Chess Artist

A python script and exe file that can analyze games in the pgn file, annotate epd file, test engine on chess puzzles or problems and can generate chess puzzles.

A. Setup

Either use the windows exe file or use the python source and required modules.

1. Download the windows exe file

https://github.com/fsmosca/chess-artist/releases

2. Use python source and modules

• Install python >= 3.8

• Install python-chess

  new:

  • pip install python-chess

  upgrade:

  • pip install --force-reinstall chess

  or

  • pip install -r requirements.txt

B. Help

Use the command line: python chess_artist.py -h or see here.

C. Command lines

1. Analyze games in pgn file

a) Basic command line

python chess_artist.py --infile ./PGN/sample.pgn --outfile out_sample.pgn --enginefile ./Engine/stockfish/stockfish_12_x64_modern.exe --eval search --job analyze --movetime 1000

b) Add options to engine, use --engineoptions

python chess_artist.py --infile sample.pgn --outfile out_sample.pgn --enginefile Stockfish.exe --engineoptions "Hash value 128, Threads value 1" --eval search --job analyze --movetime 2000

Will analyze games inside sample.pgn and output the analyzed games in out_sample.pgn. It uses engine Stockfish.exe with Hash 128MB and Threads 1 at 2000 millisec or 2 sec analysis time per position. chess_artist.py and Stockfish.exe engine must be in same directory.

c) Analyze game with polyglot book

python chess_artist.py --infile sample.pgn --outfile out_sample.pgn --enginefile Stockfish.exe --engineoptions "Hash value 128, Threads value 1" --bookfile gm2600.bin --eval search --job analyze --movetime 2000

d) Analyze game with polyglot book in book dir at chess_artist folder

python chess_artist.py --infile sample.pgn --outfile out_sample.pgn --enginefile Stockfish.exe --engineoptions "Hash value 128, Threads value 1" --bookfile book/book.bin --eval search --job analyze --movetime 2000

e) Analyze between move 12 to move 20

python chess_artist.py --infile sample.pgn --outfile out_sample.pgn --enginefile Stockfish.exe --eval search --job analyze --movetime 2000 --movestart 12 --moveend 20

f) Analyze certain player name say Carlsen, Magnus

python chess_artist.py --infile sample.pgn --outfile out_sample.pgn --enginefile Stockfish.exe --eval search --job analyze --movetime 2000 --movestart 12 --moveend 20 --player "Carlsen, Magnus"

g) Analyze games of Carlsen, Magnus where he is black

python chess_artist.py --infile iommast19.pgn --outfile magnus_black.pgn --enginefile Stockfish.exe --eval search --job analyze --movetime 2000 --movestart 12 --moveend 20 --player "Carlsen, Magnus" --color black

h) Analyze lost games of Movsesian, Sergei, also analyze moves of its opponent, using --player-and-opp option

chess_artist.exe --infile iommast19.pgn --outfile sergie_loses.pgn --enginefile Stockfish.exe --eval search --job analyze --movetime 5000 --player-and-opp "Movsesian, Sergei" --loss

i) Sample analyzed Fischer Random Chess game

[Event "World Fischer Random 2019"]
[Site "Hovikodden NOR"]
[Date "2019.11.02"]
[Round "3.5"]
[White "Carlsen, Magnus"]
[Black "So, Wesley"]
[Result "1/2-1/2"]
[BlackElo "2767"]
[BlackFideId "5202213"]
[BlackTitle "GM"]
[EventDate "2019.10.04"]
[FEN "nrkbqnbr/pppppppp/8/8/8/8/PPPPPPPP/NRKBQNBR w HBhb - 0 1"]
[SetUp "1"]
[Variant "chess 960"]
[WhiteElo "2876"]
[WhiteFideId "1503014"]
[WhiteTitle "GM"]
[Annotator "engine: Stockfish 2019-12-30 64 POPCNT, prog: Chess Artist v2.9"]

