Puzzle Kit

This repository provides 60+ useful, efficient, and problem‑specific solvers for a variety of logic puzzles. The underlying solving engines are open‑source tools such as ortools.

The repository also includes a structured dataset (32k+ instances) covering 100+ specific and popular puzzle types (like Nonogram, Slitherlink, Akari, Fillomino, Hitori, Kakuro, Kakuro) from Raetsel's Janko and puzz.link, as detailed in the table below. More data, along with related analytics, will be added over time.

Most of solvers implemented in this repo are both effective and efficient. They have been tested in 20k+ instances, most of which can be easily solved 0.5 s, even grids with a scale of 20x20. The detailed table of puzzles, datasets and solver performance are shown below.

Table of puzzles, datasets and solvers.

S&V means "Solved and Verified",

C? indicates whether the crawler of the puzzle is provided.

Problem and Solution means the number of instances and its answers in dataset respectively.

Max Size shows the maximum scale of existing puzzles.

Performance of solvers (solving time, number of constrs and variables would be added in the future). Some solvers are not refactored, thus they are marked as unsolved and unverified.

No.	Puzzle Name	Problems	Solutions	Max Size	Solver?	Avg Time	Max Time	Correct
1	ABCEndView	607	607	8x8	✅	0.015s	0.061s	592
2	Akari	970	970	100x100	✅	0.014s	0.478s	970
3	BalanceLoop	70	70	17x17	✅	0.062s	0.243s	70
4	Battleship	0	0	-	❌	-	-	-
5	Binairo	380	380	14x14	✅	0.007s	0.015s	380
6	Bosanowa	38	38	11x16	✅	0.016s	0.214s	38
7	Bricks	210	210	8x8	❌	-	-	-
8	Buraitoraito	101	100	15x15	✅	0.007s	0.123s	100
9	Burokku	270	270	10x10	❌	-	-	-
10	ButterflySudoku	77	77	12x12	✅	0.008s	0.011s	77
11	Clueless1Sudoku	29	29	27x27	✅	0.029s	0.041s	29
12	Clueless2Sudoku	40	40	27x27	✅	0.033s	0.082s	40
13	ConsecutiveSudoku	211	211	9x9	❌	-	-	-
14	Corral	419	419	25x25	❌	-	-	-
15	CountryRoad	270	270	15x15	✅	0.026s	0.084s	270
16	Creek	440	440	40x50	❌	-	-	-
17	CurvingRoad	190	190	14x14	❌	-	-	-
18	Detour	80	80	13x12	✅	0.024s	0.357s	80
19	DiffNeighbors	140	140	15x15	❌	-	-	-
20	Dominos	580	579	10x11	✅	0.004s	0.009s	579
21	DotchiLoop	60	60	17x17	❌	-	-	-
22	DoubleBack	100	100	26x26	✅	0.024s	0.262s	100
23	EntryExit	170	170	16x16	✅	0.036s	0.068s	170
24	Eulero	290	290	5x5	✅	0.004s	0.011s	290
25	EvenOddSudoku	129	129	9x9	✅	0.004s	0.006s	129
26	Fillomino	840	840	50x64	❌	-	-	-
27	Fobidoshi	250	250	12x12	✅	0.052s	0.128s	250
28	Foseruzu	310	310	30x45	❌	-	-	-
29	Fuzuli	160	160	8x8	✅	0.010s	0.017s	160
30	Gappy	429	427	18x18	✅	0.018s	0.059s	427
31	Gattai8Sudoku	120	120	21x33	✅	0.020s	0.028s	120
32	GokigenNaname	780	780	24x36	❌	-	-	-
33	GrandTour	350	350	15x15	✅	0.018s	0.059s	350
34	Hakoiri	140	140	12x12	❌	-	-	-
35	Hakyuu	480	480	30x45	✅	0.042s	0.935s	480
36	Heyawake	787	787	31x45	✅	0.470s	22.841s	786
37	Hidoku	510	510	10x10	❌	-	-	-
38	Hitori	941	941	25x25	✅	0.210s	2.084s	940
39	JigsawSudoku	680	680	9x9	✅	0.003s	0.008s	665
40	Juosan	80	80	30x45	❌	-	-	-
41	Kakkuru	400	400	9x9	❌	-	-	-
42	Kakurasu	280	280	11x11	✅	0.003s	0.016s	280
43	Kakuro	999	999	31x46	✅	0.011s	0.189s	999
44	KillerSudoku	810	810	9x9	✅	0.006s	0.091s	586
45	Kuromasu	560	560	31x45	❌	-	-	-
46	Kuroshuto	210	210	14x14	✅	0.146s	0.892s	210
47	LITS	419	419	40x57	❌	-	-	-
48	Linesweeper	310	310	16x16	✅	0.017s	0.046s	310
49	Magnetic	439	439	12x12	✅	0.010s	0.028s	439
50	Makaro	190	190	15x15	❌	-	-	-
51	MarginSudoku	149	149	9x9	❌	-	-	-
52	Masyu	828	828	40x58	✅	0.065s	0.777s	828
53	Maze-a-pix	0	0	-	❌	-	-	-
54	Minesweeper	360	360	14x24	✅	0.005s	0.023s	360
55	MoonSun	200	200	30x45	❌	-	-	-
56	Mosaic	165	104	118x100	✅	0.015s	0.123s	104
57	Munraito	360	360	12x12	✅	0.010s	0.026s	360
58	Nanbaboru	270	270	9x9	❌	-	-	-
59	Nondango	110	110	14x14	✅	0.004s	0.009s	110
60	Nonogram	2338	2337	30x40	✅	0.305s	1.216s	2337
61	Norinori	289	289	36x54	✅	0.008s	0.083s	288
62	NumberCross	170	170	8x8	❌	-	-	-
63	NumberLink	580	580	35x48	❌	-	-	-
64	NumberSnake	70	70	10x10	❌	-	-	-
65	OneToX	58	58	10x10	✅	0.011s	0.138s	58
66	Patchwork	211	211	12x12	✅	0.020s	0.032s	211
67	Pfeilzahlen	360	360	10x10	✅	0.012s	0.027s	358
68	Pills	164	163	10x10	✅	0.007s	0.010s	163
69	Polyiamond	0	0	-	❌	-	-	-
70	Polyminoes	0	0	-	❌	-	-	-
71	RegionalYajilin	70	70	10x18	❌	-	-	-
72	Renban	150	150	9x9	✅	0.005s	0.066s	150
73	SamuraiSudoku	272	272	21x21	✅	0.011s	0.021s	272
74	Shikaku	500	500	31x45	✅	0.010s	0.056s	497
75	Shimaguni	266	266	30x45	❌	-	-	-
76	Shingoki	100	100	21x21	❌	-	-	-
77	Shirokuro	110	110	17x17	❌	-	-	-
78	ShogunSudoku	90	90	21x45	✅	0.030s	0.049s	90
79	Shugaku	130	130	30x45	❌	-	-	-
80	SimpleLoop	70	70	17x18	✅	0.020s	0.053s	70
81	Skyscraper	470	470	8x8	❌	-	-	-
82	SkyscraperSudoku	50	50	9x9	❌	-	-	-
83	Slitherlink	1176	1153	60x60	✅	0.067s	1.797s	1149
84	Snake	230	230	12x12	❌	-	-	-
85	SoheiSudoku	120	120	21x21	✅	0.010s	0.021s	120
86	SquareO	120	80	15x15	✅	0.004s	0.022s	80
87	Starbattle	307	307	15x15	✅	0.009s	0.025s	307
88	Sternenhimmel	29	29	17x17	❌	-	-	-
89	Str8t	560	560	9x9	❌	-	-	-
90	Sudoku	125	125	16x16	✅	0.010s	0.020s	125
91	Suguru	200	200	10x10	✅	0.008s	0.015s	200
92	SumoSudoku	110	110	33x33	✅	0.032s	0.048s	110
93	Tatamibari	150	150	14x14	❌	-	-	-
94	TennerGrid	375	374	6x10	✅	0.007s	0.012s	374
95	Tent	706	706	30x30	✅	0.006s	0.026s	706
96	TerraX	80	80	17x17	✅	0.009s	0.018s	80
97	Thermometer	250	250	10x10	✅	0.004s	0.007s	250
98	TilePaint	377	377	16x16	✅	0.004s	0.069s	377
99	Trinairo	60	60	12x12	❌	-	-	-
100	WindmillSudoku	150	150	21x21	✅	0.012s	0.019s	150
101	Yajilin	610	610	39x57	✅	0.053s	0.542s	610
102	YinYang	170	170	14x14	❌	-	-	-
103	Yonmasu	120	120	10x10	❌	-	-	-
104	fivecells	0	0	-	❌	-	-	-
	Total	32330	32199	-	-	-	-	-

