Reference / API¶
Datasets¶
Loading/saving a dataset
Dataset (name, metadata[, observations, …]) |
A dataset of many astronomical objects. |
Dataset.load (name[, metadata_only, chunk, …]) |
Load a dataset that has been saved in HDF5 format in the data directory. |
Dataset.from_objects (name, objects, **kwargs) |
Load a dataset from a list of AstronomicalObject instances. |
Dataset.path |
Return the path to where this dataset should lie on disk |
Dataset.write ([overwrite]) |
Write the dataset out to disk. |
Retrieving objects from the dataset
Dataset.get_object ([index, object_class, …]) |
Parse keywords to pull a specific object out of the dataset |
Plotting lightcurves of objects in the dataset
Dataset.plot_light_curve (*args, **kwargs) |
Plot the light curve for an object in the dataset. |
Dataset.plot_interactive () |
Make an interactive plot of the light curves in the dataset. |
Extracting features from objects in the dataset
Dataset.extract_raw_features (featurizer[, …]) |
Extract raw features from the dataset. |
Dataset.get_raw_features_path ([tag]) |
Return the path to where the raw features for this dataset should lie on disk |
Dataset.write_raw_features ([tag]) |
Write the raw features out to disk. |
Dataset.load_raw_features ([tag]) |
Load the raw features from disk. |
Dataset.select_features (featurizer) |
Select features from the dataset for classification. |
Classifying objects in the dataset
Dataset.predict (classifier) |
Generate predictions using a classifier. |
Dataset.get_predictions_path ([classifier]) |
Return the path to where the predicitons for this dataset for a given classifier should lie on disk. |
Dataset.write_predictions ([classifier]) |
Write predictions for this classifier to disk. |
Dataset.load_predictions ([classifier]) |
Load the predictions for a classifier from disk. |
Dataset.label_folds ([num_folds, random_state]) |
Separate the dataset into groups for k-folding |
Astronomical objects¶
AstronomicalObject (metadata, observations) |
An astronomical object, with metadata and a lightcurve. |
AstronomicalObject.bands |
Return a list of bands that this object has observations in |
AstronomicalObject.subtract_background () |
Subtract the background levels from each band. |
AstronomicalObject.preprocess_observations ([…]) |
Apply preprocessing to the observations. |
AstronomicalObject.fit_gaussian_process ([…]) |
Fit a Gaussian Process model to the light curve. |
AstronomicalObject.get_default_gaussian_process () |
Get the default Gaussian Process. |
AstronomicalObject.predict_gaussian_process (…) |
Predict the Gaussian process in a given set of bands and at a given set of times. |
AstronomicalObject.plot_light_curve ([…]) |
Plot the object’s light curve |
AstronomicalObject.print_metadata () |
Print out the object’s metadata in a nice format. |
Dataset augmentation¶
Augmentor API
Augmentor (**cosmology_kwargs) |
Class used to augment a dataset. |
Augmentor.augment_object (reference_object[, …]) |
Generate an augmented version of an object. |
Augmentor.augment_dataset (augment_name, …) |
Generate augmented versions of all objects in a dataset. |
Augmentor Implementations
plasticc.PlasticcAugmentor () |
Implementation of an Augmentor for the PLAsTiCC dataset |
Augmentor methods to implement in subclasses
Augmentor._augment_metadata (reference_object) |
Generate new metadata for the augmented object. |
Augmentor._choose_sampling_times (…[, …]) |
Choose the times at which to sample for a new augmented object. |
Augmentor._choose_target_observation_count (…) |
Choose the target number of observations for a new augmented light curve. |
Augmentor._simulate_light_curve_uncertainties (…) |
Simulate the observation-related noise for a light curve. |
Augmentor._simulate_detection (observations, …) |
Simulate the detection process for a light curve. |
Classification¶
Classifier API
Classifier (name) |
Classifier used to classify the different objects in a dataset. |
Classifier.train (dataset) |
Train the classifier on a dataset |
Classifier.predict (dataset) |
Generate predictions for a dataset |
Classifier.path |
Get the path to where a classifier should be stored on disk |
Classifier.write ([overwrite]) |
Write a trained classifier to disk |
Classifier.load (name) |
Load a classifier that was previously saved to disk |
Classifier Implementations
LightGBMClassifier (name, featurizer[, …]) |
Feature based classifier using LightGBM to classify objects. |
Weights and metrics
evaluate_weights_flat (dataset[, class_weights]) |
Evaluate the weights to use for classification on a dataset. |
evaluate_weights_redshift (dataset[, …]) |
Evaluate redshift-weighted weights to use to generate a rates-independent classifier. |
weighted_multi_logloss (true_classes, predictions) |
Evaluate a weighted multi-class logloss function. |
Feature extraction¶
Featurizer API
Featurizer |
Class used to extract features from objects. |
Featurizer.extract_raw_features (…[, …]) |
Extract raw features from an object |
Featurizer.select_features (raw_features) |
Select features to use for classification |
Featurizer.extract_features (astronomical_object) |
Extract features from an object. |
Featurizer Implementations
plasticc.PlasticcFeaturizer |
Class used to extract features for the PLAsTiCC dataset. |
Exceptions¶
AvocadoException |
The base class for all exceptions raised in avocado. |