Merge pull request #1172 from zizhang-qiu:fix_bridge_observation_tensor

PiperOrigin-RevId: 606230032
Change-Id: Ia765dbfd3b84d80fec43d5896b3a125b8e8fb42e

Author: lanctot

open_spiel/games/bridge/bridge.cc

@@ -314,7 +314,8 @@ void BridgeState::WriteObservationTensor(Player player,
   auto ptr = values.begin();
   if (num_cards_played_ > 0) {
     // Observation for play phase
-    if (phase_ == Phase::kPlay) ptr[2] = 1;
+    const bool defending = (partnership != Partnership(contract_.declarer));
+    if (phase_ == Phase::kPlay) ptr[2 + defending] = 1;
     ptr += kNumObservationTypes;
 
     // Contract

open_spiel/integration_tests/playthroughs/bridge(use_double_dummy_result=false).txt

@@ -738,9 +738,9 @@ ObservationString(1) = "Vul: None\nS 9\nH QT863\nD A75\nC 982\n\nWest  North Eas
 ObservationString(2) = "Vul: None\nS KQJT2\nH none\nD 82\nC Q7643\n\nWest  North East  South\n      1S    3H    4S    \n5H    Dbl   Pass  5S    \nPass  Pass  Pass  \n\nN  E  S  W  N  E  S\n   HK H2 \n\nDeclarer tricks: 0"
 ObservationString(3) = "Vul: None\nS 76\nH J75\nD KQT9643\nC K\n\nWest  North East  South\n      1S    3H    4S    \n5H    Dbl   Pass  5S    \nPass  Pass  Pass  \n\nN  E  S  W  N  E  S\n   HK H2 \n\nDeclarer tricks: 0"
 ObservationTensor(0): binvec(571, 0x104148804e4004a3002d620221004644000000000000000000000000080800000000000000000000000000000000000000000000000000000000000000000000000000002001000)
-ObservationTensor(1): binvec(571, 0x104141a081092a480811620221004644000000000000808000000000000000000000000000000000000000000000000000000000000000000000000000000000000000002001000)
+ObservationTensor(1): binvec(571, 0x84141a081092a480811620221004644000000000000808000000000000000000000000000000000000000000000000000000000000000000000000000000000000000002001000)
 ObservationTensor(2): binvec(571, 0x10414296202210046441620221004644080000000000000000000000000000000000000000000000000800000000000000000000000000000000000000000000000000002001000)
-ObservationTensor(3): binvec(571, 0x104144811094c1109301620221004644000000000000000000000000000000000000008080000000000000000000000000000000000000000000000000000000000000002001000)
+ObservationTensor(3): binvec(571, 0x84144811094c1109301620221004644000000000000000000000000000000000000008080000000000000000000000000000000000000000000000000000000000000002001000)
 Rewards() = [0, 0, 0, 0]
 Returns() = [0, 0, 0, 0]
 LegalActions() = [14, 22, 38]
@@ -831,9 +831,9 @@ ObservationString(1) = "Vul: None\nS none\nH QT86\nD A75\nC 982\n\nWest  North E
 ObservationString(2) = "Vul: None\nS QJ2\nH none\nD 82\nC Q764\n\nWest  North East  South\n      1S    3H    4S    \n5H    Dbl   Pass  5S    \nPass  Pass  Pass  \n\nN  E  S  W  N  E  S\n   HK H2 H5 HA \nS3 S9 ST S7 \n      SK S6 S5 H3 \n      C3 CK \n\nDeclarer tricks: 3"
 ObservationString(3) = "Vul: None\nS none\nH J7\nD KQT9643\nC none\n\nWest  North East  South\n      1S    3H    4S    \n5H    Dbl   Pass  5S    \nPass  Pass  Pass  \n\nN  E  S  W  N  E  S\n   HK H2 H5 HA \nS3 S9 ST S7 \n      SK S6 S5 H3 \n      C3 CK \n\nDeclarer tricks: 3"
 ObservationTensor(0): binvec(571, 0x104148800e0004a30025420221000640000000000000000000000000000200000000000000000000002000004000000000080000000000000000000000400000400000000401000)
-ObservationTensor(1): binvec(571, 0x104141a001092a080811420221000640000000000000002000000000000000000000020000000000000000800000000000000000000004000004000000000004000000000401000)
