zenodo: more author metadata, and ccp-petmr community

[casperdcl]

• add Zenodo metadata for automated release (https://doi.org/10.5281/zenodo.2707911)
  • list of authors, ORCIDs, affiliations
  • communities: CCP-PETMR
  • licence: Apache-2.0 Other (for now, until GPL from FFTW is removed)
  • related DOI: 10.1109/NSSMIC.2017.8532815

[casperdcl]

@KrisThielemans regarding foobar2016 previously appearing in Zenodo:

1. it's no longer an issue if we're specifying authors in a metadata file
2. apparently it's likely because a contributor used to have that username
3. author lists are generated based on https://github.com/CCPPETMR/SIRF/graphs/contributors

[codecov]

Codecov Report

Merging #391 into master will not change coverage.
The diff coverage is n/a.

@@           Coverage Diff           @@
##           master     #391   +/-   ##
=======================================
  Coverage   50.05%   50.05%           
=======================================
  Files           2        2           
  Lines        1720     1720           
=======================================
  Hits          861      861           
  Misses        859      859

---

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 ea4a56f...0314fff. Read the comment docs.

[codecov]

Codecov Report

Merging #391 into master will decrease coverage by 0.47%.
The diff coverage is n/a.

@@            Coverage Diff             @@
##           master     #391      +/-   ##
==========================================
- Coverage   50.64%   50.17%   -0.48%     
==========================================
  Files           2        2              
  Lines        1629     1720      +91     
==========================================
+ Hits          825      863      +38     
- Misses        804      857      +53

Impacted Files	Coverage Δ
src/xGadgetron/pGadgetron/Gadgetron.py	47.13% <0%> (-0.89%)	⬇️
src/xSTIR/pSTIR/STIR.py	52.83% <0%> (-0.09%)	⬇️

---

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 8ff0bea...f124e74. Read the comment docs.

[ashgillman]

Unrelated, but how did Johannes get an 'X' in his ORCID? I checked and it links correctly.

[casperdcl]

lol I noticed that

[casperdcl]

response from Zenodo support:

You can see the code that does the extraction here: https://github.com/inveniosoftware/invenio-github/blob/master/invenio_github/utils.py#L80-L108 

I've tried to debug it, and I've traced the foobar2016 to this GitHub API response:

https://api.github.com/repos/CCPPETMR/SIRF/stats/contributors
(https://developer.github.com/v3/repos/#list-contributors)
forbar2016 is listed there as having 3 contributions to the project

How the user ended up there I didn't have time to look into yet.

[KrisThielemans]

Currently, our license still has to be GPL 3 due to use of FFTW. can we say something like "GPL 3, but most parts also licensed as Apache 2.0"?

[KrisThielemans]

@casperdcl I actually cannot see foobar2016 on our list of contributors. (Indeed, that user has no contributions to any repo anywhere). mystery.

[casperdcl]

can we say something like "GPL 3, but most parts also licensed as Apache 2.0"?

no but we can leave it as the default (Other - Open)

I actually cannot see foobar2016 on our list of contributors

good point, neither can I when I access the API. Sent a follow-up to Zenodo. They must have some weird caching issues.

[ashgillman]

NB: You might have to accept #392 before this PR

[rijobro]

NB: You might have to accept #392 before this PR

Good spot, done.

[casperdcl]

@paskino you may want to check this file to find out author details

[casperdcl]

Follow-up from Zenodo:

I apologise, I sent the wrong [...] URL. The URL is https://api.github.com/repos/CCPPETMR/SIRF/contributors (without stats) [...] which includes foobar2016.

[...] you can see the commits here:
https://api.github.com/repos/CCPPETMR/SIRF/commits?author=foobar2016
Specifically [...]:
ddc363c
18b7e18
819d8a1
[...] authored by:
Evgueni Ovtchinnikov <me@example.com>

Why GitHub is associating the GitHub user foobar2016 with Evgueni Ovtchinnikov <me@example.com>, I don't know. Perhaps because of the email address "me@example.com" being used?

[ashgillman]

Some quick edits on a cluster or another computer perhaps?

I've had similar experiences when making edits on the cluster and no email was set, GitHub doesn't associate the commits with my account.

You can modify these with
git config --global user.email "me@example.com" and also user.name.

[KrisThielemans]

my first thought was that github is a bit silly but I then realised that the commiter email is probably the only way they can do this. After all, you can have lots of local commits made by lots of different people, which then one github user pushes. Those commits should not be assigned to the github user of course. Therefore, github has to look at git content, which stores user.email and user.name for each commit. Makes sense.

We cannot fix this as it'd mean invalidating all commits since then (replacing them with new ones), so we'll live with it.

Moral of the store: be careful how you do your git config!

@casperdcl please thank your zenodo contact. Very helpful!