[Quantum Chinese Chess] Add class Jump(QuantumEffect). #164

madcpf · 2023-10-17T23:20:06Z

Add Jump(QuantumEffect) to handle the jump cases;
Add method unhook() to QuantumWorld, so that a given object could be replaced with a new ancilla with value 0;
Add update_board_by_sampling() in QuantumChineseChess to sample and update the classical properties of the pieces;
Update Piece.reset() method to set the is_entangled field;
Flip the labeling of the board (row <-> num_row - 1 - row) to make it more natural;
Add some test util functions.

dstrain115 · 2023-10-18T23:27:40Z

unitary/alpha/quantum_world.py

+        self.circuit = self.circuit.transform_qubits(
+            lambda q: qubit_remapping_dict.get(q, q)
+        )
+        return


Should this return the ancilla?

The method works like reset. (Maybe I should rename it to reset()?) So I don't think user should apply more effects on the ancilla. They should move forward with the reset object, which since now has a new reset value.

dstrain115 · 2023-10-19T20:56:25Z

unitary/alpha/quantum_world.py

@@ -305,6 +305,20 @@ def _interpret_result(self, result: Union[int, Iterable[int]]) -> int:
            return result_list[0]
        return result

+    def unhook(self, object: QuantumObject) -> None:
+        """Replaces all usages of the given object in the circuit with a new ancilla with value=0."""


Maybe leave off the value=0 portion or move it to a new sentence to clarify what you mean. The value=0 is the initial state of the ancilla but it won't have value=0 after all the operations act on it.

Since this is part of the alpha interface, we may want to further explain what this does in more detail.

Thanks Doug. It's updated. Is it clearer now?

dstrain115 · 2023-10-19T20:59:01Z

unitary/examples/quantum_chinese_chess/board.py

@@ -74,15 +74,11 @@ def from_fen(cls, fen: str = _INITIAL_FEN) -> "Board":
                        name, SquareState.OCCUPIED, piece_type, color
                    )
                    col += 1
-            row_index -= 1
+            row_index += 1


Next time, please separate this into two PRs, since the print change is independent of the board display change. It's fine to keep it here this time though.

Thanks Doug. I'll send separate changes with concentric topic next time.

dstrain115 · 2023-10-25T12:04:37Z

unitary/examples/quantum_chinese_chess/chess.py

@@ -381,6 +386,9 @@ def apply_move(self, str_to_parse: str) -> None:
            quantum_pieces_1,
        )

+        print(move_type, " ", move_variant)


Is this a debug statement?

Yes, it's now changed to only print above certain debug level.

dstrain115 · 2023-10-25T12:15:22Z

unitary/examples/quantum_chinese_chess/move.py

+        # i.e. pawn would move only if the target is there, i.e. CNOT(t, s), and an entanglement could be
+        # created. This could be a general game setting, i.e. we could allow players to choose if they
+        # want the source piece to move (in case of capture) if the target piece is not there.
+        source_0, target_0 = objects


It might be worth putting a check that the len(objects) is 2.

Thanks Doug. I think it's already been checked since we override num_objects()?

dstrain115 · 2023-10-25T12:16:06Z

unitary/examples/quantum_chinese_chess/move.py

+        source_0, target_0 = objects
+        world = source_0.world
+        if self.move_variant == MoveVariant.CAPTURE:
+            # We peek and force measure source_0.


nit, peek -> pop.

Thanks Doug. I think pop() includes peek and force measurement?

dstrain115 · 2023-10-25T12:19:42Z

unitary/examples/quantum_chinese_chess/move.py

+            target_0.reset()
+        elif self.move_variant == MoveVariant.CLASSICAL:
+            if target_0.type_ != Type.EMPTY:
+                # For classical moves with target_0 occupied, we replace the qubit of target_0 with


You might be correct here, but why do we need a new ancilla if it is a classical move?

Thanks Doug. By "world.unhook(target_0)" I'm trying to reset target_0 to the initial value (i.e. 0), before doing the PhasedMove below. You are right that we don't need an ancilla in this case, I'm changing it to flip(target_0) instead.

dstrain115 · 2023-10-25T12:22:49Z

unitary/examples/quantum_chinese_chess/move_test.py

+    assert_fifty_fifty(board_probabilities, locations_to_bitboard(["a2", "b1"]))
+    assert_fifty_fifty(board_probabilities, locations_to_bitboard(["a3", "b1"]))
+    Jump(MoveVariant.CAPTURE)(world["a2"], world["b1"])
+    # pop() will break the supersition and only one of the following two states are possible.


nit:superposition

Updated. Thanks.

dstrain115 · 2023-10-25T12:23:20Z

unitary/examples/quantum_chinese_chess/move_test.py

+    else:
+        assert_samples_in(board, [locations_to_bitboard(["a3", "b1"])])
+
+    # Target is in quantum state.


Consider splitting this to a new test. (and below too)

Thanks Doug. Updated into separate tests.

unitary/examples/quantum_chinese_chess/test_utils.py

initial set

0e6a382

madcpf marked this pull request as draft October 17, 2023 23:20

Pengfei Chen and others added 9 commits October 17, 2023 16:26

update

1b69d6f

update test

8c18f5f

update

46348d8

test

bcdb75a

update test

cf35e05

update

640b685

update

e8cce5a

update

576b13c

update

0ee28c1

madcpf requested a review from dstrain115 October 18, 2023 21:42

madcpf marked this pull request as ready for review October 18, 2023 21:42

update

40463a1

madcpf changed the title ~~Add class Jump(QuantumEffect).~~ [Quantum Chinese Chess] Add class Jump(QuantumEffect). Oct 23, 2023

dstrain115 reviewed Oct 25, 2023

View reviewed changes

Pengfei Chen added 2 commits October 25, 2023 12:48

up

e3a7fbc

update

8d5bee5

madcpf requested a review from dstrain115 October 25, 2023 19:50

Merge branch 'quantumlib:main' into board

a69abd7

dstrain115 approved these changes Oct 30, 2023

View reviewed changes

madcpf merged commit ab75eb3 into quantumlib:main Oct 30, 2023
6 checks passed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

[Quantum Chinese Chess] Add class Jump(QuantumEffect). #164

[Quantum Chinese Chess] Add class Jump(QuantumEffect). #164

madcpf commented Oct 17, 2023 •

edited

Loading

dstrain115 Oct 18, 2023

madcpf Oct 25, 2023

dstrain115 Oct 19, 2023

madcpf Oct 25, 2023

dstrain115 Oct 19, 2023

madcpf Oct 25, 2023

dstrain115 Oct 25, 2023

madcpf Oct 25, 2023

dstrain115 Oct 25, 2023

madcpf Oct 25, 2023

dstrain115 Oct 25, 2023

madcpf Oct 25, 2023 •

edited

Loading

dstrain115 Oct 25, 2023

madcpf Oct 25, 2023

dstrain115 Oct 25, 2023

madcpf Oct 25, 2023

dstrain115 Oct 25, 2023

madcpf Oct 25, 2023

[Quantum Chinese Chess] Add class Jump(QuantumEffect). #164

[Quantum Chinese Chess] Add class Jump(QuantumEffect). #164

Conversation

madcpf commented Oct 17, 2023 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

madcpf Oct 25, 2023 • edited Loading

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

Choose a reason for hiding this comment

madcpf commented Oct 17, 2023 •

edited

Loading

madcpf Oct 25, 2023 •

edited

Loading