As I wrote the Japanese crossword puzzle in 4 hours

SAT- () (1, 2, 3),(-1, 2 ..), (True/False).

, , .

( , ); , , 8x8 = 64 , .

(1 2 3… 8) — 1 . — [1,2,3...8]
— (9… 16) ..

— 1 !
[ 1 2] ( — ). [-1,-2],[-1,-3] .

, , . , .

( ) — (1 2… 8) ( 1 2) (...). , — .

for cnter in range(4):
  sol = pycosat.solve(clauses)
  if isinstance(sol, list):
    print(sol)
    clauses.append([-x for x in sol]) #   -      
  else: #    - 
    return
#     ,    ,   .

, «-» ?

/ /, — . , .

[1] 3 , :
[X__],[_X_],[__X]
, ,
[1,-2,-3],[-1,2,-3],[-1,-2,3]

, …
:
[1 -2 -3] [-1, 2...] [...]

, . , . — , !

We submit the test input of DNF: lstall = [[1,2,3]]. That is, we expect that only [1,2,3] (corresponds to the full line [XXX]) will be a valid solution. The

library gives KNF: cnf = [[-1, -2, -3, 4], [1 , -4], [2, -4], [3, -4]]
It is not clear, but very interesting. Well, let's say I believe you ...

We ask pycosat to find all the solutions for cnf:
[1, 2, 3, 4]
[1, 2, -3, -4]
[1, -2, -3, -4]
[1 , -2, 3, -4]
[-1, -2, -3, -4]
[-1, -2, 3, -4]
[-1, 2, -3, -4]
[-1, 2, 3, -4]

Instead of the expected one, there are eight of them! What to do?

, , 4 True. ?

: cnf2 = [[-1, -2, -3, 4], [1, -4], [2, -4], [3, -4], [4]]
pycosat cnf2:
1: [1, 2, 3, 4]

, , ! , , .
— !