Automatic author list generation

[drphilmarshall]

This PR will address #25. To get started, I made a contributions.csv file for cookiecutter to fill in. I couldn't remember the columns that mkauthlist wants, and came up one short - maybe you can correct this, @kadrlica ? I think the contributions should be formatted such that the combination $INITIALS $CONTRIBUTION makes a complete sentence (as in, "PJM started the paper."). We may need mkauthlist --lsstdesc to lowercase the first character of the contribution statement, and extract the initials of the author from their name. Thoughts?

[kadrlica]

I've added an example author list file. I'm not sure which information you want to fill with the cookiecutter. Some of the columns are currently ignored by mkauthlist.

[kadrlica]

Ok, I've update the version of mkauthlist. How do we want to do the pip installation? Can cookiecutter do it?

[drphilmarshall]

I don't think so: it needs to go in the Makefile somehow. Ideally makeauthlist would pip install the first time you tried to make authors and then never again.

On Thu, Nov 17, 2016 at 11:33 AM, Alex Drlica-Wagner < notifications@github.com> wrote:

Ok, I've update the version of mkauthlist. How do we want to do the pip installation? Can cookiecutter do it?

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub https://github.com/DarkEnergyScienceCollaboration/start_paper/pull/37#issuecomment-261345585, or mute the thread https://github.com/notifications/unsubscribe-auth/AArY93C9h7NzZfL_UolrmI1rO3Ac_I-Oks5q_KwNgaJpZM4K1iwg .

[drphilmarshall]

Thanks @kadrlica ! I renamed your authors.csv file to be example_contributions.csv . I'm actually wondering if we might just add these 4 lines to the contributions.csv file, as an example - do you think people would figure out to delete us? I like the simplicity of only having one file of each kind...

BTW I think the file that we input to mkauthlist does need to be called contributions.csv, to emphasize that this is where the names and credits go.

[drphilmarshall]

Ah, just seen your comment in https://github.com/DarkEnergyScienceCollaboration/start_paper/issues/25#issuecomment-261301428 To be clearer, I think it's important that the author information is stored in contributions.csv because we want people paying attention to the contributions - because this is new, unfamiliar, required by the policy, and at risk of not being done well...

[kadrlica]

For the time being (at least until an automated publication database is in place) I think it would be better to leave a fleshed out example in contributions.csv. Actually, I think that mkauthlist will respect the # comment character, so what we could do is fill the first author from the cookiecutter and comment out the other lines as examples. The other option would be to leave the lines uncommented, but to change the content to junk like Lastname1,Firstname1,F.~Lastname1,....

I still prefer authors.csv, but I understand your point. When we talked about this in PA, the idea was that the mkauthlist would take this file (whatever we call it) and spit out authors.tex and contributions.tex. The problems that I see with calling it contributions.csv is:

• It is primarily a list of people not a list of contributions
• The existing machinery (i.e., mkauthlist) will operate just fine if there are no contributions in this file
• This is the file we are passing to a tool called mkauthlist after all...

Maybe we can brainstorm other alternatives? Something like contributors.csv, authlist.csv, people.csv, usual-suspects.csv, blame-list.csv.

A major pending issue is to revamp mkauthlist to actually deal with the new formatting and additional feature that we are requesting. Let the mkauthlist issues begin...

[drphilmarshall]

I like the idea of the commented out example rows, below the cookiecutter-input one. OK, I am persuaded by your arguments about the filename. I'll fix things up!

[kadrlica]

ok, I went maybe 5 steps forward and 1 step back. The author.csv now has the cookiecutter author fields commented out and the example authors uncommented. This was mostly for testing and I would not object to having things changed back.

I also need to commit my changes to mkauthlist...

[kadrlica]

Ok @drphilmarshall , it seems like it's probably to code review this monster...

[drphilmarshall]

I will be delighted to do this in the New Year! Thanks very much, Alex - happy holidays!

[drphilmarshall]

Hi Alex,

Sorry to miss you at the DESC meeting this week! Seth tells me that he is going to point people at start_paper in his "Pub Board Update" plenary on Thursday, so I'm wondering if we should try and merge this PR by then? The main motivation is to get people working on their contributions...

For this "launch" to go well, I think we need to be sure of the following things:

• People with old start_paper folders need to be able to make update and get the new code, including mkauthlist, without breaking anything (like their author lists)

• The installation of mkauthlist needs to work cleanly, or else Make needs to run without mkauthlist

What do you think? Are there other behaviors we need to see before launching the new authors.csv feature? Shoudl we wait until after this week (when we will have more users)?

Hope things are going well!

Phil

[kadrlica]

Hi Phil, I'm traveling and have limit ability to develop this week, but I thought that these requirements were satisfied by this PR. Would you be able to test this and tell me if any of these break? Targeted fixes should be possible.

[drphilmarshall]

Hey @kadrlica I got mkauthlist working! The meta-Note in the start_paper repo now uses it to make a nice author list based on its authors.csv file, and everything works. Travis is failing though, I think because the tex outputs from mkauthlist are not wrapped in % {% raw %} and % {% endraw %} statements, perhaps? I'll see if I can fix this by avoiding underscores in the cookiecutter variables.

[drphilmarshall]

OK cool - I changed the names of the cookiecutter variables and everything works now. I'm going to merge this, and try upgrading the CCL note with it. If I hit problems I'll issue them afresh. Thanks, Alex!