Files
NexQuant/rdagent/app/data_science/loop.py
T
Linlang 805f337fc4 chore: add a rdagent server with UI & logger storage refinement(#553)
* change_log_object

* lint code

* delete comments

* change_log_object

* change_log_object

* fix import test error

* update code

* update code

* fix bugs

* skip mypy error

* skip mypy error

* skip mypy error

* Start the flask server before running the demo.

* achieve front and back interaction

* fix github-advanced-security comments

* fix github-advanced-security comments

* tmp ignore

* fix CI

* move some logic

* change format

* adjust logic

* log2json changes

* tmp

* fix

* fix bug

* refine log2json between 5 scenarios

* fix

* refine codes

* fix logic

* use localhost

* add loop & all_duration param for old scenario startup

* merge control logic

* add README for server ui api

* update README

* reuse code in logger

* add loop_n and all_duration param

* fix upload

* ui server now use port in setting

* fix port setting

* fix port setting

* fix mypy check

* refine logger and log storage

* fix ruff error

* fix CI

* refine logger, loop, storage

* bind one FileStorage with one logger

* not truncate log storage

* refine LoopBase.load(), use `checkout` instead of `output_path` and `do_truncate`

* clear session folder when loading loop to run

* move component info init step to ExpGen Class

* Update rdagent/utils/workflow.py

* move truncate_session function to LoopBase class

* add checkout param for other scenarios

* fix bug

* move WebStorage to UI

* change web_storage name

* add randomname to requirements

* add typer

* fix requirements

---------

Co-authored-by: WinstonLiyte <1957922024@qq.com>
Co-authored-by: Bowen Xian <xianbowen@outlook.com>
Co-authored-by: you-n-g <you-n-g@users.noreply.github.com>
2025-06-06 18:52:19 +08:00

74 lines
3.0 KiB
Python

from pathlib import Path
import fire
from rdagent.app.data_science.conf import DS_RD_SETTING
from rdagent.core.utils import import_class
from rdagent.log import rdagent_logger as logger
from rdagent.scenarios.data_science.loop import DataScienceRDLoop
def main(
path: str | None = None,
checkout: bool | str | Path = True,
step_n: int | None = None,
loop_n: int | None = None,
competition="bms-molecular-translation",
timeout=None,
replace_timer=True,
exp_gen_cls: str | None = None,
):
"""
Parameters
----------
path :
A path like `$LOG_PATH/__session__/1/0_propose`. This indicates that we restore the state after finishing step 0 in loop 1.
checkout :
Used only when a path is provided.
Can be True, False, or a path.
Default is True.
- If True, the new loop will use the existing folder and clear logs for sessions after the one corresponding to the given path.
- If False, the new loop will use the existing folder but keep the logs for sessions after the one corresponding to the given path.
- If a path (or a str like Path) is provided, the new loop will be saved to that path, leaving the original path unchanged.
step_n :
Number of steps to run; if None, the process will run indefinitely until an error or KeyboardInterrupt occurs.
loop_n :
Number of loops to run; if None, the process will run indefinitely until an error or KeyboardInterrupt occurs.
- If the current loop is incomplete, it will be counted as the first loop for completion.
- If both step_n and loop_n are provided, the process will stop as soon as either condition is met.
competition :
Competition name.
replace_timer :
If a session is loaded, determines whether to replace the timer with session.timer.
exp_gen_cls :
When there are different stages, the exp_gen can be replaced with the new proposal.
Auto R&D Evolving loop for models in a Kaggle scenario.
You can continue running a session by using the command:
.. code-block:: bash
dotenv run -- python rdagent/app/data_science/loop.py [--competition titanic] $LOG_PATH/__session__/1/0_propose --step_n 1 # `step_n` is an optional parameter
rdagent kaggle --competition playground-series-s4e8 # This command is recommended.
"""
if competition is not None:
DS_RD_SETTING.competition = competition
if not DS_RD_SETTING.competition:
logger.error("Please specify competition name.")
if path is None:
kaggle_loop = DataScienceRDLoop(DS_RD_SETTING)
else:
kaggle_loop: DataScienceRDLoop = DataScienceRDLoop.load(path, checkout=checkout, replace_timer=replace_timer)
# replace exp_gen if we have new class
if exp_gen_cls is not None:
kaggle_loop.exp_gen = import_class(exp_gen_cls)(kaggle_loop.exp_gen.scen)
kaggle_loop.run(step_n=step_n, loop_n=loop_n, all_duration=timeout)
if __name__ == "__main__":
fire.Fire(main)