[Bug] mutable default is not allowed in @dataclass in python > 3.9.

[ga92xug]

🐛 Bug

Description

mutable default is not allowed in @dataclass in python > 3.9. I would be happy to do a pull request.

Checklist

• I checked on the latest version of Hydra
• I created a minimal repro.

To reproduce

** Minimal Code/Config snippet to reproduce **
min.py

from typing import Optional
import hydra
from omegaconf import DictConfig 

@hydra.main(version_base="1.3", config_path=".", config_name="conf.yaml")
def main(cfg: DictConfig) -> Optional[float]:

    print(cfg.pretty())
    return None

if __name__ == "__main__":
    main()

conf.yaml

# @package hydra.sweeper
_target_: hydra_plugins.hydra_ax_sweeper.ax_sweeper.AxSweeper
max_batch_size: null
ax_config:
  max_trials: 10
  early_stop:
    minimize: true
    max_epochs_without_improvement: 10
    epsilon: 1.0e-05
  experiment:
    name: null
    objective_name: objective
    minimize: true
    parameter_constraints: null
    outcome_constraints: null
    status_quo: null
  client:
    verbose_logging: false
    random_seed: null
  is_noisy: true
  params: {}

** Stack trace/error message **

Traceback (most recent call last):
  File "/home/stefan/git/min_example_hydra_ax_bug/min.py", line 12, in <module>
    main()
  File "/home/stefan/anaconda3/envs/hydra_ax_bug_py_3_11/lib/python3.11/site-packages/hydra/main.py", line 94, in decorated_main
    _run_hydra(
  File "/home/stefan/anaconda3/envs/hydra_ax_bug_py_3_11/lib/python3.11/site-packages/hydra/_internal/utils.py", line 337, in _run_hydra
    search_path = create_automatic_config_search_path(
                  ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/stefan/anaconda3/envs/hydra_ax_bug_py_3_11/lib/python3.11/site-packages/hydra/_internal/utils.py", line 190, in create_automatic_config_search_path
    return create_config_search_path(search_path_dir)
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/stefan/anaconda3/envs/hydra_ax_bug_py_3_11/lib/python3.11/site-packages/hydra/_internal/utils.py", line 203, in create_config_search_path
    search_path_plugins = Plugins.instance().discover(SearchPathPlugin)
                          ^^^^^^^^^^^^^^^^^^
  File "/home/stefan/anaconda3/envs/hydra_ax_bug_py_3_11/lib/python3.11/site-packages/hydra/core/plugins.py", line 46, in instance
    ret = Singleton.instance(Plugins, *args, **kwargs)
          ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/stefan/anaconda3/envs/hydra_ax_bug_py_3_11/lib/python3.11/site-packages/hydra/core/singleton.py", line 17, in instance
    return cls(*args, **kwargs)
           ^^^^^^^^^^^^^^^^^^^^
  File "/home/stefan/anaconda3/envs/hydra_ax_bug_py_3_11/lib/python3.11/site-packages/hydra/core/singleton.py", line 13, in __call__
    cls._instances[cls] = super(Singleton, cls).__call__(*args, **kwargs)
                          ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/stefan/anaconda3/envs/hydra_ax_bug_py_3_11/lib/python3.11/site-packages/hydra/core/plugins.py", line 54, in __init__
    self._initialize()
  File "/home/stefan/anaconda3/envs/hydra_ax_bug_py_3_11/lib/python3.11/site-packages/hydra/core/plugins.py", line 71, in _initialize
    scanned_plugins, self.stats = self._scan_all_plugins(modules=top_level)
                                  ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/stefan/anaconda3/envs/hydra_ax_bug_py_3_11/lib/python3.11/site-packages/hydra/core/plugins.py", line 197, in _scan_all_plugins
    spec.loader.exec_module(loaded_mod)
  File "<frozen importlib._bootstrap_external>", line 940, in exec_module
  File "<frozen importlib._bootstrap>", line 241, in _call_with_frames_removed
  File "/home/stefan/anaconda3/envs/hydra_ax_bug_py_3_11/lib/python3.11/site-packages/hydra_plugins/hydra_ax_sweeper/ax_sweeper.py", line 8, in <module>
    from .config import AxConfig
  File "/home/stefan/anaconda3/envs/hydra_ax_bug_py_3_11/lib/python3.11/site-packages/hydra_plugins/hydra_ax_sweeper/config.py", line 44, in <module>
    @dataclass
     ^^^^^^^^^
  File "/home/stefan/anaconda3/envs/hydra_ax_bug_py_3_11/lib/python3.11/dataclasses.py", line 1230, in dataclass
    return wrap(cls)
           ^^^^^^^^^
  File "/home/stefan/anaconda3/envs/hydra_ax_bug_py_3_11/lib/python3.11/dataclasses.py", line 1220, in wrap
    return _process_class(cls, init, repr, eq, order, unsafe_hash,
           ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/stefan/anaconda3/envs/hydra_ax_bug_py_3_11/lib/python3.11/dataclasses.py", line 958, in _process_class
    cls_fields.append(_get_field(cls, name, type, kw_only))
                      ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
  File "/home/stefan/anaconda3/envs/hydra_ax_bug_py_3_11/lib/python3.11/dataclasses.py", line 815, in _get_field
    raise ValueError(f'mutable default {type(f.default)} for field '
ValueError: mutable default <class 'hydra_plugins.hydra_ax_sweeper.config.EarlyStopConfig'> for field early_stop is not allowed: use default_factory

