Skip to content
This repository has been archived by the owner on Aug 6, 2024. It is now read-only.

Latest commit

 

History

History
360 lines (237 loc) · 11.2 KB

CHANGES.rst

File metadata and controls

360 lines (237 loc) · 11.2 KB
Raw

0.1 (2020-11-17)

Adopt Project-template from @nstarman

New Features

discO.type_hints

  • Add types for type hinting: [#17]

    • EllipsisType : type(Ellipsis) b/c ellipsis fails.

    • UnitType : the type of Astropy's UnitBase and FunctionUnitBasse

    • RepresentationOrDifferentialType : the type of Astropy's BaseRepresentationOrDifferential [#34]

    • RepresentationType : the type of Astropy's BaseRepresentation [#34]

    • RepresentationLikeType : RepresentationType or str [#42]

    • DifferentialType : the type of Astropy's BaseDifferential [#34]

    • FrameType : the type of Astropy's BaseCoordinateFrame

    • SkyCoordType : the type of Astropy's SkyCoord

    • CoordinateType : the union of FrameType & SkyCoordType

    • GenericPosiionType : RepresentationOrDifferentialType or CoordinateType [#34]

    • FrameLikeType : the union of CoordinateType & parseable str

      • anything that can be used in frame= in Skycoord(...,frame=)

    • TableType : the type of Astropy's Table [#34]

    • QTableType : the type of Astropy's QTable [#34]

    • UnitType : the type of Astropy's UnitBase or FunctionUnitType [#34]

    • UnitLikeType : UnitType or parseable str [#34]

    • QuantityType : the type of astropy's Quantity

    • QuantityType : QuantityType or parseable str [#34]

  • changed location from common to type_hints [#34]

discO.config

  • Add configuration for default frame ("icrs"). [#17]
  • Add configuration for default representation type ("cartesian") [#45]

discO.core

Where the unified architecture is defined.

Modules:

  • core : the base class. [#17]
  • sample : for sampling from a Potential. [#17]
  • measurement : for resampling, given observational errors. [#17]
  • fitter : for fitting a Potential given a sample [#20]
  • residual : residual between fit and fiducial potential
  • statistic : statistic functions for collapsing a residual to a single number [#50]
  • pipeline : for comboing the analysis [#19]
  • wrapper : for wrapping potentials [#45]

discO.core.core

The base class samplers, fitters, and most everything else. If a package is defined as a class argument, it parses the package and stores it in the class. This is used for registering classes into registry classes.

subclasses must override the _registry and __call__ methods.

discO.core.core

  • CommonBase : base class

    • provides tools for working with class registries

discO.core.sample

  • PotentialSampler : base class for sampling potentials [#17]

    • registers subclasses. Each subclass is for sampling from potentials from a different package. Eg. GalpyPotentialSampler for sampling galpy potentials.
    • PotentialSampler can be used to initialize any of its subclasses.
    • Takes a potential and a frame (astropy CoordinateFrame). The potential is used for sampling, but the resulting points are not located in any reference frame, which we assign with frame.
    • Can also specify representation type [#43]
    • __call__ and sample are used to sample the potential
    • sample samples the potential many times. This can be done for many iterations and different sample number points.
    • frame and representation_type can be None or Ellipse or anything that works with resolve_framelike. [#45]

  • MeshGridSampler : sample an arbitrary potential with a weighted grid [#59]

discO.core.fitter

  • PotentialFitter : base class for fitting potentials [#20]

    • registers subclasses.
    • PotentialFitter can be used to initialize any of its subclasses. [#44]
    • Takes a potential_cls and key argument which are used to figure out the desired subclass, and how to fit the potential.
    • __call__ and fit are used to fit the potential, with the latter working on N-D samples (multiple iterations).
    • returns a PotentialWrapper [#40]
    • Allow for frame and representation. Care should be taken this matches the sampling frame. [#45]
    • frame and representation_type can be None or Ellipse or anything that works with resolve_framelike. [#45]

discO.core.measurement

  • MeasurementErrorSampler : base class for resampling a potential given measurement errors [#17]

    • registers subclasses. Each subclass is for resampling in a different way.
    • MeasurementErrorSampler is a registry wrapper class and can be used in-place of any of its subclasses.
    • Add method resample for ND array samples from PotentialSampler [#38]
    • frame and representation_type can be None or Ellipse or anything that works with resolve_framelike. [#45]
    • c_err must be a keyword argument. [#45]

  • RVS_Continuous : scipy rv_continuous distribution [#42]

    • Any scipy rv_continuous distribution.
    • rvs must be a keyword argument. [#45]

  • GaussianMeasurementError : Gaussian rvs distribution [#42]

    • should work for any normal distribution (if has "norm") in name.

