search

mne-tools / mne-bids-pipeline

Automatically process entire electrophysiological datasets using MNE-Python.
https://mne.tools/mne-bids-pipeline/
BSD 3-Clause "New" or "Revised" License
140 stars 66 forks source link

STY: Name of package #265

Closed larsoner closed 3 years ago

larsoner commented 3 years ago

I think we should rename mne-study-template to mne-bids-pipeline.

The "study template" name gives the wrong impression. When describing it to people I have to tell them it's a misnomer. I think from the implementation end we can make a redirect for mne.tools/mne-study-template and GitHub will automatically redirect when a repo gets renamed, so it's mostly a question of how many existing users we will annoy versus the benefits we will get from (I think much) better visibility from a more appropriate name. It seems worth it to me.

If people agree that renaming is worthwhile, does mne-bids-pipeline seem like a good choice? It seems most appropriate given the current tagline for the name is:

The MNE Study Template is a full-flegded processing pipeline for your MEG and EEG data.

It operates on raw data stored according to the Brain Imaging Data Structure (BIDS). Processing is controlled using a simple human-readable configuration file.

So mne-bids-pipeline to me captures the two essential components: being a pipeline, and working with BIDS data.

From offline discussion with @agramfort and @hoechenberger they're on board with a name change already. So people feel free to comment on mne-bids-pipeline specifically and/or give alternatives, or speak up if you're opposed to this move overall!

jasmainak commented 3 years ago

I'm not much involved here but I like it :)

SophieHerbst commented 3 years ago

I don't fully oversee which complications it might cause, but agree with @larsoner and I like mne-bids-pipeline!

drammock commented 3 years ago

+1 from me

agramfort commented 3 years ago

such a consensus on an MNE issue ❤

larsoner commented 3 years ago

I'll probably carry out the rename on Friday unless anyone objects in the meantime!

hoechenberger commented 3 years ago

MNE BIDS Pipeline is good. Should we move the repo to mne-tools/bids-pipeline?

Otherwise it might appear as an "MNE-BIDS Pipeline" (mind the hyphen), but that's sorta misleading...

agramfort commented 3 years ago

Keep mne

larsoner commented 3 years ago

I would call the repo mne-bids-pipeline, package mne_bids_pipeline, and write about it in docs as MNE-BIDS-Pipeline or MNE BIDS Pipeline. My preference is to have all hyphenated because it's most consistent with MNE-Python

hoechenberger commented 3 years ago

Ok!

hoechenberger commented 3 years ago

Clarification:

I'm ok with your proposal:

I would call the repo mne-bids-pipeline, package mne_bids_pipeline, and write about it in docs as MNE-BIDS-Pipeline

I'm only slightly concerned that "MNE-BIDS-Pipeline" is rather longish … can we find a more concise or jazzy term than "pipeline"?

hoechenberger commented 3 years ago

How about … 

MNE-BIDS-Plumber
MNE-Plumber

because, you know … pipeline … pipes … plumber?

xx

Voilà !

Logo & copyright infringement lawsuit right there!

britta-wstnr commented 3 years ago

MNE-Pypes :smile:

larsoner commented 3 years ago

I still prefer MNE-BIDS-Pipeline but could live with shortening it to MNE-Pipeline or so. I enjoy the more creative/clever names that avoid the word "pipeline" but so far don't prefer them because they hurt discoverability / make people have to think harder or read more about what the package actually does to know.

dengemann commented 3 years ago

I think pipeline is a misnomer, no? Isn't this supposed to be a template that is copied in each study and edited or extended according to the particular needs of the study? A pipeline would something that you use like a library. MNE-BIDS-template, MNE-pipline-maker, MNE-BIDS-maker, MNE-BIDS-assistant ...

drammock commented 3 years ago

I think pipeline is a misnomer, no?

I think it makes sense to call it a pipeline because it manages and performs several sequential preprocessing steps.

hoechenberger commented 3 years ago

Isn't this supposed to be a template that is copied in each study and edited or extended according to the particular needs of the study?

In principle yes, but over the course of the past year we've moved more and more in the direction of a package-like thing as complexity grew.

I'll give my okay for MNE-BIDS-Pipeline :)

agramfort commented 3 years ago

In principle yes, but over the course of the past year we've moved more and more in the direction of a package-like thing as complexity grew.

I'll give my okay for MNE-BIDS-Pipeline :)

+1

I agree the code is now written and thought so that people will not just clone, modify locally and keep changes locally. We are moving towards a community consolidated pipeline

hoechenberger commented 3 years ago

@larsoner while you're breaking things anyway 😜 feel free to also rename the master branch to main if you want! We wanted to do that anyway

larsoner commented 3 years ago

I just had to fetch 500 MB to update my repo, which is not great. I think the gh-pages branch should be moved to another dedicated repo and deleted from here. @hoechenberger do you want to take care of that part? Maybe a new mne-bids-pipeline-gh-pages or something?

agramfort commented 3 years ago

I would stick to gh-pages but in this repo but I would squash commits regularly. It's the simplest thing to do

hoechenberger commented 3 years ago

How come we're not running into this issue with MNE..?

larsoner commented 3 years ago

I would stick to gh-pages but in this repo but I would squash commits

Should I start just by squashing all to a single commit now?

How come we're not running into this issue with MNE..?

We use a separate repo for MNE-Python

https://github.com/mne-tools/mne-tools.github.io

agramfort commented 3 years ago

Should I start just by squashing all to a single commit now?

+1

How come we're not running into this issue with MNE..?

We use a separate repo for MNE-Python

https://github.com/mne-tools/mne-tools.github.io

yes but it's the URL of the org and it has a domain name.

if you don't use gh-pages maybe it could become

mne-bids-pipeline.mne.tools

but I prefer that we stick to:

mne.tools/mne-bids-pipeline

sappelhoff commented 3 years ago

I would stick to gh-pages but in this repo but I would squash commits

Should I start just by squashing all to a single commit now?

We are using a gh-pages branch in mne-bids too (235 commits by now) --> https://github.com/mne-tools/mne-bids/tree/gh-pages

so should we "squash" that too?

I don't really get though why that would reduce the download size. Is the previous history of commits lost during squashing?

Is this something like git rebase -i, using squash for all but one of the commits? Would you mind posting your commands @larsoner, please ? :-)

agramfort commented 3 years ago

it's to avoid having to download all the web site versions when you git clone the origin repo

larsoner commented 3 years ago

I don't really get though why that would reduce the download size. Is the previous history of commits lost during squashing?

Yes squashing (and then push-force, which is what I did with this repo) loses the commit history. In MNE-Python mne-tools.github.io nowadays we do one huge squash at each release to keep repo size down while still maintaining some amount of history.

Is this something like git rebase -i, using squash for all but one of the commits? Would you mind posting your commands @larsoner, please ? :-)

It was something like:

$ git fetch upstream
$ git checkout -b gh-pages upstream/gh-pages
$ git shortlog | wc -l
181
$ git rebase -i HEAD~181
... # do :%s/pick/squash/g in vim, then change the topmost one back to pick
sappelhoff commented 3 years ago

thanks a lot, makes sense!