Still under construction
This package allows one to import a number of observational datasets in a format compatible with IMAGINE.
One can the available datasets using the package's show_available
function.
import imagine_datasets as img_data
# Shows a list of available packages
img_data.show_available()
This will print the details of all the available datasets to screen. For example, the part of this output corresponding to the Oppermann2012 Faraday depth map is:
-------------------imagine_datasets.HEALPix.fd.Oppermann2012--------------------
Class: imagine_datasets.HEALPix.fd.Oppermann2012
Observable: fd
Type: HEALPix
Bibliographic ref: Oppermann et al. (2012) A&A, 542, A93
URL: https://ui.adsabs.harvard.edu/abs/2012A&A...542A..93O/abstract
A dataset can be used in an IMAGINE inference pipeline by simply
instantiating its corresponding class and including it in a
Measurements
object, for example,
import imagine as img
import imagine_datasets as img_data
dset = img_data.HEALPix.fd.Oppermann2012(Nside=64)
Once loaded, the dataset contents can be used as a regular IMAGINE dataset:
measurements = img.observables.Measurements(dset)
Additionally, the datasets in this repository come with handy attributes which point to the original publications
citation = dset.ref
citation_url = dset.ref_url
Datasets in this repository also support caching. Usually, a dataset is downloaded when it is requested (i.e. instantiated) for the first time. If a cache directory is set, the data is saved to disk and, thus, will not need to be downloaded again.
One can choose cache directory setting the module variable
imagine_datasets.cache_dir
or using the environment variable
IMAGINE_DATASETS_CACHE_PATH
, for example including the following line in
your .bashrc file:
export IMAGINE_DATASETS_CACHE_DIR=foo/bar/DatasetCacheDir