Reference / API¶

Pre-processing¶

Performing feature extraction for 1 light curve

`LightCurve`()	Light Curve object, holding meta and photometric data.
`LightCurve.load_snpcc_lc`(path_to_data)	Reads one LC from SNPCC data.
`LightCurve.fit_bazin`(band)	Extract Bazin features for one filter.
`LightCurve.fit_bazin_all`()	Perform Bazin fit for all filters independently and concatenate results.
`LightCurve.plot_bazin_fit`([save, show, …])	Plot data and Bazin fitted function.

Fitting an entire data set

fit_snpcc_bazin(path_to_data_dir, features_file) Fit Bazin functions to all filters in training and test samples.

Basic light curve analysis tools

`bazin.bazin`(time, a, b, t0, tfall, trise)	Parametric light curve function proposed by Bazin et al., 2009.
`bazin.errfunc`(params, time, flux)	Absolute difference between theoretical and measured flux.
`bazin.fit_scipy`(time, flux)	Find best-fit parameters using scipy.least_squares.

The Canonical object for holding the entire sample.

`Canonical`()	Canonical sample object.
`Canonical.snpcc_get_canonical_info`(…[, …])	Load SNPCC metada data required to characterize objects.
`Canonical.snpcc_identify_samples`()	Identify training and test sample.
`Canonical.find_neighbors`()	Identify 1 nearest neighbor for each object in training.

Functions to populate the Canonical object

`build_snpcc_canonical`(path_to_raw_data, …)	Build canonical sample for SNPCC data.
`plot_snpcc_train_canonical`(sample[, …])	Plot comparison between training and canonical samples.

`SNPCCPhotometry`()	Handles photometric information for entire SNPCC data.
`SNPCCPhotometry.get_lim_mjds`(raw_data_dir)	Get minimum and maximum MJD for complete sample.
`SNPCCPhotometry.create_daily_file`(…[, header])	Create one file for a given day of the survey.
`SNPCCPhotometry.build_one_epoch`(…[, …])	Fit bazin for all objects with enough points in a given day.

Object upon which the learning process is performed

`DataBase`()	DataBase object, upon which the active learning loop is performed.
`DataBase.load_bazin_features`(path_to_bazin_file)	Load Bazin features from file.
`DataBase.load_features`(path_to_file[, …])	Load features according to the chosen feature extraction method.
`DataBase.build_samples`(initial_training[, …])	Separate train and test samples.
`DataBase.classify`([method, screen, n_est, …])	Apply a machine learning classifier.
`DataBase.evaluate_classification`([…])	Evaluate results from classification.
`DataBase.make_query`([strategy, batch, seed, …])	Identify new object to be added to the training sample.
`DataBase.update_samples`(query_indx, loop[, …])	Add the queried obj(s) to training and remove them from test.
`DataBase.save_metrics`(loop, …[, batch])	Save current metrics to file.
`DataBase.save_queried_sample`(…[, …])	Save queried sample to file.

random_forest(train_features, train_labels, …) Random Forest classifier.

`random_sampling`(test_ids, queryable_ids[, …])	Randomly choose an object from the test sample.
`uncertainty_sampling`(class_prob, test_ids, …)	Search for the sample with highest uncertainty in predicted class.

Individual metrics

`accuracy`(label_pred, label_true)	Calculate accuracy.
`efficiency`(label_pred, label_true[, ia_flag])	Calculate efficiency.
`purity`(label_pred, label_true[, ia_flag])	Calculate purity.
`fom`(label_pred, label_true[, ia_flag, penalty])	Calculate figure of merit.

Metrics agregated by category or use

get_snpcc_metric(label_pred, label_true[, …]) Calculate the metric parameters used in the SNPCC.

Full light curve

learn_loop(nloops, strategy, …[, …]) Perform the active learning loop.

Time domain

`get_original_training`(path_to_features[, …])	Read original full light curve training sample
`time_domain_loop`(days, output_metrics_file, …)	Perform the active learning loop.

`Canvas`()	Canvas object, handles and plot information from multiple strategies.
`Canvas.load_metrics`(path_to_files, …[, …])	Load and identify set of metrics.
`Canvas.set_plot_dimensions`()	Set directives for plot sizes.
`Canvas.plot_metrics`(output_plot_file, …)	Generate plot for all metrics in files and strategies given as input.

`build_canonical`(user_choices)	Build canonical sample for SNPCC data set fitted with Bazin features.
`build_time_domain`(user_choice)	Generates features files for a list of days of the survey.
`fit_dataset`(user_choices)	Fit the entire sample with the Bazin function.
`make_metrics_plots`(user_input)	Generate metric plots.
`run_loop`(args)	Command line interface to run the active learning loop.
`run_time_domain`(user_choice)	Command line interface to the Time Domain Active Learning scenario.