geowatch.utils.lightning_ext.lightning_cli_ext module

This module is an exension of jsonargparse and lightning CLI that will respect scriptconfig style arguments

References

https://github.com/Lightning-AI/lightning/issues/15038

class geowatch.utils.lightning_ext.lightning_cli_ext.LightningArgumentParser_Extension(*args: Any, description: str = 'Lightning Trainer command line tool', env_prefix: str = 'PL', default_env: bool = False, **kwargs: Any)[source]

Bases: ArgumentParserPatches, LightningArgumentParser

CommandLine

xdoctest -m geowatch.utils.lightning_ext.lightning_cli_ext LightningArgumentParser_Extension

Example

>>> from geowatch.utils.lightning_ext.lightning_cli_ext import *  # NOQA
>>> LightningArgumentParser_Extension()
Refactor references:

~/.pyenv/versions/3.10.5/envs/pyenv3.10.5/lib/python3.10/site-packages/pytorch_lightning/cli.py ~/.pyenv/versions/3.10.5/envs/pyenv3.10.5/lib/python3.10/site-packages/jsonargparse/core.py ~/.pyenv/versions/3.10.5/envs/pyenv3.10.5/lib/python3.10/site-packages/jsonargparse/signatures.py

Initialize argument parser that supports configuration file input.

For full details of accepted arguments see ArgumentParser.__init__.

Parameters:

  • description – Description of the tool shown when running --help.

  • env_prefix – Prefix for environment variables. Set default_env=True to enable env parsing.

  • default_env – Whether to parse environment variables.

class geowatch.utils.lightning_ext.lightning_cli_ext.LightningCLI_Extension(model_class: ~typing.Type[~pytorch_lightning.core.module.LightningModule] | ~typing.Callable[[...], ~pytorch_lightning.core.module.LightningModule] | None = None, datamodule_class: ~typing.Type[~pytorch_lightning.core.datamodule.LightningDataModule] | ~typing.Callable[[...], ~pytorch_lightning.core.datamodule.LightningDataModule] | None = None, save_config_callback: ~typing.Type[~pytorch_lightning.cli.SaveConfigCallback] | None = <class 'pytorch_lightning.cli.SaveConfigCallback'>, save_config_kwargs: ~typing.Dict[str, ~typing.Any] | None = None, trainer_class: ~typing.Type[~pytorch_lightning.trainer.trainer.Trainer] | ~typing.Callable[[...], ~pytorch_lightning.trainer.trainer.Trainer] = <class 'pytorch_lightning.trainer.trainer.Trainer'>, trainer_defaults: ~typing.Dict[str, ~typing.Any] | None = None, seed_everything_default: bool | int = True, parser_kwargs: ~typing.Dict[str, ~typing.Any] | ~typing.Dict[str, ~typing.Dict[str, ~typing.Any]] | None = None, subclass_mode_model: bool = False, subclass_mode_data: bool = False, args: ~typing.List[str] | ~typing.Dict[str, ~typing.Any] | ~jsonargparse.namespace.Namespace | None = None, run: bool = True, auto_configure_optimizers: bool = True)[source]

Bases: LightningCLI

Our customized LightningCLI extension.

Receives as input pytorch-lightning classes (or callables which return pytorch-lightning classes), which are called / instantiated using a parsed configuration file and / or command line args.

Parsing of configuration from environment variables can be enabled by setting parser_kwargs={"default_env": True}. A full configuration yaml would be parsed from PL_CONFIG if set. Individual settings are so parsed from variables named for example PL_TRAINER__MAX_EPOCHS.

For more info, read the CLI docs.

Parameters:

  • model_class – An optional LightningModule class to train on or a callable which returns a LightningModule instance when called. If None, you can pass a registered model with --model=MyModel.

  • datamodule_class – An optional LightningDataModule class or a callable which returns a LightningDataModule instance when called. If None, you can pass a registered datamodule with --data=MyDataModule.

  • save_config_callback – A callback class to save the config.

  • save_config_kwargs – Parameters that will be used to instantiate the save_config_callback.

  • trainer_class – An optional subclass of the Trainer class or a callable which returns a Trainer instance when called.

  • trainer_defaults – Set to override Trainer defaults or add persistent callbacks. The callbacks added through this argument will not be configurable from a configuration file and will always be present for this particular CLI. Alternatively, configurable callbacks can be added as explained in the CLI docs.

  • seed_everything_default – Number for the seed_everything() seed value. Set to True to automatically choose a seed value. Setting it to False will avoid calling seed_everything.

  • parser_kwargs – Additional arguments to instantiate each LightningArgumentParser.

  • subclass_mode_model – Whether model can be any subclass of the given class.

  • subclass_mode_data – Whether datamodule can be any subclass of the given class.

  • args – Arguments to parse. If None the arguments are taken from sys.argv. Command line style arguments can be given in a list. Alternatively, structured config options can be given in a dict or jsonargparse.Namespace.

  • run – Whether subcommands should be added to run a Trainer method. If set to False, the trainer and model classes will be instantiated only.

init_parser(**kwargs)[source]
parse_arguments(parser: LightningArgumentParser, args) → None[source]

Parses command line arguments and stores it in self.config.