Issue merging Psych-DS spec with OSF structure and functionality

[Peder Isager]

Hi Melissa/Psych-DS peeps,

The Psychological Science Accelerator is currently drafting a standard structure for our project data repositories on OSF: https://osf.io/kx6e4/?view_only=8a18e45f7f554d94bddebe959a2c2ee6

For more streamlined data management, we intend to utilize the Psych-DS specification for all our projects if possible. However, we have encountered an issue. We would like to make use of the component project feature on OSF for the "raw_data/" sub-directory, because we need to be able to change the privacy settings for this directory if sensitive data are contained within. Because we need to split the project into sub-components on OSF, we are basically splitting up the Psych-DS database and breaking the spec. This means that users will have to manually download and patch together each component before it can strictly be considered spec compliant. 

The problem seems easy to solve. All we would need is to be able to treat component projects as sub-directories of the main OSF project, and download them all as one zipped file (ignoring private project components). That way, we could freely make use of the OSF component feature, and data downloaders would still be able to extract the up-to-spec database without any patching. However, such a project+component project download is not currently possible in OSF it seems.

Does anyone know if there are currently any plans for making Psych-DS and OSF more compatible? Alternatively, do you have any advice for how we could mitigate the issue? 

ICCing Nick Fox and Ian Sullivan here as well, in case they have any insights into the issue from the OSF-side of things. 

All the best, 

Peder M. Isager

Board member, Data and Methods Committee

Psychological Science Accelerator

[Melissa Kline]

Hi Peder - sorry for my slow response! Ian and I were just able to talk about this, and the good news is that there is an immediate solution; we also hope to be able to build tighter integration in the future! (No immediate plans, but I'll be having some conversations in the next few weeks about how we can start moving forward.)

The short version is that there is a newly updated R package, osfr, that allows partial access to the OSF API. This means that it's possible to write a script that runs locally on your machine to look at a directory, make an OSF project with the same name, make components for every top-level child, and upload all files in the right place. And likewise, you could write a script that can re-download everything into a 'real' Psych-DS compliant project. 

If this sounds like it would do what you want, I can work with Nick (who is making a similar script for our SCORE projects) to learn what we need to make these adaptations!

Thanks for being in touch,

Melissa

[Melissa Kline]

Excellent, I'll be happy to collaborate on this! Anyone else on the listserv interested in joining in on building either this tool or the instructions? 

On Mon, Mar 11, 2019, 11:33 AM Peder Isager <peder...@gmail.com> wrote:

Oh man, that's awesome! We would probably need to write a step-by-step tutorial for users that are not familiar with API-style packages (if such a tutorial does not exist already), but yes, that definitely sounds like a solution to the problem.

Thanks for the quick reply (your baseline for a slow response is way too strict), and for providing an immediate solution. 

Cheers, 

Peder

[Melissa Kline]

That would be great! 

On Mon, Mar 11, 2019, 11:58 AM Seehuus, Martin O. <msee...@middlebury.edu> wrote:

I’m happy to help with either; my skills are probably more in line with the writing instructions/tutorial than in the coding itself, but I can at least troubleshoot.

 

Martin

 

From: psych-data...@googlegroups.com <psych-data...@googlegroups.com> On Behalf Of Melissa Kline
Sent: Monday, March 11, 2019 11:57 AM
To: Peder Isager <peder...@gmail.com>
Cc: Patrick S Forscher <fors...@uark.edu>; Nicholas Fox <nw...@psych.rutgers.edu>; Ian Sullivan <i...@cos.io>; Psych-Data-Standards <psych-data...@googlegroups.com>
Subject: Re: Issue merging Psych-DS spec with OSF structure and functionality

 

Notice: This email is from an external sender. Please use caution before clicking links or opening attachments.

 

--
You received this message because you are subscribed to the Google Groups "Psych-Data-Standards" group.
To unsubscribe from this group and stop receiving emails from it, send an email to psych-data-stand...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/psych-data-standards/CAF%3DPoJOyGDLykrTr4oJYvZ5VHPsk5__Scjh4cMoGKLNc7LRBsA%40mail.gmail.com.
For more options, visit https://groups.google.com/d/optout.

[Seehuus, Martin O.]

I’m happy to help with either; my skills are probably more in line with the writing instructions/tutorial than in the coding itself, but I can at least troubleshoot.

 

Martin

 

From: psych-data...@googlegroups.com <psych-data...@googlegroups.com> On Behalf Of Melissa Kline
Sent: Monday, March 11, 2019 11:57 AM
To: Peder Isager <peder...@gmail.com>
Cc: Patrick S Forscher <fors...@uark.edu>; Nicholas Fox <nw...@psych.rutgers.edu>; Ian Sullivan <i...@cos.io>; Psych-Data-Standards <psych-data...@googlegroups.com>
Subject: Re: Issue merging Psych-DS spec with OSF structure and functionality

 

Notice: This email is from an external sender. Please use caution before clicking links or opening attachments.

 

Excellent, I'll be happy to collaborate on this! Anyone else on the listserv interested in joining in on building either this tool or the instructions? 

--

[Ruben Arslan]

It might make sense to upload our shared our implementation examples in the PsychDS repo to OSF to show how this works right? If we go all the way, we'll notice what steps in the process are difficult. Do you have code to share, Melissa? I could give it to my RA who worked through the example.
Had we planned to review each others implementation examples?

[Peder Isager]

Oh man, that's awesome! We would probably need to write a step-by-step tutorial for users that are not familiar with API-style packages (if such a tutorial does not exist already), but yes, that definitely sounds like a solution to the problem.

Thanks for the quick reply (your baseline for a slow response is way too strict), and for providing an immediate solution. 

Cheers, 

Peder

Den man. 11. mar. 2019 kl. 15:34 skrev Melissa Kline <melissa...@gmail.com>:

[Melissa Kline]

No code yet, but I'll share what I've been playing with in a day or 2 - OSFR is a pretty new package so I"m not positive if it'll support everything we want to do yet (it can all be done via direct API calls, but osfr makes things MUCH more accessible to people like...me, so), but I have a question out to the developers to ask about a few specifics before I get started. 

Fortunately Nick has already been writing a closely related script for our giant COS project, so anything we learn for either of these avenues will help both :)

- Melissa

On Mon, Mar 11, 2019 at 12:19 PM Ruben Arslan <ruben...@gmail.com> wrote:

It might make sense to upload our shared our implementation examples in the PsychDS repo to OSF to show how this works right? If we go all the way, we'll notice what steps in the process are difficult. Do you have code to share, Melissa? I could give it to my RA who worked through the example.
Had we planned to review each others implementation examples?

--

You received this message because you are subscribed to the Google Groups "Psych-Data-Standards" group.
To unsubscribe from this group and stop receiving emails from it, send an email to psych-data-stand...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/psych-data-standards/CALv3xztmy9%3Dtzu7oMWQEP0czUGy8%3D%3DHvoh6eUvtK%3DnUqDard4Q%40mail.gmail.com.