Unpin Numpy

[sf1919]

Description of work.
A few months ago we pinned numpy until we could resolve any issues with newer versions. This is the PR that fixes any issues and removes the pin.

To test:

1. Check tests pass
2. Code Review

@ajjackson - please can you take a look at the fix for the Abins tests?

There is no associated issue.

This does not require release notes because the changes should not affect users

---

Reviewer

Please comment on the following (full description):

Code Review

• Is the code of an acceptable quality?
• Does the code conform to the coding standards?
• Are the unit tests small and test the class in isolation?
• If there is GUI work does it follow the GUI standards?
• If there are changes in the release notes then do they describe the changes appropriately?
• Are the release notes saved in a separate file, using Issue or PR number for file name and in the correct location?

Functional Tests

• Do changes function as described? Add comments below that describe the tests performed?
• Do the changes handle unexpected situations, e.g. bad input?
• Has the relevant (user and developer) documentation been added/updated?

Does everything look good? Mark the review as Approve. A member of @mantidproject/gatekeepers will take care of it.

[ajjackson]

As discussed I think the nicest way to fix the Abins tests is to move the array creation down into the
with self.assertRaises(ValueError): blocks. Then we are still checking that ragged arrays will be stopped, and don't really care whether it is Numpy or Abins that stops them.

I would suggest that we make a note to remove this check once we know Mantid is not being run on any older Numpy, but there is already a roadmap to remove these classes so it should happen naturally.

[martyngigg]

Could I make a suggestion that we match conda-forge here unless there is anything that we need from a much more up to date verison? numpy is core to so many projects and we use a large set of libraries from conda-forge so having our minimum requires be the same gives us maximum compatibility.

Conda Forge maintain a pinning repository where they define these versions.

For numpy they define a version per Python version which is accomplished by matching the list entries for Python and numpy, e.g 3.8 uses 1.20.

[sf1919]

I've now made the changes requested by @martyngigg and @ajjackson .

[robertapplin]

Looks good to me 👍