overhaul infrastructure

[knaaptime]

No description provided.

[martinfleis]

What is the selection logic here?

[knaaptime]

no idea. its who was in base. hoping reviewers have opinions

[martinfleis]

In geopandas, in "authors" is only Kelsey Jordahl who founded the project, no one else. Maybe that would be the solution here as well? So only @sjsrey? Just a suggestion.

[knaaptime]

most of the other packages have two, so i think its probably appropriate to have it be serge and levi. I'll change it to that unless they think otherwise

[martinfleis]

Yeah, I don't know the history here.

[knaaptime]

well i get the logic, but maybe its a question that should be revisited occasionally instead of solved once. Serge, Ran and I were once the authors of spopt, but now it makes more sense for the authors to be James and Xin. ...but also a lot of the code was ported over from region and originally written by juanca, serge, and aleks. So its tricky

[knaaptime]

i'd probably characterize it as brute force rather than elegance, but sometimes any solution is preferable to nothing :P. By using founders as a replacement for authors, we're just saying our answer is to not address the question. That's fine, but it's a particular position (that assigning authorship is too hard to handle).

to throw out another few analogies, Guido is the author of Python, but at this point im sure he's the author of almost nothing in stdlib. Serge and Luc will forever be the authors of PySAL, but i think of authorship among the subpackages as additive, so the 'authors' are the respective contribution trees

i also think of pysal a bit like an edited handbook on spatial analytical methods. Serge and Luc are the editors of the handbook, and they wrote several chapters of the first edition. Many editions are pretty similar, where the same authors update their chapters in light of new evidence and techniques. But the content of the collection changes a bit over time as the discipline evolves and some of the authors rotate out, so chapters are eventually written by others. The subpackages in pysal are like the chapters, and lib is maybe the introduction to the series.

i think theres also an argument that the four of us sat down, scoped out, wrote, and revised this draft of the package (the graph-based version anyway), which makes us the authors of this version of the chapter. Eventually somebody else will probably write a new draft, at which point they assume 'authorship' (but we will always be in the contributor tree (and picked up by the zotero citation, for whatever thats worth)

[jGaboardi]

I think that sounds reasonable. I think it's your call at this point @knaaptime.

[knaaptime]

i'm the resident dev with a community planning degree, so its only my job to raise questions. Answers are for the community :P. But jokes aside i'd probably just leave it at serge and levi

[martinfleis]

I'm fine with any solution. Just didn't want to exclude someone if we decide to name authors based on contributions. FYI, pandas has just Pandas development team as authors.

[martinfleis]

We need to add pip install . to the GHA to ensure we have access to __version__. That will fix the CI failure.

[codecov]

Codecov Report

Merging #579 (e885c09) into main (3576691) will increase coverage by 1.5%.
The diff coverage is n/a.

@@           Coverage Diff           @@
##            main    #579     +/-   ##
=======================================
+ Coverage   82.4%   83.9%   +1.5%     
=======================================
  Files        129     128      -1     
  Lines      15358   15021    -337     
=======================================
- Hits       12655   12610     -45     
+ Misses      2703    2411    -292     

see 1 file with indirect coverage changes

[martinfleis]

One minor nit. Otherwise looks good!

[martinfleis]

I think that the latest is v4 now (am on phone now)

[jGaboardi]

Are we good to merge for now then, or are we waiting for something else?

[martinfleis]

Two approvals are in and authorship can be discussed in the dev call and changed later if needed. Lgtm