Added boolean flag to control IPFS upload

.env.example

@@ -3,4 +3,5 @@ BET_FROM_PRIVATE_KEY=
 OPENAI_API_KEY=
 GRAPH_API_KEY=
 PINATA_API_KEY=
-PINATA_API_SECRET=
+PINATA_API_SECRET=
+ENABLE_IPFS_UPLOAD=

prediction_market_agent_tooling/config.py

@@ -45,6 +45,7 @@ class APIKeys(BaseSettings):
     LANGFUSE_HOST: t.Optional[str] = None
     LANGFUSE_DEPLOYMENT_VERSION: t.Optional[str] = None
 
+    ENABLE_IPFS_UPLOAD: bool = False
     PINATA_API_KEY: t.Optional[SecretStr] = None
     PINATA_API_SECRET: t.Optional[SecretStr] = None
 
@@ -151,6 +152,12 @@ def default_enable_langfuse(self) -> bool:
             and self.LANGFUSE_HOST is not None
         )
 
+    @property
+    def enable_ipfs_upload(self) -> bool:
+        return check_not_none(
+            self.ENABLE_IPFS_UPLOAD, "ENABLE_IPFS_UPLOAD missing in the environment."
+        )
+
     @property
     def pinata_api_key(self) -> SecretStr:
         return check_not_none(

prediction_market_agent_tooling/deploy/agent.py

@@ -11,6 +11,7 @@
 from pydantic import BaseModel, BeforeValidator, computed_field
 from typing_extensions import Annotated
 from web3 import Web3
+from web3.constants import HASH_ZERO
 
 from prediction_market_agent_tooling.config import APIKeys
 from prediction_market_agent_tooling.deploy.betting_strategy import (
@@ -302,7 +303,6 @@ def __init__(
     ) -> None:
         super().__init__(enable_langfuse=enable_langfuse)
         self.place_bet = place_bet
-        self.ipfs_handler = IPFSHandler(APIKeys())
 
     def get_betting_strategy(self, market: AgentMarket) -> BettingStrategy:
         user_id = market.get_user_id(api_keys=APIKeys())
@@ -522,16 +522,21 @@ def store_prediction(
             if processed_market.answer.reasoning
             else ""
         )
-        ipfs_hash = self.ipfs_handler.store_agent_result(
-            IPFSAgentResult(reasoning=reasoning)
-        )
+
+        ipfs_hash_decoded = HexBytes(HASH_ZERO)
+        if keys.enable_ipfs_upload:
+            logger.info("Storing prediction on IPFS.")
+            ipfs_hash = IPFSHandler(keys).store_agent_result(
+                IPFSAgentResult(reasoning=reasoning)
+            )
+            ipfs_hash_decoded = ipfscidv0_to_byte32(ipfs_hash)
 
         tx_hashes = [
             HexBytes(HexStr(i.id)) for i in processed_market.trades if i.id is not None
         ]
         prediction = ContractPrediction(
             publisher=keys.public_key,
-            ipfs_hash=ipfscidv0_to_byte32(ipfs_hash),
+            ipfs_hash=ipfs_hash_decoded,
             tx_hashes=tx_hashes,
             estimated_probability_bps=int(processed_market.answer.p_yes * 10000),
         )

tests_integration_with_local_chain/markets/omen/test_omen.py

@@ -6,6 +6,7 @@
 import numpy as np
 import pytest
 from web3 import Web3
+from web3.constants import HASH_ZERO
 
 from prediction_market_agent_tooling.config import APIKeys
 from prediction_market_agent_tooling.gtypes import (
@@ -455,25 +456,32 @@ def get_position_balance_by_position_id(
     )
 
 
-def test_add_predictions(local_web3: Web3, test_keys: APIKeys) -> None:
+@pytest.mark.parametrize(
+    "ipfs_hash",
+    ["0x3750ffa211dab39b4d0711eb27b02b56a17fa9d257ee549baa3110725fd1d41b", HASH_ZERO],
+)
+def test_add_predictions(local_web3: Web3, test_keys: APIKeys, ipfs_hash: str) -> None:
     agent_result_mapping = OmenAgentResultMappingContract()
     market_address = test_keys.public_key
     dummy_transaction_hash = (
         "0x3750ffa211dab39b4d0711eb27b02b56a17fa9d257ee549baa3110725fd1d41b"
     )
+    stored_predictions = agent_result_mapping.get_predictions(
+        market_address, web3=local_web3
+    )
     p = ContractPrediction(
         tx_hashes=[HexBytes(dummy_transaction_hash)],
         estimated_probability_bps=5454,
-        ipfs_hash=HexBytes(dummy_transaction_hash),
+        ipfs_hash=HexBytes(ipfs_hash),
         publisher=test_keys.public_key,
     )
 
     agent_result_mapping.add_prediction(test_keys, market_address, p, web3=local_web3)
-    stored_predictions = agent_result_mapping.get_predictions(
+    updated_stored_predictions = agent_result_mapping.get_predictions(
         market_address, web3=local_web3
     )
-    assert len(stored_predictions) == 1
-    assert stored_predictions[0] == p
+    assert len(updated_stored_predictions) == len(stored_predictions) + 1
+    assert updated_stored_predictions[-1] == p
 
 
 def test_place_bet_with_prev_existing_positions(