Bring the extension up to date with datalad-extension-template

[mslw]

This WIP PR attempts to bring the extension in line with the datalad-extension-template; fix #99. In summary, the biggest changes affect:

• build / packaging process (addition of versioneer, use of setup.cfg, modernised build support functions)
• metadata and config files (eg. addition of zenodo.json, tweaks to appveyor CI config, addition of github actions)

What was done and how

• Set up versioneer from scratch, using its latest version
  • Also made a minor change in versioneer to prefer setuptools over distutils (reflecting changes in datalad-extension-template and an unmerged change in versioneer itself)
  • Was easier to do it this way (rather than bring it in from the template), also ended up with a newer version
• Made the remaining alignments by merging datalad-extension-template onto this one (git merge --allow-unrelated-histories --squash) and using best judgement to resolve conflicts.
  • The most involving change was to move contents from setup.py to setup.cfg, and act on related files
  • This also brought in more meta files such as info for zenodo and config for code climate
• Added contents to zenodo.json (all contributors from GitHub page, in alphabetic order, with affiliations and orcids taken from other datalad repos)
• Made a small fix to the way the template builds manpages and examples
• Tests: nosetests passes locally (3 skipped)

I also tried to keep the commit messages descriptive, so that may be a good way to learn what was changed.

What I was able to verify

• Packaging / versioning: from what I can see, python setup.py sdist bdist_wheel produces very similar content to what it did, except now it is tagged in a versioneer way, ie. something like datalad_neuroimaging-0.3.1+24.gc24e4fc for a build from source.
• Docs: build successfuly (locally and in github actions)
• Tests: pass

Attention for the reviewers

• In doing the moves above, the history of the files added from the template was lost and now I'm shown as the only contributor. Versioneer (autogenerated) was also added in my name. Is that ok?
• Please check the contents of zenodo.json, and see if that's ok (all contributors from GitHub page, in alphabetic order, with affiliations and orcids taken from other datalad repos)
• Would be awesome if you could confirm actual functioning, I haven't gone beyond checking the import.

[codecov[bot]]

Codecov Report

Merging #101 (69c8a68) into master (d04807c) will increase coverage by 85.12%. The diff coverage is 100.00%.

@@             Coverage Diff             @@
##           master     #101       +/-   ##
===========================================
+ Coverage        0   85.12%   +85.12%     
===========================================
  Files           0       21       +21     
  Lines           0     1049     +1049     
===========================================
+ Hits            0      893      +893     
- Misses          0      156      +156     

Impacted Files	Coverage Δ
datalad_neuroimaging/__init__.py	100.00% <100.00%> (ø)
datalad_neuroimaging/_version.py	100.00% <100.00%> (ø)
datalad_neuroimaging/tests/test_procedure.py	100.00% <0.00%> (ø)
datalad_neuroimaging/tests/test_register.py	100.00% <0.00%> (ø)
datalad_neuroimaging/tests/test_dicomconv.py	96.42% <0.00%> (ø)
...atalad_neuroimaging/extractors/tests/test_dicom.py	94.54% <0.00%> (ø)
datalad_neuroimaging/extractors/tests/test_nidm.py	100.00% <0.00%> (ø)
datalad_neuroimaging/bids2scidata.py	73.35% <0.00%> (ø)
datalad_neuroimaging/tests/test_bids2scidata.py	96.77% <0.00%> (ø)
... and 12 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more Δ = absolute <relative> (impact), ø = not affected, ? = missing data Powered by Codecov. Last update d04807c...69c8a68. Read the comment docs.

[mslw]

I've updated my PR description, and I think this is ready - for review at least.

As a side note, I had to restart the linux build on appveyor because it failed after being stuck for like an hour on "Build started", but on the second run it completed without any issues (3,5 minutes) so it must have been some glitch in appveyor.