pyabsa.tasks.UniversalSentimentAnalysis

Subpackages

Package Contents

Classes

USAConfigManager

Simple object for storing attributes.

USATrainingDataset

USADatasetList

The following datasets are for aspect polarity classification task.

USATrainingInstructor

USAModelList

Built-in mutable sequence.

USAPredictor

USATrainer

Trainer class for training PyABSA models
class pyabsa.tasks.UniversalSentimentAnalysis.USAConfigManager(args, **kwargs)[source]

Bases: pyabsa.framework.configuration_class.configuration_template.ConfigManager

Simple object for storing attributes.

Implements equality by attribute names and values, and provides a simple string representation.

static set_usa_config(configType: str, newitem: dict)[source]
static set_usa_config_template(newitem)[source]
static set_usa_config_base(newitem)[source]
static set_usa_config_english(newitem)[source]
static set_usa_config_chinese(newitem)[source]
static set_usa_config_multilingual(newitem)[source]
static get_usa_config_template()[source]
static get_usa_config_base()[source]
static get_usa_config_english()[source]
static get_usa_config_chinese()[source]
static get_usa_config_multilingual()[source]
class pyabsa.tasks.UniversalSentimentAnalysis.USATrainingDataset(config, tokenizer, dataset_type='train', **kwargs)[source]
load_data_from_dict(dataset_dict, **kwargs)[source]
load_data_from_file(dataset_file, **kwargs)[source]
tokenize_function_inputs(sample)[source]

Udf to tokenize the input dataset.

class pyabsa.tasks.UniversalSentimentAnalysis.USADatasetList[source]

Bases: list

The following datasets are for aspect polarity classification task. The datasets are collected from different sources, you can use the id to locate the dataset.

Laptop14
Restaurant14
Restaurant15
Restaurant16
Chinese_Zhang
Multilingual
Synthetic
class pyabsa.tasks.UniversalSentimentAnalysis.USATrainingInstructor(config)[source]

Bases: pyabsa.framework.instructor_class.instructor_template.BaseTrainingInstructor

_load_dataset_and_prepare_dataloader()[source]

Load the dataset and prepare the dataloader. This method should be implemented in a subclass.

run()[source]
class pyabsa.tasks.UniversalSentimentAnalysis.USAModelList[source]

Bases: list

Built-in mutable sequence.

If no argument is given, the constructor creates a new empty list. The argument must be an iterable if specified.

GenerationModel
class pyabsa.tasks.UniversalSentimentAnalysis.USAPredictor(checkpoint=None, **kwargs)[source]

Bases: pyabsa.framework.prediction_class.predictor_template.InferenceModel

task_code
batch_infer(target_file=None, print_result=True, save_result=False, ignore_error=True, **kwargs)[source]

A deprecated version of batch_predict method.

Parameters:

  • target_file (str) – the path to the target file for inference

  • print_result (bool) – whether to print the result

  • save_result (bool) – whether to save the result

  • ignore_error (bool) – whether to ignore the error

Returns:

a dictionary of the results

Return type:

result (dict)

infer(text: str = None, print_result=True, ignore_error=True, **kwargs)[source]

A deprecated version of the predict method.

Parameters:

  • text (str) – the text to predict

  • print_result (bool) – whether to print the result

  • ignore_error (bool) – whether to ignore the error

Returns:

a dictionary of the results

Return type:

result (dict)

batch_predict(target_file=None, print_result=True, save_result=False, ignore_error=True, **kwargs)[source]

Predict the sentiment from a file of sentences. param: target_file: the file path of the sentences to be predicted. param: print_result: whether to print the result. param: save_result: whether to save the result. param: ignore_error: whether to ignore the error when predicting. param: kwargs: other parameters.

predict(text: str | list = None, print_result=True, ignore_error=True, **kwargs)[source]

Predict the sentiment from a sentence or a list of sentences. param: text: the sentence to be predicted. param: print_result: whether to print the result. param: ignore_error: whether to ignore the error when predicting. param: kwargs: other parameters.

_run_prediction(save_path=None, print_result=True, **kwargs)[source]

This method should be implemented in the subclass for running predictions using the trained model.

Parameters:

kwargs – additional keyword arguments

Returns:

predicted labels or other prediction outputs

clear_input_samples()[source]
class pyabsa.tasks.UniversalSentimentAnalysis.USATrainer(config: pyabsa.tasks.UniversalSentimentAnalysis.configuration.configuration.USAConfigManager = None, dataset=None, from_checkpoint: str = None, checkpoint_save_mode: int = ModelSaveOption.SAVE_MODEL_STATE_DICT, auto_device: bool | str = DeviceTypeOption.AUTO, path_to_save=None, load_aug=False)[source]

Bases: pyabsa.framework.trainer_class.trainer_template.Trainer

Trainer class for training PyABSA models