EasyTPP Config Modules

class config_factory.Config[source]

Bases: Registrable

save_to_yaml_file(fn)[source]

Save the config into the yaml file ‘fn’.

Parameters:: fn (str) – Target filename.

Returns:

static build_from_yaml_file(yaml_fn, **kwargs)[source]

Load yaml config file from disk.

Parameters:: yaml_fn (str) – Path of the yaml config file.
Returns:: Config object corresponding to cls.
Return type:: EasyTPP.Config

abstract get_yaml_config()[source]

Get the yaml format config from self.

Returns:

abstract static parse_from_yaml_config(yaml_config)[source]

Parse from the yaml to generate the config object.

Parameters:: yaml_config (dict) – configs from yaml file.
Returns:: Config class for data.
Return type:: EasyTPP.Config

abstract copy()[source]

Get a same and freely modifiable copy of self.

Returns:

update(config)[source]

Update the config.

Parameters:: config (dict) – config dict.
Returns:: Config class for data.
Return type:: EasyTPP.Config

pop(key: str, default_var: Any)[source]

pop out the key-value item from the config.

Parameters:

key (str) – key name.
default_var (Any) – default value to pop.

Returns:

value to pop.

Return type:

Any

get(key: str, default_var: Any)[source]

Retrieve the key-value item from the config.

Parameters:

key (str) – key name.
default_var (Any) – default value to pop.

Returns:

value to get.

Return type:

Any

set(key: str, var_to_set: Any)[source]

Set the key-value item from the config.

Parameters:

key (str) – key name.
var_to_set (Any) – default value to pop.

Returns:

value to get.

Return type:

Any

class config_factory.DataConfig(train_dir, valid_dir, test_dir, specs=None)[source]

Bases: Config

__init__(train_dir, valid_dir, test_dir, specs=None)[source]

Initialize the DataConfig object.

Parameters:

train_dir (str) – dir of tran set.
valid_dir (str) – dir of valid set.
test_dir (str) – dir of test set.
specs (dict, optional) – specs of dataset. Defaults to None.

get_yaml_config()[source]

Return the config in dict (yaml compatible) format.

Returns:: config of the data in dict format.
Return type:: dict

static parse_from_yaml_config(yaml_config)[source]

Parse from the yaml to generate the config object.

Parameters:: yaml_config (dict) – configs from yaml file.
Returns:: Config class for data.
Return type:: EasyTPP.DataConfig

copy()[source]

Copy the config.

Returns:: a copy of current config.
Return type:: EasyTPP.DataConfig

get_data_dir(split)[source]

Get the dir of the source raw data.

Parameters:: split (str) – dataset split notation, ‘train’, ‘dev’ or ‘valid’, ‘test’.
Returns:: dir of the source raw data file.
Return type:: str

class config_factory.DataSpecConfig(**kwargs)[source]

Bases: Config

__init__(**kwargs)[source]: Initialize the Config class.

get_yaml_config()[source]

Return the config in dict (yaml compatible) format.

Returns:: config of the data specs in dict format.
Return type:: dict

static parse_from_yaml_config(yaml_config)[source]

Parse from the yaml to generate the config object.

Parameters:: yaml_config (dict) – configs from yaml file.
Returns:: Config class for data specs.
Return type:: DataSpecConfig

copy()[source]

Copy the config.

Returns:: a copy of current config.
Return type:: DataSpecConfig

class config_factory.ModelConfig(**kwargs)[source]

Bases: Config

__init__(**kwargs)[source]: Initialize the Config class.

get_yaml_config()[source]

Return the config in dict (yaml compatible) format.

Returns:: config of the model config specs in dict format.
Return type:: dict

static parse_from_yaml_config(yaml_config)[source]

Parse from the yaml to generate the config object.

Parameters:: yaml_config (dict) – configs from yaml file.
Returns:: Config class for trainer specs.
Return type:: ModelConfig

copy()[source]

Copy the config.

