feat: update the policy if the tool fails to respond

packages/valory/skills/decision_maker_abci/behaviours/decision_receive.py

@@ -27,9 +27,11 @@
 from typing import Any, Dict, Generator, List, Optional, Tuple, Union
 
 from packages.valory.skills.decision_maker_abci.behaviours.base import (
-    DecisionMakerBaseBehaviour,
     remove_fraction_wei,
 )
+from packages.valory.skills.decision_maker_abci.behaviours.storage_manager import (
+    StorageManagerBehaviour,
+)
 from packages.valory.skills.decision_maker_abci.io_.loader import ComponentPackageLoader
 from packages.valory.skills.decision_maker_abci.models import (
     BenchmarkingMockData,
@@ -59,7 +61,7 @@
 TOKEN_PRECISION = 10**18
 
 
-class DecisionReceiveBehaviour(DecisionMakerBaseBehaviour):
+class DecisionReceiveBehaviour(StorageManagerBehaviour):
     """A behaviour in which the agents receive the mech response."""
 
     matching_round = DecisionReceiveRound
@@ -93,6 +95,17 @@ def mech_response(self) -> MechInteractionResponse:
             return MechInteractionResponse(error=error)
         return self._mech_response
 
+    @property
+    def is_invalid_response(self) -> bool:
+        """Check if the response is invalid."""
+        if self.mech_response.result is None:
+            self.context.logger.warning(
+                "Trying to check whether the mech's response is invalid but no response has been detected! "
+                "Assuming invalid response."
+            )
+            return True
+        return self.mech_response.result == self.params.mech_invalid_response
+
     def _next_dataset_row(self) -> Optional[Dict[str, str]]:
         """Read the next row from the input dataset which is used during the benchmarking mode.
 
@@ -527,12 +540,17 @@ def async_act(self) -> Generator:
         """Do the action."""
 
         with self.context.benchmark_tool.measure(self.behaviour_id).local():
+            success = yield from self._setup_policy_and_tools()
+            if not success:
+                return None
+
             prediction_response = self._get_decision()
             is_profitable = None
             bet_amount = None
             next_mock_data_row = None
             bets_hash = None
             decision_received_timestamp = None
+            policy = None
             if prediction_response is not None and prediction_response.vote is not None:
                 is_profitable, bet_amount = yield from self._is_profitable(
                     prediction_response
@@ -552,6 +570,14 @@ def async_act(self) -> Generator:
                     bet_amount,
                 )
 
+            if prediction_response is not None:
+                self.policy.tool_responded(
+                    self.synchronized_data.mech_tool,
+                    self.synced_timestamp,
+                    self.is_invalid_response,
+                )
+                policy = self.policy.serialize()
+
             # always remove the processed trade from the benchmarking input file
             # now there is one reader pointer per market
             if self.benchmarking_mode.enabled:
@@ -572,7 +598,9 @@ def async_act(self) -> Generator:
                 prediction_response.confidence if prediction_response else None,
                 bet_amount,
                 next_mock_data_row,
+                policy,
                 decision_received_timestamp,
             )
 
+        self._store_all()
         yield from self.finish_behaviour(payload)

packages/valory/skills/decision_maker_abci/payloads.py

@@ -35,6 +35,7 @@ class DecisionReceivePayload(UpdateBetsPayload):
     confidence: Optional[float]
     bet_amount: Optional[int]
     next_mock_data_row: Optional[int]
+    policy: Optional[str]
     decision_received_timestamp: Optional[int]
 
 
@@ -119,3 +120,10 @@ class BetPlacementPayload(MultisigTxPayload):
     """Represents a transaction payload for placing a bet."""
 
     wallet_balance: Optional[int] = None
+
+
+@dataclass(frozen=True)
+class HandleFailedTxPayload(VotingPayload):
+    """Represents a transaction payload for placing a bet."""
+
+    tx_submitter: str

packages/valory/skills/decision_maker_abci/states/decision_receive.py

@@ -49,6 +49,7 @@ class DecisionReceiveRound(CollectSameUntilThresholdRound):
         get_name(SynchronizedData.confidence),
         get_name(SynchronizedData.bet_amount),
         get_name(SynchronizedData.next_mock_data_row),
+        get_name(SynchronizedData.policy),
     )
     collection_key = get_name(SynchronizedData.participant_to_decision)
 

packages/valory/skills/decision_maker_abci/tests/states/test_decision_receive.py

@@ -63,6 +63,7 @@ def get_payloads(
     next_mock_data_row: Optional[int],
     is_profitable: Optional[bool],
     bets_hash: str,
+    policy: str,
 ) -> Mapping[str, UpdateBetsPayload]:
     """Get payloads."""
     return {
@@ -73,7 +74,8 @@ def get_payloads(
             bet_amount=bet_amount,
             next_mock_data_row=next_mock_data_row,
             is_profitable=is_profitable,
-            bets_hash=bets_hash,  # Added bets_hash parameter
+            bets_hash=bets_hash,
+            policy=policy,
             decision_received_timestamp=int(datetime.datetime.utcnow().timestamp()),
         )
         for participant in get_participants()
@@ -110,6 +112,7 @@ class TestDecisionReceiveRound(BaseCollectSameUntilThresholdRoundTest):
                     bet_amount=100,
                     next_mock_data_row=1,
                     is_profitable=True,
+                    policy="",
                     bets_hash=DUMMY_BETS_HASH,  # Added bets_hash
                 ),
                 final_data={
@@ -131,6 +134,7 @@ class TestDecisionReceiveRound(BaseCollectSameUntilThresholdRoundTest):
                     bet_amount=50,
                     next_mock_data_row=2,
                     is_profitable=False,
+                    policy="",
                     bets_hash=DUMMY_BETS_HASH,  # Added bets_hash
                 ),
                 final_data={
@@ -152,6 +156,7 @@ class TestDecisionReceiveRound(BaseCollectSameUntilThresholdRoundTest):
                     bet_amount=None,
                     next_mock_data_row=None,
                     is_profitable=True,
+                    policy="",
                     bets_hash=DUMMY_BETS_HASH,  # Added bets_hash
                 ),
                 final_data={},
@@ -168,6 +173,7 @@ class TestDecisionReceiveRound(BaseCollectSameUntilThresholdRoundTest):
                     bet_amount=None,
                     next_mock_data_row=None,
                     is_profitable=True,
+                    policy="",
                     bets_hash=DUMMY_BETS_HASH,  # Added bets_hash
                 ),
                 final_data={},
@@ -184,6 +190,7 @@ class TestDecisionReceiveRound(BaseCollectSameUntilThresholdRoundTest):
                     bet_amount=None,
                     next_mock_data_row=None,
                     is_profitable=True,
+                    policy="",
                     bets_hash=DUMMY_BETS_HASH,  # Added bets_hash
                 ),
                 final_data={},

packages/valory/skills/decision_maker_abci/tests/test_payloads.py

@@ -50,6 +50,7 @@
                 "confidence": 0.90,
                 "bet_amount": 1,
                 "next_mock_data_row": 1,
+                "policy": "dummy policy",
                 "decision_received_timestamp": int(datetime.utcnow().timestamp()),
             },
         ),