cogwheel.utils¶

Utility functions.

Functions

`clear_caches`	Clear caches of functions decorated with `lru_cache`.
`exp_normalize`	Return normalized probabilities from unnormalized log probabilities.
`get_eventdir`	Return pathlib.Path object for a directory of the form parentdir/prior_name/eventname/.
`get_priordir`	Return pathlib.Path object for a directory of the form `parentdir/prior_name`.
`get_rundir`	Return path to a directory where a new inference run can take place.
`handle_scalars`	Return a function that outputs a scalar instead of 0-d array.
`import_lal`	Import lal, disabling stdout/stderr redirection if in ipython.
`lru_cache`	Cache function results and register it for easy cache clearing.
`merge_dictionaries_safely`	Merge multiple dictionaries into one.
`mkdirs`	Create directory and its parents if needed.
`mod`	Modulus operation, but allows to specify the domain of the output.
`n_effective`	Calculate the effective sample size from a set of weights.
`quantile`	Compute the q-th quantile of the data.
`read_json`	Return a class instance that was saved to json.
`real_matmul`	Return real part of complex matrix multiplication.
`replace`	Replace specified elements in a sequence with new values.
`resample_equal`	Resample a weighted DataFrame to produce equal-weight samples.
`rundir_number`	Return first strech of numbers in rundir as `int`.
`sorted_rundirs`	Return rundirs sorted by number (i.e. 'run_2' before 'run_10').
`submit_condor`	Generic function to submit a job using HTCondor.
`submit_lsf`	Generic function to submit a job using IBM Spectrum LSF.
`submit_slurm`	Generic function to submit a job using slurm.
`temporarily_change_attributes`	Change attributes of an object temporarily in a context manager.
`update_dataframe`	Modify a dataframe in-place by adding the columns from another.
`weighted_avg_and_std`	Return average and standard deviation of values with weights.

Classes

`ClassProperty`	Like `@property` but for class attributes instead of instance attributes.
`CogwheelDecoder`	Decoder for subclasses of JSONMixin.
`CogwheelEncoder`	Encoder for subclasses of JSONMixin.
`InitDictMixin`	Provide methods `get_init_dict` and `reinstantiate`.
`JSONMixin`	Provide JSON output to subclasses.
`NumpyEncoder`	Encoder for numpy data types.