+ObservationTensor(1): binvec(571, 0x84141a001092a080811420221000640000000000000002000000000000000000000020000000000000000800000000000000000000004000004000000000004000000000401000)
 ObservationTensor(2): binvec(571, 0x10414294202210006401420221000640020000000000000000000000200000000000000000000000000000000000000040000040000000000040000000000800000000000401000)
-ObservationTensor(3): binvec(571, 0x10414481101081109101420221000640000000000002000000000000000000000000000020000000000000000400000000000400000000008000000000000000000000040401000)
+ObservationTensor(3): binvec(571, 0x8414481101081109101420221000640000000000002000000000000000000000000000020000000000000000400000000000400000000008000000000000000000000040401000)
 Rewards() = [0, 0, 0, 0]
 Returns() = [0, 0, 0, 0]
 LegalActions() = [12, 32, 36, 48]
@@ -940,9 +940,9 @@ ObservationString(1) = "Vul: None\nS none\nH QT8\nD A75\nC none\n\nWest  North E
 ObservationString(2) = "Vul: None\nS QJ2\nH none\nD 82\nC 7\n\nWest  North East  South\n      1S    3H    4S    \n5H    Dbl   Pass  5S    \nPass  Pass  Pass  \n\nN  E  S  W  N  E  S\n   HK H2 H5 HA \nS3 S9 ST S7 \n      SK S6 S5 H3 \n      C3 CK CA C2 \nC5 C8 CQ D3 \n      C4 D4 CJ C9 \nCT H6 C6 D6 \nH4 \n\nDeclarer tricks: 7"
 ObservationString(3) = "Vul: None\nS none\nH J7\nD KQT9\nC none\n\nWest  North East  South\n      1S    3H    4S    \n5H    Dbl   Pass  5S    \nPass  Pass  Pass  \n\nN  E  S  W  N  E  S\n   HK H2 H5 HA \nS3 S9 ST S7 \n      SK S6 S5 H3 \n      C3 CK CA C2 \nC5 C8 CQ D3 \n      C4 D4 CJ C9 \nCT H6 C6 D6 \nH4 \n\nDeclarer tricks: 7"
 ObservationTensor(0): binvec(571, 0x10414880040004810005400021000440000800000000000000000000000000000000000000000000000000000000200000000080000000000020000000000001000000000041000)
-ObservationTensor(1): binvec(571, 0x10414180010108080811400021000440000000000000000000000000000000000000000000800000000000000800000000000200000000000010000000000000000200000041000)
+ObservationTensor(1): binvec(571, 0x8414180010108080811400021000440000000000000000000000000000000000000000000800000000000000800000000000200000000000010000000000000000200000041000)
 ObservationTensor(2): binvec(571, 0x10414294000210004401400021000440000000000000000000000000000008000000000000000000000000002000000000000100000000000000002000000000800000000041000)
-ObservationTensor(3): binvec(571, 0x10414480000081109101400021000440000000000000000080000000000000000000000000000000000000001000000000000000020000000008000000000002000000000041000)
+ObservationTensor(3): binvec(571, 0x8414480000081109101400021000440000000000000000080000000000000000000000000000000000000001000000000000000020000000008000000000002000000000041000)
 Rewards() = [0, 0, 0, 0]
 Returns() = [0, 0, 0, 0]
 LegalActions() = [26, 34, 42]
@@ -1006,9 +1006,9 @@ ObservationString(1) = "Vul: None\nS none\nH QT\nD A75\nC none\n\nWest  North Ea
 ObservationString(2) = "Vul: None\nS QJ\nH none\nD 82\nC none\n\nWest  North East  South\n      1S    3H    4S    \n5H    Dbl   Pass  5S    \nPass  Pass  Pass  \n\nN  E  S  W  N  E  S\n   HK H2 H5 HA \nS3 S9 ST S7 \n      SK S6 S5 H3 \n      C3 CK CA C2 \nC5 C8 CQ D3 \n      C4 D4 CJ C9 \nCT H6 C6 D6 \nH4 H8 S2 H7 \n      C7 \n\nDeclarer tricks: 8"
