambda prob: prob[0][8], reverse=True

?

Done, also magic numbers removed from __sort_pieces_list

[davidmallasen]

Great, thanks @nayan2167 !