Add test script to test pyspedas releases as installed from pypi

[jameswilburlewis]

Our current github actions test scripts install pyspedas via 'git clone' rather than 'pip install'. This won't catch packaging problems, as recently happened when some directories got moved without updating MANIFEST.in to ensure that necessary non-Python files were still included in the release. So we should do some testing on the 'pip install'-able packages.

Of course, we would want to catch such problems before releasing to pypi. So we may want to alter our release workflow, for example by releasing a beta package to pypi first (that 'pip install' won't see by default), testing it, then promoting to a 'production' release after the beta passes its tests. Or, I believe there's some sort of 'test' server for pypi. So we can deploy a release to the test pypi server, test it, then upload the identical files to production-pypi after it passes.

And this might in turn lead to us creating release-specific branches and tags, rather than building and releasing directly from the master branch.