{Hash 256mb, Threads 2, analysis 15.0s per position, move score is in pawn unit,
positive is good for white and negative is good for black}
1. Nb3 $6 {-0.16} (1. c3 {+0.80}) 1... f5 $6 {+0.46} (1...Nb6 2. e4 e5 3. c3 Ra8 {+0.00}) 
2. f3 $3 {-0.11, planning Bxa7} 2... Nb6 {+0.46} 
3. e4 {+0.00} 3... fxe4 {+0.39} 
4. fxe4 {-0.19, with a better piece mobility} 4... e5 {+0.40} 
5. Ne3 $6 {-0.33} (5. Bf3 Ne6 6. O-O-O d6 7. d3 {+0.39}) 5... Ne6 $1 {+0.51, followed by h5} 
6. c3 $6 {-0.44} (6. d3 {+0.41}) 6... Bg5 $6 {+0.71} (6...Ra8 {-0.45}) 
7. Bc2 $3 {+0.20, followed by h4} 7... O-O-O $6 {+0.64} (7...Bf7 8. h4 Bxe3 9. Bxe3 O-O-O {-0.09}) 
8. O-O-O $2 {-0.07} (8. a4 {+0.53}) 8... Bf7 $1 {+0.41, with the idea of Bh5} 
9. Kb1 {+0.00} (9. h4 Bxe3 10. Qxe3 d6 11. d3 {+0.45}) 9... Bh5 $3 {+0.17, followed by Bxd1} 
10. Rc1 {+0.00} 10... Rf8 {+0.11} (10...Kb8 11. h4 Bxe3 12. Bxe3 d6 {+0.00}) 
11. h4 {-0.13} (11. Nf5 {+0.21}) 11... Bxe3 $5 {+0.05} 
12. Bxe3 $5 {-0.27} 12... Nf4 $3 {+0.11, with the idea of Nxg2} 
13. Rg1 $6 {-0.45} (13. Bxf4 Rxf4 14. g3 Rf6 15. Rf1 {+0.17}) 13... Bg4 $2 {+0.00} (13...Ne2 14. Bg5 Nxg1 15. Qxg1 Kb8 {-0.59}) 
14. Bxf4 $6 {-0.49} (14. g3 Ne2 15. Bg5 Nxc1 16. Kxc1 {-0.10}) 14... Rxf4 {-0.01} 
15. g3 $6 {-0.47} (15. d4 d6 16. g3 Rf6 17. Nd2 {-0.10}) 15... Rf6 {+0.00} (15...Rf8 16. Rf1 Rxf1 17. Qxf1 d6 {-0.36}) 
16. d4 {-0.45} (16. Qe3 d6 17. Rgf1 Qf7 18. Rxf6 {-0.19}) 16... d6 $3 {-0.10, with the idea of Qb5} 
17. Bd3 $6 {-0.43} (17. Nd2 Qb5 18. Qe3 Qe2 19. Qg5 {+0.00}) 17... Kb8 $2 {-0.06} (17...Qf7 {-0.66}) 
18. Ka1 $6 {-0.51} (18. Qd2 Qe7 19. Qg5 h5 20. Rgf1 {-0.09}) 18... Qf7 {-0.15} (18...Rf3 19. Bc2 Nc4 20. Qe2 Ne3 {-0.44}) 
19. d5 $2 {-1.13} (19. Qe3 {-0.15}) 19... Rf3 $1 {-0.85} (19...Nd7 {-0.84}) 
20. Bb1 $2 {-1.38} (20. Qd2 {-0.38}) 20... Rf8 $2 {-0.58} (20...Nc4 21. Nd2 Nxd2 22. Qxd2 Rf8 {-1.30}) 
21. c4 {-0.69} 21... Nd7 $5 {-0.48} 
22. Qb4 $4 {-1.52} (22. c5 Nxc5 23. Nxc5 dxc5 24. Rxc5 {-0.47}) 22... b6 {-1.39} 
23. Na5 $4 {-2.57} (23. a3 {-1.13}) 23... Nc5 {-2.24} 
24. Nc6+ {-3.26} 24... Kb7 {-2.69} (24...Ka8 {-3.11}) 
25. Na5+ {-3.07} 25... Ka8 {-2.60} 
26. Nc6 {-3.47} 26... Kb7 $2 {-2.56} (26...Bd7 {-3.26}) 
27. Na5+ {-3.32} 27... Ka8 {-3.15} 28. Nc6 {-3.22} {WhiteBlunder=2, BlackBlunder=0, WhiteBad=3, BlackBad=4} 1/2-1/2