Returns:: a copy of current config.
Return type:: ModelConfig

class config_factory.BaseConfig(**kwargs)[source]

Bases: Config

__init__(**kwargs)[source]: Initialize the Config class.

static set_backend(backend)[source]

get_yaml_config()[source]

Return the config in dict (yaml compatible) format.

Returns:: config of the base config specs in dict format.
Return type:: dict

static parse_from_yaml_config(yaml_config)[source]

Parse from the yaml to generate the config object.

Parameters:: yaml_config (dict) – configs from yaml file.
Returns:: Config class for trainer specs.
Return type:: BaseConfig

copy()[source]

Copy the config.

Returns:: a copy of current config.
Return type:: BaseConfig

class config_factory.RunnerConfig(base_config, model_config, data_config, trainer_config)[source]

Bases: Config

__init__(base_config, model_config, data_config, trainer_config)[source]

Initialize the Config class.

Parameters:

base_config (EasyTPP.BaseConfig) – BaseConfig object.
model_config (EasyTPP.ModelConfig) – ModelConfig object.
data_config (EasyTPP.DataConfig) – DataConfig object.
trainer_config (EasyTPP.TrainerConfig) – TrainerConfig object

get_yaml_config()[source]

Return the config in dict (yaml compatible) format.

Returns:: config of the runner config in dict format.
Return type:: dict

static parse_from_yaml_config(yaml_config, **kwargs)[source]

Parse from the yaml to generate the config object.

Parameters:: yaml_config (dict) – configs from yaml file.
Returns:: Config class for trainer specs.
Return type:: RunnerConfig

ensure_valid_config()[source]: Do some sanity check about the config, to avoid conflicts in settings.

update_config()[source]: Updated config dict.

get_metric_functions()[source]

get_metric_direction(metric_name='rmse')[source]

copy()[source]

Copy the config.

Returns:: a copy of current config.
Return type:: RunnerConfig

class config_factory.HPOConfig(framework_id, storage_uri, is_continuous, num_trials, num_jobs)[source]

Bases: Config

__init__(framework_id, storage_uri, is_continuous, num_trials, num_jobs)[source]

Initialize the HPO Config

Parameters:

framework_id (str) – hpo framework id.
storage_uri (str) – result storage dir.
is_continuous (bool) – whether to continuously do the optimization.
num_trials (int) – num of trails used in optimization.
num_jobs (int) – num of the jobs.

property storage_protocol

Get the storage protocol

Returns:: the dir of the storage protocol.
Return type:: str

property storage_path

Get the storage protocol

Returns:: the dir of the hpo data storage.
Return type:: str

get_yaml_config()[source]

Return the config in dict (yaml compatible) format.

Returns:: config of the HPO specs in dict format.
Return type:: dict

static parse_from_yaml_config(yaml_config, **kwargs)[source]

Parse from the yaml to generate the config object.

Parameters:: yaml_config (dict) – configs from yaml file.
Returns:: Config class for HPO specs.
Return type:: EasyTPP.HPOConfig

copy()[source]

Copy the config.

Returns:: a copy of current config.
Return type:: EasyTPP.HPOConfig

class config_factory.HPORunnerConfig(hpo_config, runner_config)[source]

Bases: Config

__init__(hpo_config, runner_config)[source]

Initialize the config class

Parameters:

hpo_config (EasyTPP.HPOConfig) – hpo config class.
runner_config (EasyTPP.RunnerConfig) – runner config class.

static parse_from_yaml_config(yaml_config, **kwargs)[source]

Parse from the yaml to generate the config object.

Parameters:: yaml_config (dict) – configs from yaml file.
Returns:: Config class for HPO specs.
Return type:: EasyTPP.HPORunnerConfig

copy()[source]

Copy the config.

Returns:: a copy of current config.
Return type:: EasyTPP.HPORunnerConfig