Feat/benchmarking convergence #382
Conversation
- Simulating bet placement and queue status update in bet placement behaviour for benchamrking
- move bet to fresh if benchmarking mode is enabled as for the given bet no transaction was found in the dataset
- as process queue can be small and we might not reach number of mech calls, we will start considering reprocessed bets also.
Adamantios
force-pushed
the
feat/benchmarking-convergence
branch
from
3e985ad to
a7e01fd
Compare
| elif "id" in data_attributes: | ||
| common_attributes = set(bet_annotations) & set(data_attributes) | ||
| data = {key: data[key] for key in common_attributes} | ||
| if "queue_status" in data: | ||
| data["queue_status"] = QueueStatus(data["queue_status"]) | ||
| return Bet(**data) |
There was a problem hiding this comment.
Can you explain this part?
There was a problem hiding this comment.
Please also add comments to explain this in the code.
There was a problem hiding this comment.
@Adamantios it was during the time we had a discussion that bet.json didn't have backward compatibility and all you and Anna were still working on the sell token feature and were going to handle this later. basically the new investments field in Bets dataclass was also being treated as a dataclass instance and files couldn't be read during benchmarking mode.
@annasambrook this can be replaced with the current logic of bets on main, this was a backport fix for then because this branch was created out of sell token feature branch that had investments as a dict with investment values for both yes and no.
I can see that chnage has been reverted so this code is not needed.
packages/valory/skills/decision_maker_abci/behaviours/bet_placement.py
Outdated
Show resolved
Hide resolved
This branch fixes the Benchmarking mode running at the trader. It introduces some changes regarding