[Quantum Chinese Chess] Add SplitJump and MergeJump #166
Conversation
| alpha.PhasedMove(-0.5)(source_0, target_0) | ||
| alpha.PhasedMove(-0.5)(source_1, target_0) | ||
| # Pass the classical properties of the source pieces to the target piece. | ||
| target_0.reset(source_0) |
There was a problem hiding this comment.
This semantic is kind of confusing. I am not totally sure if this is correct. Remind me what reset is supposed to do again?
There was a problem hiding this comment.
It copies the "type", "color" and "is_entangled" properties from source_0 to target_0.
There was a problem hiding this comment.
I see. And we are assuming that type, color are the same for source_0 and source_1.
Couldn't is_entangled be different for source_0 and source_1 though?
Also, do we ever fix the type/color for the source squares? Like, they may be empty now, right?
I don't think this logic is correct.
There was a problem hiding this comment.
Hi Doug, yes, earlier in chess.py we've checked to make sure source_0 and source_1 have the same type, color, and is_entangled = true (maybe later we could support classical to merge with entangled piece).
The type/color/is_entangled for source squares are fixed through the method "update_board_by_sampling()" in chess.py, which will be called after each move. There are cases that they won't be reset to empty (see test cases for those scenarios).
The source piece for SplitJump will also be fixed through update_board_by_sampling().
|
|
||
| def test_merge_jump_imperfect_merge_scenario_3(): | ||
| # Imperfect merge scenario 3 | ||
| # This is a simplied version of the scenario above, where we unhook a3 and b3. |
There was a problem hiding this comment.
Consider changing the line comments here and above to docstrings with triple quotes. I think they will show up in IDEs better.
There was a problem hiding this comment.
Thanks Doug! It's updated.
Add SplitJump and MergeJump as new moves.