Expected Behavior

System information

• Hydra Version : 1.3.2
• Python version : 3.11.8
• Virtual environment type and version : conda
• Operating system : ubuntu 22.04

Fix

config.py

# Copyright (c) Facebook, Inc. and its affiliates. All Rights Reserved
from dataclasses import dataclass, field
from typing import Any, Dict, List, Optional

from hydra.core.config_store import ConfigStore


@dataclass
class EarlyStopConfig:
    minimize: bool = True

    max_epochs_without_improvement: int = 10

    # An improvement larger than epsilon is considered significant
    epsilon: float = 0.00001


@dataclass
class ExperimentConfig:
    # Experiment name
    name: Optional[str] = None

    # Name of metric to optimize or null if you only return one metric
    objective_name: str = "objective"

    # Defaults to minimize, set to false to maximize
    minimize: bool = True

    # For the remaining parameters, refer the Ax documentation: https://ax.dev/api/core.html#experiment
    parameter_constraints: Optional[List[str]] = None
    outcome_constraints: Optional[List[str]] = None
    status_quo: Optional[Dict[str, Any]] = None


@dataclass
class ClientConfig:

    verbose_logging: bool = False

    # set random seed here to make Ax results reproducible
    random_seed: Optional[int] = None


@dataclass
class AxConfig:

    # max_trials is application-specific. Tune it for your use case
    max_trials: int = 10
    early_stop: EarlyStopConfig = field(default_factory=EarlyStopConfig)
    experiment: ExperimentConfig = field(default_factory=ExperimentConfig)
    client: ClientConfig = field(default_factory=ClientConfig)
    params: Dict[str, Any] = field(default_factory=dict)
    # is_noisy = True indicates measurements have unknown uncertainty
    # is_noisy = False indicates measurements have an uncertainty of zero
    is_noisy: bool = True


@dataclass
class AxSweeperConf:
    _target_: str = "hydra_plugins.hydra_ax_sweeper.ax_sweeper.AxSweeper"
    # Maximum number of trials to run in parallel
    max_batch_size: Optional[int] = None
    ax_config: AxConfig = field(default_factory=AxConfig)


ConfigStore.instance().store(
    group="hydra/sweeper", name="ax", node=AxSweeperConf, provider="ax_sweeper"
)

[Jasha10]

Possibly related: #2777

[ga92xug]

Ah yes that is right. Overlooked that. I will close that.