Update build system to flit_core

[takluyver]

Hi, I'm opening this pull request as part of a push to modernise how packages use Flit as a Python build backend.

Using flit_core as the backend in place of flit is recommended in the docs, and will make it faster for tools like pip & build to build your package from source, as it has fewer dependencies to install.

Specifying an explicit version range (>=2,<4) helps to ensure that your package can still be readily built from source despite changes in future major versions of Flit, because it will still use version 3.x. For instance, a future version is likely to drop support for the [tool.flit.metadata] table, in favour of the now-standardised [project] table for metadata. This is also in the docs, along with details of which versions support which features.

Most users probably install your package from a pre-built 'wheel' on PyPI, so this changes won't affect them at all. But people who install from a git checkout, for instance, will benefit.

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 77.73%. Comparing base (d4f59a0) to head (8e3c20a).

❗ There is a different number of reports uploaded between BASE (d4f59a0) and HEAD (8e3c20a). Click for more details.

HEAD has 7 uploads less than BASE

Flag	BASE (d4f59a0)	HEAD (8e3c20a)
	12	5

Additional details and impacted files

@@             Coverage Diff             @@
##             main     #114       +/-   ##
===========================================
- Coverage   89.76%   77.73%   -12.04%     
===========================================
  Files         140      140               
  Lines        8102     8102               
  Branches     1499     1499               
===========================================
- Hits         7273     6298      -975     
- Misses        596     1660     +1064     
+ Partials      233      144       -89     

☔ View full report in Codecov by Harness.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.