2. Test engine with test suite

Use movetime of 500ms

python chess_artist.py --infile ./EPD/wacnew.epd --outfile result_wacnew.txt --enginefile stockfish_10.exe --engineoptions "Hash value 128, Threads value 1" --eval search --job test --movetime 500

Output file result_wacnew.txt

:: EPD wacnew.epd TEST RESULTS ::
Engine        : Stockfish 10 64 POPCNT
Time/pos (sec): 0.5

Total epd lines       : 300
Total tested positions: 300
Total correct         : 274
Correct percentage    : 91.33

Use depth of 10 without movetime

python chess_artist.py --infile wacnew.epd --outfile result_wacnew.txt --enginefile stockfish_10.exe --engineoptions "Hash value 128, Threads value 1" --eval search --job test --depth 10 --movetime 0

3. Annotate the epd file

Use Stockfish to annotate epd file at movetime 30s or 30000ms

python2 chess_artist.py --infile repertoire.epd --outfile out_repertoire.epd --enginefile stockfish_10.exe --engineoptions "Hash value 128, Threads value 1" --eval search --job analyze --movetime 30000

Example output:

rn1q1rk1/pbp2ppp/1p1ppn2/6B1/2PP4/P1Q1P3/1P3PPP/R3KBNR w KQ - c0 "Nimzo-Indian"; acd 26; acs 30; bm f3; ce +46; Ae "Stockfish 10 64 POPCNT";
rnbq1rk1/pp2ppbp/6p1/2p5/3PP3/2P2N2/P4PPP/1RBQKB1R w K - c0 "Gruenfeld"; acd 26; acs 30; bm Be2; ce +74; Ae "Stockfish 10 64 POPCNT";

Use Stockfish to annotate epd file with depth equals 16 and no movetime

python2 chess_artist.py --infile repertoire.epd --outfile out_repertoire.epd --enginefile stockfish_10.exe --engineoptions "Hash value 128, Threads value 1" --eval search --job analyze --depth 16 --movetime 0

4. Generate chess puzzles or test positions

chess_artist.exe --infile iommast19.pgn --outfile out_iommast19.pgn --enginefile C:\chess\engines\stockfish_19100908_x64_modern.exe --engineoptions "Hash value 128" --movestart 8 --movetime 10000 --job createpuzzle --eval search

positions are saved in puzzle.epd

Example output:

r2qkb1r/pp3p2/1np1b2p/3p2p1/3PP3/2N2NP1/PPQ2PP1/2KR1B1R w kq - bm Bd3; Ubm f1d3;
r3kb1r/pp3pq1/1np1b2p/3pP3/3P2p1/2NB2P1/PPQ2PPN/2KR3R w kq - bm Kb1; Ubm c1b1;
1rbk1b1r/pp1ppppp/4qn2/2R5/2Q5/2N1P3/P1P2PPP/4KBNR w K - bm Qxe6; Ubm c4e6;
1r1k3r/1bqpbpp1/p3pn2/1p5p/8/1R1BPN1Q/P1P1NPPP/5RK1 b - - bm h4; Ubm h5h4;

The Ubm refers to best move in uci move format.

D. Credits

• Python-Chess
• Stockfish
• Multi Variant Stockfish
• TheWeekInChess