API Reference

This contains a reference document to the floatCSEP API.

Commands

The commands and entry-points with which to call floatcsep from the terminal are:

run(config, **kwargs)
plot(config, **kwargs)
reproduce(config, **kwargs)

Experiment

The experiment is defined using the Experiment class.

Experiment([name, time_config, ...])	Main class that handles an Experiment's context.
Experiment.set_models(model_config[, order])	Parse the models' configuration file/dict.
Experiment.get_model(name)	Returns a Model by its name string
Experiment.stage_models()	Stages all the experiment's models.
Experiment.set_tests(test_config)	Parse the tests' configuration file/dict.
Experiment.catalog	Returns a CSEP catalog loaded from the given query function or a stored file if it exists.
Experiment.set_test_cat(tstring)	Filters the complete experiment catalog to a test sub-catalog bounded by the test time-window.
Experiment.set_tasks()	Lazy definition of the experiment core tasks by wrapping instances, methods and arguments.
Experiment.run()	Run the task tree
Experiment.read_results(test, window)	Reads an Evaluation result for a given time window and returns a list of the results for all tested models.
Experiment.plot_results()	Plots all evaluation results
Experiment.plot_catalog([dpi, show])	Plots the evaluation catalogs
Experiment.plot_forecasts()	Plots and saves all the generated forecasts
Experiment.generate_report()	Creates a report summarizing the Experiment's results
Experiment.to_yml(filename, **kwargs)	Serializes the Experiment instance into a .yml file.
Experiment.from_yml(config_yml[, reprdir])	Initializes an experiment from a .yml file.

Models

A model is defined using the Model class.

Model(name, model_path[, forecast_unit, ...])	Class defining a forecast generating Model.
Model.get_source([zenodo_id, giturl, force])	Search, download or clone the model source in the filesystem, zenodo and git, respectively.
Model.stage([timewindows])	Pre-steps to make the model runnable before integrating
Model.init_db([dbpath, force])	Initializes the database if use_db is True.
Model.rm_db()	Clean up the generated HDF5 File
Model.get_forecast([tstring, region])	Wrapper that just returns a forecast, which should hide the access method (db storage, ti_td, etc.) under the hood
Model.create_forecast(tstring, **kwargs)	Creates a forecast from the model source and a given time window
Model.forecast_from_func(start_date, ...)
Model.forecast_from_file(start_date, ...)	Generates a forecast from a file, by parsing and scaling it to the desired time window.
Model.from_dict(record, **kwargs)	Returns a Model instance from a dictionary containing the required atrributes.

Evaluations

A test is defined using the Evaluation class.

Evaluation(name, func[, func_kwargs, ...])	Class representing a Scoring Test, which wraps the evaluation function, its arguments, parameters and hyper-parameters.
Evaluation.type	Returns the type of the test, mapped from the class attribute Evaluation._TYPES
Evaluation.get_catalog(catalog_path, forecast)	Reads the catalog(s) from the given path(s).
Evaluation.prepare_args(timewindow, catpath, ...)	Prepares the positional argument for the Evaluation function.
Evaluation.compute(timewindow, catalog, ...)	Runs the test, structuring the arguments according to the
Evaluation.write_result(result, path)	Dumps a test result into a json file.
Evaluation.from_dict(record)	Parses a dictionary and re-instantiate an Evaluation object

Accessors

query_gcmt(start_time, end_time[, ...])
from_zenodo(record_id, folder[, force])	Download data from a Zenodo repository.
from_git(url, path[, branch, depth])	Clones a shallow repository from a git url

Extras

Additional pyCSEP functionalities

sequential_likelihood(gridded_forecasts, ...)	Performs the likelihood test on Gridded Forecast using an Observed Catalog.
sequential_information_gain(...[, seed, ...])	param gridded_forecasts: list csep.core.forecasts.GriddedForecast
vector_poisson_t_w_test(forecast, ...)	Computes Student's t-test for the information gain per earthquake over a list of forecasts and w-test for normality
brier_score(forecast, catalog[, ...])
negative_binomial_number_test(...)	Computes "negative binomial N-Test" on a gridded forecast.
binomial_joint_log_likelihood_ndarray(...)	Computes Bernoulli log-likelihood scores, assuming that earthquakes follow a binomial distribution.
binomial_spatial_test(gridded_forecast, ...)	Performs the binary spatial test on the Forecast using the Observed Catalogs.
binomial_conditional_likelihood_test(...[, ...])	Performs the binary conditional likelihood test on Gridded Forecast using an Observed Catalog.
binary_paired_t_test(forecast, ...[, alpha, ...])	Computes the binary t-test for gridded earthquake forecasts.
log_likelihood_point_process(observation, ...)	Log-likelihood for point process
paired_ttest_point_process(forecast, ...[, ...])	Function for T test based on Point process LL.

Utilities

parse_csep_func(func)	Searchs in pyCSEP and floatCSEP a function or method whose name matches the provided string.
parse_timedelta_string(window[, exp_class])	Parses a float or string representing the testing time window length
timewindows_ti([start_date, end_date, ...])	Creates the testing intervals for a time-independent experiment.
timewindows_td([start_date, end_date, ...])	Creates the testing intervals for a time-dependent experiment.
Task(instance, method, **kwargs)
Task.run()
Task.check_exist()
timewindow2str(datetimes)	Converts a time window (list/tuple of datetimes) to a string that represents it.
plot_sequential_likelihood(evaluation_results)
magnitude_vs_time(catalog)

Readers

ForecastParsers.dat(filename)
ForecastParsers.xml(filename[, verbose])
ForecastParsers.quadtree(filename)
ForecastParsers.csv(filename)
ForecastParsers.hdf5(filename[, group])
HDF5Serializer.grid2hdf5(rates, region, mag)
serialize()