Skip to content

[DOC] Document less common in-memory data formats #7386

@fkiraly

Description

@fkiraly

The recent PR #7231 adds an API reference page for in-memory data representation.

With this, detailed documentation on the most commonly used time series machine types were also added.

However, documentation for the less common formats is still missing - we should add this.

Recipe:

  1. pick one type and help document it! (one pull request per type). See below for a list of undocumented types.
  2. use get_examples from sktime.datatypes to generate some examples, and look at MTYPE_REGISTER for a short description. You can also use convert_to to see how already documented types convert to the documented one.
  3. add a docstring to the class in one of the _check modules, e.g., sktime.datatypes._panel._check for one of the Panel mtypes. The name of the class is the same as below. All modules already contain complete examples, so "follow the pattern".

To identify a missing case, see the list below.

Alternatively, you can also:

  • search for classes with empty docstrings in one of the datasets _check modules, within the _series, _panel, _hierarchical or _table submodules.
  • look at the "Data Format Specifications" page on the API reference (link), click through concrete representations and search for empty docstrings.

Docstrings should be completed with a description of the in-memory type. Example descriptions can be found for the more common types, in the API reference, or the aforementioned modules.

Classes currently empty:

  • SeriesXArray
  • SeriesDask
  • SeriesPolarsEager
  • SeriesGluontsList
  • SeriesGluontsPandas
  • PanelDask
  • PanelPolarsEager
  • PanelGluontsList
  • PanelGluontsPandas
  • HierarchicalDask
  • HierarchicalPolarsEager
  • TablePdDataFrame
  • TablePdSeries
  • TableNp1D
  • TableNp2D
  • TableListOfDict
  • TablePolarsEager

FYI @Abhay-Lejith and @pranavvp16, as you wrote some of the respective plug-ins - helping with the docs would be much appreciated, since you know most about these!

Metadata

Metadata

Assignees

No one assigned

    Labels

    documentationDocumentation & tutorialsgood first issueGood for newcomersmodule:datatypesdatatypes module: data containers, checkers & converters

    Type

    No type

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions