tests/test_data · ocelot · Tangled

emily.space / ocelot

[READ-ONLY] Mirror of https://github.com/emilyhunt/ocelot. A toolbox for working with observations of star clusters. ocelot-docs.org

ocelot tests test_data

4 folders · 10 files

at feat-model-docs

Emily Hunt · Merge pull request #2 from emilyhunt/feat--cluster-simulation · 1y ago

Merge pull request #2 from emilyhunt/feat--cluster-simulation

Changed the DataPartitioner to work with HEALPix pixels. - Re-wrote much of the code to now use HEALPix. It is *not* backward compatible! - Made some efficiency improvements to how it calculates parallax members (more stuff is saved to the class for later use) - Unit tested and verified that the new changes work as intended. - Wrote some really neat plotting code for the unit testing that I should totally integrate into the class itself...

Added a new metrics subpackage to ocelot.cluster. - Added a very basic (and probably hilariously slow) 5 parameter metric to cluster.epsilon. This commit is experimental work towards 0.3.0.

simulated_populations

Work on synthetic cluster generation. - Synthetic populations are now being correctly generated! The King profile was changed to be drawn from 3 times at once (i.e. in 3D.) - The class' runtime was sped up by a factor of ~15, by creating a sped up (check-free) version of the King profile, dramatically reducing the number of function calls by drawing a total number of points based on the expectation value of the accept/rejects, and by dropping stars below a conservative limiting magnitude very early in the process (making distant & large clusters much smaller.) This commit is experimental work for 0.3.0

blanco_1_gaia_dr2_gmag_18_cut_cluster_only.pickle

Work on ocelot.calculate. - Implemented a basic version of .radius - Wrote unit tests for all functions in the submodule - Fixed a handful of small bugs in the implementations - Re-factored .all_statistics() so it can use different modes in the future. Eventually, it would be lovely to infer many of these parameters in a Bayesian way - All functions pass all tests =) This release satisfies 0.3.0 feature c.

blanco_1_gaia_dr2_gmag_18_cut_LABELS.pickle

Unit tests added for ocelot.plot along with a number of bugfixes

blanco_1_gaia_dr2_gmag_18_cut.pickle

Unit tests added for ocelot.plot along with a number of bugfixes

crossmatch_match_all_axes

Final, final work on ocelot.crossmatch to fix bugs and make it as usable as possible. - Now using the ocelot standard parameter names in output match tables from Catalogue.crossmatch() for consistency, instead of using the column names of the literature catalogue. This commit satisfies 0.3.0 feature d.

crossmatch_match_position_only

Final work on ocelot.crossmatch to fix bugs and make it as usable as possible. - Added a systematic position error estimate parameter to the Catalogue class at __init__ - Changed Catalogue.crossmatch() to be much simpler to call and to now use ocelot standard parameter names, which are now stored in ocelot.calculate.constants - These parameter names can be overriden (or even added to!) with Catalogue.override_default_ocelot_parameter_names() - Increased the stats in the crossmatch process - now powered only by sigma values, ignoring tidal radii and looking for the mean or max sigma to pick a best match - Catalogue.crossmatch() is now unit tested (more of an integration test =p) after lots of checking values by hand and looking for bugs! This commit satisfies 0.3.0 feature d.

crossmatch_summary_all_axes

Final work on ocelot.crossmatch to fix bugs and make it as usable as possible. - Added a systematic position error estimate parameter to the Catalogue class at __init__ - Changed Catalogue.crossmatch() to be much simpler to call and to now use ocelot standard parameter names, which are now stored in ocelot.calculate.constants - These parameter names can be overriden (or even added to!) with Catalogue.override_default_ocelot_parameter_names() - Increased the stats in the crossmatch process - now powered only by sigma values, ignoring tidal radii and looking for the mean or max sigma to pick a best match - Catalogue.crossmatch() is now unit tested (more of an integration test =p) after lots of checking values by hand and looking for bugs! This commit satisfies 0.3.0 feature d.

crossmatch_summary_position_only

Final work on ocelot.crossmatch to fix bugs and make it as usable as possible. - Added a systematic position error estimate parameter to the Catalogue class at __init__ - Changed Catalogue.crossmatch() to be much simpler to call and to now use ocelot standard parameter names, which are now stored in ocelot.calculate.constants - These parameter names can be overriden (or even added to!) with Catalogue.override_default_ocelot_parameter_names() - Increased the stats in the crossmatch process - now powered only by sigma values, ignoring tidal radii and looking for the mean or max sigma to pick a best match - Catalogue.crossmatch() is now unit tested (more of an integration test =p) after lots of checking values by hand and looking for bugs! This commit satisfies 0.3.0 feature d.

healpix_12237.csv

Added HEALPix support to ocelot.cluster.recenter_dataset. - Can now specify pixel_id instead of the center of a field for HEALPix pixels - Can rotate the HEALPix pixel to have the edges of the boundary aligned with the latitude/longitude axes. - Passed its tests! This commit is additional contributing work for 0.3.0.

mwsc_ii_catalogue

Work on ocelot.crossmatch - Have now written a basic crossmatch utility in the form of a class with methods, ocelot.crossmatch.Catalogue - It should be able to crossmatch with an arbitrary number of features, depending on the catalogue to be matched against - It requires testing! This commit partially satisfies 0.3.0 feature d.

mwsc_ii_perturbed

Work on ocelot.crossmatch - Basic bugfixing to ocelot.crossmatch.Catalogue (it now works as written and intended), but I'd like to refactor it to work in more sensible ways (e.g. using an estimated systematic error on position to calculate positional probabilities instead of using the tidal radius as a proxy.) - Next commit will be quite different! This commit partially satisfies 0.3.0 feature d.