 ObservationString(3) = "Vul: None\nS none\nH J\nD KQT9\nC none\n\nWest  North East  South\n      1S    3H    4S    \n5H    Dbl   Pass  5S    \nPass  Pass  Pass  \n\nN  E  S  W  N  E  S\n   HK H2 H5 HA \nS3 S9 ST S7 \n      SK S6 S5 H3 \n      C3 CK CA C2 \nC5 C8 CQ D3 \n      C4 D4 CJ C9 \nCT H6 C6 D6 \nH4 H8 S2 H7 \n      C7 \n\nDeclarer tricks: 8"
 ObservationTensor(0): binvec(571, 0x10414880040004810005000001000440000000000000000000000000000000020000000000000000000000080000000000000000800000040000000000000000080000000021000)
-ObservationTensor(1): binvec(571, 0x10414180010100080811000001000440000000000000000000200000000000000000000000000000000000000008000000400000000000000000800000000080000000000021000)
+ObservationTensor(1): binvec(571, 0x8414180010100080811000001000440000000000000000000200000000000000000000000000000000000000008000000400000000000000000800000000080000000000021000)
 ObservationTensor(2): binvec(571, 0x10414290000010004401000001000440000002000000000000000000000000000000000000000000000004000000000000000008000000000800000000000000008000000021000)
-ObservationTensor(3): binvec(571, 0x10414480000001109101000001000440000000000000000000000000000000000000000000002000000000000080000000008000000000000000080000004000000000000021000)
+ObservationTensor(3): binvec(571, 0x8414480000001109101000001000440000000000000000000000000000000000000000000002000000000000080000000008000000000000000080000004000000000000021000)
 Rewards() = [0, 0, 0, 0]
 Returns() = [0, 0, 0, 0]
 LegalActions() = [29, 33, 38, 41, 45]
@@ -1060,9 +1060,9 @@ ObservationString(1) = "Vul: None\nS none\nH QT\nD A75\nC none\n\nWest  North Ea
 ObservationString(2) = "Vul: None\nS QJ\nH none\nD 82\nC none\n\nWest  North East  South\n      1S    3H    4S    \n5H    Dbl   Pass  5S    \nPass  Pass  Pass  \n\nN  E  S  W  N  E  S\n   HK H2 H5 HA \nS3 S9 ST S7 \n      SK S6 S5 H3 \n      C3 CK CA C2 \nC5 C8 CQ D3 \n      C4 D4 CJ C9 \nCT H6 C6 D6 \nH4 H8 S2 H7 \n      C7 D9 \n\nDeclarer tricks: 8"
 ObservationString(3) = "Vul: None\nS none\nH J\nD KQT\nC none\n\nWest  North East  South\n      1S    3H    4S    \n5H    Dbl   Pass  5S    \nPass  Pass  Pass  \n\nN  E  S  W  N  E  S\n   HK H2 H5 HA \nS3 S9 ST S7 \n      SK S6 S5 H3 \n      C3 CK CA C2 \nC5 C8 CQ D3 \n      C4 D4 CJ C9 \nCT H6 C6 D6 \nH4 H8 S2 H7 \n      C7 D9 \n\nDeclarer tricks: 8"
 ObservationTensor(0): binvec(571, 0x10414880040004810005000001000440000000000000000000000000000000020000000000000010000000080000000000000000800000040000000000000000080000000021000)
-ObservationTensor(1): binvec(571, 0x10414180010100080811000001000440000000000000000000200000000000000100000000000000000000000008000000400000000000000000800000000080000000000021000)
+ObservationTensor(1): binvec(571, 0x8414180010100080811000001000440000000000000000000200000000000000100000000000000000000000008000000400000000000000000800000000080000000000021000)
 ObservationTensor(2): binvec(571, 0x10414290000010004401000001000440000002000000000000001000000000000000000000000000000004000000000000000008000000000800000000000000008000000021000)
-ObservationTensor(3): binvec(571, 0x10414480000000109101000001000440000000010000000000000000000000000000000000002000000000000080000000008000000000000000080000004000000000000021000)
+ObservationTensor(3): binvec(571, 0x8414480000000109101000001000440000000010000000000000000000000000000000000002000000000000080000000008000000000000000080000004000000000000021000)
 Rewards() = [0, 0, 0, 0]
 Returns() = [0, 0, 0, 0]
 LegalActions() = [11, 27, 30, 37, 51]