The read function

pesummary provides functionality to read in nearly all result file formats. This is done with the function. Below we show how the read function works and some of the functions and properties that it includes.

Reading a result file

Below, we show how to read in a result file,

>>> from import read
>>> data = read("example.dat", package="core")
>>> json_load = read("example.json", package="core")
>>> hdf5_load = read("example.hdf5", package="core")
>>> txt_load = read("example.txt", package="core")

pesummary is able to read in json, dat, txt and hdf5 file formats.

Extracting data

Once read in, pesummary offers properties and functions to extract the data that is stored in the input file. As pesummary is able to read in both a file that contains the samples for a single analysis and a pesummary metafile which is able to store samples from multiple analyses, the way to extract the samples is slightly different for the two situations.

For both cases, the stored samples can be inspected through the .samples_dict property

Single analysis file

When the input file only contains a single analysis, the samples are stored All samples are stored in a subclass of the Python builtin dict: pesummary.utils.samples_dict.SamplesDict,

>>> samples_dict = data.samples_dict
>>> type(samples_dict)
<class 'pesummary.utils.samples_dict.SamplesDict'>
>>> print(samples_dict.keys())
dict_keys(['a', 'b',...])

Each marginalized posterior distribution is stored as a subclass of numpy.ndarray: pesummary.utils.samples_dict.Array. For more information about the SamplesDict class, see the SamplesDict docs. For more information about the Array class, see the Array docs. This structure provides direct access to the optimised array functions from numpy and the usability and familiarity of dictionaries.

PESummary metafile

When the input file is a pesummary metafile, this time, the samples are stored in a different subclass of the Python builtin dict: pesummary.utils.samples_dict.MultiAnalysisSamplesDict,

>>> samples_dict = data.samples_dict
>>> type(samples_dict)
<class 'pesummary.utils.samples_dict.MultiAnalysisSamplesDict'>
>>> print(samples_dict.keys())
dict_keys(['analysis_1', 'analysis_2',...])
>>> type(samples_dict['analysis_1'])
<class 'pesummary.utils.samples_dict.SamplesDict'>

The samples for a given analysis and now assigned a label in order to distinguish them from another analysis. The samples for a given analysis are stored as a pesummary.utils.samples_dict.SamplesDict object. For more information about the MultiAnalysisSamplesDict class, see the MultiAnalysisSamplesDict socs.

For details about how to extract additional information from the pesummary metafile, see the extract information from a pesummary file docs.

Changing file format

We may convert the result file to a different file format by using the .write() method. For example, we may convert to a .dat file with,

>>> data.write(package="core", file_format="dat", outdir="./", filename="example.dat")


pesummary provides the ability to interact with latex and produce a latex table displaying the results or latex macros to aid in writing publications. You are able to specify which parameters you want included in the latex table as well as detailed descrptions of what each parameter is. Below is an example for how to make a latex table,

>>> description_mapping = {
...     "a": "The a parameter",
...     "b": "The b parameter"
... }
>>> data.to_latex_table(parameter_dict=description_mapping)
\begin{tabular}{l c }
The a parameter & $35.04^{+8.00}_{-5.00}$\\
The b parameter & $76.01^{+7.56}_{-0.45}$\\
>>> data.to_latex_table(save_to_file="table.tex", parameter_dict=description_mapping)

And latex macros may be generated as follows,

>>> macros_map = {"a": "A"}
>>> data.generate_latex_macros(parameter_dict=macros_map)
>>> data.generate_latex_macros(save_to_file="macros.tex", parameter_dict=macros_map)