Gallery of some puzzles (not complete!)

Unlike other solvers that rely on logical/deductive methods, the solvers here are primarily based on Constraint Programming. While I greatly admire those who can spot purely logical solutions, this project is not intended to replace human reasoning with automated solving: it’s just for fun.

Some details are greatly inspired by similar yet more sophisticated repositories like puzzle_solver for 90+puzzles by Ar-Kareem and Puzzles-Solver in action by newtomsoft.

Usage

Dependencies: Python >= 3.10, e.g.,

conda create -n py310 python=3.10.14

install the solvers via pip:

pip install puzzlekit

Then a quick tour:

import puzzlekit

data = {
        "num_rows": 20, 
        "num_cols": 20, 
        "cols": list(map(lambda x: x.split(" "), "2 8\n5 2\n2 4 2\n6 1\n2 2\n2 5\n6 1 6\n3 3 6\n2 1 1 1 3 1\n2 1 1 1\n2 1\n2 1 1 1\n1 1 1 4\n2 1 1 2 2\n2 3 1\n2 1 1 1\n2 1 1 5 1\n3 8 1\n6 6 2\n7".split("\n"))),
        "rows": list(map(lambda x: x.split(" "), "2 2\n1 1 4 4\n1 1 2 3 2\n1 1\n4 2 1 1 2\n4 1 2 2 1\n4 1 1 1\n4 1 2 2 1\n2 1 1 3 1\n1 2 2 2\n1 2 2 1\n2 3 3\n2 4\n3 1 4\n4 1 4\n4 2 4\n4 2 3\n4 1 1\n3 2 2\n13".split("\n"))),
        "grid": list()
        }

solver = puzzlekit.solver("nonogram", data)
result = solver.solve()
print(result)
# result.show(auto_close_sec = 10) 
# run this If you wanna see the final result in matplotlib ... 

Table of contents

1. Solvers for Logic Puzzles by CS-SAT. INTERESTING and brain-burning logic puzzles (at least it's hard for impatient guys like me).

2. Dataset of 100+ puzzles, One of the key features that distinguishes this repository from related works.

• Motivation: Many puzzles available online are stored in PDF or image formats, which are not readily usable for automated solving. This repository provides easy-to-use web crawlers that extract puzzle data and convert it into a structured, machine-readable format.
• Usage: The datasets can serve as benchmarks for evaluating and testing the performance of computer-aided solvers.

---

Reference

• ortools Official.
• Hakank's ORtools tutorials.
• PySCIPOpt's tutorials.
• Puzzle data source: Raetsel's Janko, Puzzle.
• Related repos like puzzle_solver and Puzzles-Solver.
• puzz.link and pzprjs.
• Nonogram solver.