Closed slundqui closed 4 months ago
Attention: Patch coverage is 86.18421%
with 21 lines
in your changes missing coverage. Please review.
Project coverage is 81.05%. Comparing base (
378a3fd
) to head (e7f7e33
). Report is 84 commits behind head on main.
:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.
This PR adds various changes to support rl bots training. Some of these changes helps take a step in allowing async trades in agent0, such that multiple transactions can be submitted on the same block.
Breaking changes
agent.liquidate
toagent.execute_liquidate
.Changes
get_policy_action
andget_liquidate_action
) and executing an agent's actions (i.e.,execute_action
) into two separate steps, withexecute_policy_action
andexecute_liquidate
calling the two steps sequentially.manual_database_sync
toLocalChain.Config
to allow for manual database syncing instead of syncing after every trade.agent.open_long
) no longer eagerly gets the wallet, and instead only gets wallet information if the trade crashes (the wallet was only being used for crash reporting).build_transaction
, and set nonce after. This fixes an issue with requiring transactions to be built in order.