Skip to content

removed future improvements list #9

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
leifvan wants to merge 1 commit into
base: main
Choose a base branch
from
Open

removed future improvements list #9

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
33 changes: 0 additions & 33 deletions README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -206,36 +206,3 @@ Another benefit from using the typed version is that tooltips in VS Code are mor
* [ ] ``get_shape_variables`` does not work if checks are disabled. This should be possible but give a performance warning, recommending not to use this feature in performance-critical applications.
* [x] Variable stack collects int values as variables and prints them, i.e. it produces error messages like ``"Shape torch.Size([1, 4, 4]) does not match descriptor (b, 4, 4) at position 0 based on already inferred variables {'b': 4, 'n': 304, 2: 2, 1: 1, 4: 4}"``

## Future Improvements
These are feature that are not implemented yet, but might be added in future releases.

* [x] dtype annotation
* [ ] add tests for autogenerated constraints and come up with a specific
syntax to enable it (or enable it by default?)
* [x] make exception messages more concise and remove currently used exception
reraise
* [ ] improve annotation handling for method overrides in subclasses
* [x] add tests for frameworks other than PyTorch
* [x] torch
* [x] numpy
* [x] jax
* [x] dask
* [x] donnx
* [x] sparse
* ~~[ ] cupy~~ (we leave this out for now because it requires CUDA)
* [x] check compatibility with static type checkers
* [ ] rewrite README to give a cleaner overview over the features
* [ ] support union of shape descriptors (but this might break the current simplicity)
* [ ] benchmark speed to understand impact in tight loops
* [x] compatibility for torch.compile (or at least auto-disable check)
* ~~[ ] device annotation~~ (device definition not standardized in Python array API 2024.12, see [this section of the specifications](https://data-apis.org/array-api/2024.12/design_topics/device_support.html#device-support))
* [ ] add variable names for dtype
* [ ] add more helpful message when parameter / output is not the expected type
* [x] (maybe instead of the one before) catch and reraise all exceptions inside
wrapper and reraise with additional info about exception location
* [x] improve hints for static type checking (currently it assumes either torch
or the object just having a .shape parameter)
* [ ] make get_shape_variables work in check modes "never" and "once" without
performance overhead
* [ ] work on performance overhead in general