  • xpercenterror_factory : to build xpercenterror function. [#36] Convenience function for construct errors with X% error in each dimension.

discO.core.pipeline

  • Pipeline : run a full analysis pipeline [#19]

    • PotentialSampler to MeasurementErrorSampler to PotentialFitter to ResidualMethod to statistic. [#19,#26]
    • Pipelines can also be created by concatenation.
    • Pipeline can take arguments frame and representation_type. [#45]
    • Calling pipeline can take arguments observer versions of frame and representation_type. [#45]
    • frame and representation_type can be None or Ellipse or anything that works with resolve_framelike. [#45]
    • convenience properties for potential, frame, representation_type, potential_frame, potential_representation_type, observer_frame, observer_representation_type, sampler, measurer, fitter, residualer, statisticer. [#45]
    • Add method run_iter to iteratively call pipeline. [#26]
    • Add method run_with_samples to skip sampling step [#51]

  • PipelineResult store results of a pipe [#37]

    • produced by Pipeline at end of a run or call.
    • convenience properties for samples, potential_frame, potential_representation_type, measured, observation_frame, observation_representation_type, fit, residual, statistic. [#45]

discO.core.residual

  • ResidualMethod : calculate a residual [#26]
    • difference between original and fit potential
  • GridResidual : calculate a residual on a pre-defined grid [#26]
    • difference between original and fit potential
    • need pre-defined grid

discO.core.statistic

  • rms : calculate the RMS [#50]

discO.core.wrapper

  • PotentialWrapper : base class for wrapping Potentials [#39]

    • unified interface for the potential and specific force. [#39,#49]
    • all methods are both instance and static methods.
    • specific force returns a vector field.
    • frame and representation_type can be None or Ellipse or anything that works with resolve_framelike. [#45]
    • total_mass function. [#45]
    • density function [#56]

discO.data

  • Add Milky_Way_Sim_100 data [#10]

discO.plugin

Where classes for external packages are held.

discO.plugin.agama

  • AGAMAPotentialSampler [#17]

    • Sample from agama potentials.
    • Subclass of PotentialSampler
    • stores the mass and potential as attributes on the returned SkyCoord

  • AGAMAPotentialFitter [#20]

    • Fit agama potentials.
    • Subclass of PotentialFitter
    • registers subclasses for different fit methods.
    • AGAMAPotentialFitter can be used to initialize any of its subclasses.
    • Takes a pot_type argument which is used to figure out the desired subclass, and how to fit the potential.
    • returns a AGAMAPotentialWrapper [#40]

  • AGAMAMultipolePotentialFitter [#20]

    • Fit agama potentials with a multipole
    • Subclass of AGAMAPotentialFitter

  • AGAMAPotentialWrapper : for wrapping Potentials [#39]

    • unified interface for the potential and specific force.
    • all methods are both instance and static methods.
    • specific force returns a vector field.
    • total_mass function. [#45]

discO.plugin.gala

  • AGAMAPotentialWrapper : for wrapping Potentials [#58]

    • unified interface for the potential and specific force.
    • all methods are both instance and static methods.
    • specific force returns a vector field.

discO.plugin.galpy

  • GalpyPotentialSampler [#17]

    • Sample from galpy potentials with a corresponding distribution function.
    • stores the mass and potential as attributes on the returned SkyCoord
  • MeshGridPositionDF : mimics a galpy DF [#59]

    For sampling positions using a disco MeshGridSampler.

  • GalpyPotentialWrapper : for wrapping Potentials [#39]

    • unified interface for the potential and specific force.
    • all methods are both instance and static methods.
    • specific force returns a vector field.
    • total_mass function. [#45]

  • GalpySCFPotentialFitter : for fitting an SCF to particles [#41]

    • fit galpy SCF potential
    • returns a GalpyPotentialWrapper with the specified frame.

discO.utils

  • resolve_framelike [#17]

    • Determine the frame and return a blank instance for anything that can be used in frame= in Skycoord(...,frame=).
    • Ellipsis resolves to the configured default frame ("icrs"). [#45]
    • None becomes UnFrame() [#45]

  • resolve_representationlike [#42]

    • Determine the representation type given a class, instance, or string name.
    • Ellipsis uses default representation type ("cartesian") [#45]

  • UnFrame : unconnected generic coordinate frame [#43]

    • For use when no reference frame is specified.

  • vector fields [#35]

    For transforming vector fields between coordinate systems (eg Cartesian to spherical). Built on top of Astropy's Representation machinery.

  • NumpyRNGContext : astropy's, extended to RandomState s [#43]

API Changes

N/A

Bug Fixes

N/A

Docs

  • Added glossary [#17]

    • 'frame-like'
    • 'coord-like'
    • 'coord scalar' and 'coord-like scalar'
    • 'coord array' and 'coord-like array'

Other Changes and Additions

  • Alphabetize name in credits [#8]

  • PR Template [#5]

    • Updated [#11]

  • Use GitHub for CI [#12]

    • On tag [#17]

  • Dependabot yml [#13]

  • Issues Templates [#14]

  • Update from project template [#18]

  • Add .mailmap [#17]

Actions