jax.numpy delivers ediff1d on mathematical functions

[BrookeMa]

Hi, Because the force push caused conflicts, so I make a new PR again. Please help me to review my code. Thank you
Close #11072

[Daniel4078]

you are supposed to add a frontend function for ediff1d of jax right? why did you add those backend experimental functions instead?

[BrookeMa]

Hi, Daniel. Because some unit tests case didn't pass. (Can't reach 100% pass) It's related to the backend. So I did it. Can I implement the backend together or separate it into two PR?

[Daniel4078]

well, I need to see your frontend version of ediff1d first

[Daniel4078]

also, you can not just add functions to our experimental module, all the functions that we plan to have there is listed here: #3856

[BrookeMa]

well, I need to see your frontend version of ediff1d first

Okay, thanks. I will delete the code related to the backend. And make a new PR. How about my frontend functions? Does it have any problems?

[Daniel4078]

It looks find to me, but I suppose it only works with your backend code right?

[BrookeMa]

It looks find to me, but I suppose it only works with your backend code right?

Yes, If I delete the backend functions, it will fail the current test case. If possible, can I only test the frontend functions?

[BrookeMa]

It looks find to me, but I suppose it only works with your backend code right?

This is my test case. I refer to diff.

/Users/yema/miniconda3/envs/ivy_dev/bin/python /Applications/PyCharm.app/Contents/plugins/python/helpers/pydev/pydevd.py --multiprocess --qt-support=auto --client 127.0.0.1 --port 50981 --file /Applications/PyCharm.app/Contents/plugins/python/helpers/pycharm/_jb_pytest_runner.py --target ivy_tests/test_ivy/test_frontends/test_jax/test_jax_numpy_math.py::test_jax_numpy_ediff1d 
Testing started at 14:31 ...
Connected to pydev debugger (build 223.8617.48)
Launching pytest with arguments ivy_tests/test_ivy/test_frontends/test_jax/test_jax_numpy_math.py::test_jax_numpy_ediff1d --no-header --no-summary -q in /Users/yema/Desktop/ivy

Use Database in ReadOnly Mode with local caching !
============================= test session starts ==============================
collecting ... collected 4 items

ivy_tests/test_ivy/test_frontends/test_jax/test_jax_numpy_math.py::test_jax_numpy_ediff1d[cpu-ivy.functional.backends.numpy-False-False] 
ivy_tests/test_ivy/test_frontends/test_jax/test_jax_numpy_math.py::test_jax_numpy_ediff1d[cpu-ivy.functional.backends.jax-False-False] 
ivy_tests/test_ivy/test_frontends/test_jax/test_jax_numpy_math.py::test_jax_numpy_ediff1d[cpu-ivy.functional.backends.tensorflow-False-False] 
ivy_tests/test_ivy/test_frontends/test_jax/test_jax_numpy_math.py::test_jax_numpy_ediff1d[cpu-ivy.functional.backends.torch-False-False] 

======================== 4 passed, 1 warning in 16.24s =========================

Process finished with exit code 0
PASSED [ 25%]PASSED [ 50%]PASSED [ 75%]PASSED [100%]

[Daniel4078]

ok, just to make sure, have you tried using the existing functions in Ivy's module directly or compositionally to achieve the desired behavior? because we might be about to skip the problem of adding a new set of backend functions and tests

[BrookeMa]

ok, just to make sure, have you tried using the existing functions in Ivy's module directly or compositionally to achieve the desired behavior? because we might be about to skip the problem of adding a new set of backend functions and tests

Yes, I didn't use ivy's functions on my implementation function. What should I do now? Delete the backend functions and make a new PR, right?

[Daniel4078]

well you can try remove the unnecessary changes on this branch if you know how to. but you are free to create a new branch and pull request

[BrookeMa]

well you can try remove the unnecessary changes on this branch if you know how to. but you are free to create a new branch and pull request

Great, thank you

[BrookeMa]

well you can try remove the unnecessary changes on this branch if you know how to. but you are free to create a new branch and pull request

I have deleted four functions about the backend of ediff1d. Can you help me review it again? Is there any unnecessary code now?

[Daniel4078]

check the comments below, thanks

[Daniel4078]

I think it is good enough. Thank you!

[BrookeMa]

I think it is good enough. Thank you!

Thanks for your review.