Skip to content

Latest commit

 

History

History
57 lines (48 loc) · 2.73 KB

TODO.md

File metadata and controls

57 lines (48 loc) · 2.73 KB
Raw

Wishlist

Here we collect ideas to improve and extend kuibit. If you want to contribute to the project, this a good place where to start. The projects are sorted in no particular order. The number of [=] along with each idea indicates the expected difficulty. This parameter increases with the level of knowledge of kuibit or of Python required to complete the task.

Features

  • Function to compute spectrogram of TimeSeries. [=]

  • The extrapolation to infinity function for gravitational waves has to be tested and can be extended to support generic strains (not only for fixed l, m). [==]

  • Improve algorithm for __call__ in Series and grid_data to be more Pythonic and faster. [==]

  • Extend Series and grid_data to support array data instead of only scalar data. [====]

  • Correctly identify and merge refinement levels in HierarchicalGridData even where there are multiple centers of refinement. [===]

  • Add layer in class hierarchy in cactus_grid_functions to save the work done in reading ASCII files containing multiple variables. [==]

  • Port cactus_parfile from PostCactus. [==]

  • Transparently handle low dimensional grid data with higher dimensional one is available (e.g. add to the xy if xyz is available). [==]

  • Add support for HDF5 for AHFinderDirect output. [==]

  • Add method to merge AHFinderDirect patches. [==]

  • Perform interpolation in shapes of AHFinderDirect to better find shapes when the cut is not on a major direction. [==]

  • Add support for all_reductions_in_one_file to cactus_scalars. [==]

  • Add support for VolumeIntegral thorns to cactus_scalars. [==]

  • Hunt for TODOs in the codebase and implement them. [=?=]

Infrastructure

  • Improve errors with dynamical name types. [==] (e.g. type(self).__name__ instead of hardconding the name)
  • Improve docstrings. [==] (According to PEP8, the first line should be short and descriptive.) Check that all the sphinx param descriptions end with a period.
  • Simplify apply_unary, apply_binary, apply_to_self in numerical.py. These functions may be turned into decorators, or at least used in a more concise way. [==]
  • Numba-ify low-level functions. [====]
  • _multipoles_from_textfiles takes a long time if there are thousands of files. [==]
  • cactus_scalars, cactus_multipoles, cactus_grid_functions have a lot of common infrastructure. There should be a way to reduce code deduplication. Also, at the moment, the three group of classes are structured in very different ways, they should be more uniform. [====]
  • Add support for logging. [===]
  • Refactor code to correctly use np.array or np.asarray (not use np.array when is not needed). [==]
  • Improve performance of HierarchicalGridData.all_components. [==]