Closed yarikoptic closed 3 years ago
FTR: Cloning this exact dataset works for me. So I must be having something that you don't have...
(osf) ╭─adina@muninn /tmp
╰─➤ datalad clone osf://q8xnk
install(ok): /tmp/q8xnk (dataset)
It also worked after upgrading to 0.13.2.dev9. I'm also attaching what else is in my venv:
although description says it that it is just "an example", IMHO would be better to tune it to use this particular dataset id (n6bgd) not some other q8xnk, so fast/silly readers like me do not just skip the preamble of the lengthy description and would start assuming that may be the ID is for the same dataset but some different version or smth
created #126 in response
Could you maybe run this with --dbg
and see if the http response has any more meaningful information? When I did #126, it contained info on required attributes I initially didn't specify, for example...
FWIW I'm getting the same error on our Windows machine right now, and I just can't get behind the cause of it.
... what working with the windows machine feels like (live from my desk)
EDIT: I now got past the "Bad request" status (and promptly remembered that a datalad clone from this repo isn't possible on Windows ;-) ). It was a malformed OSF token. Not sure whether this is true in your case, too...
The only way I can reproduce this if the credentials are not correct. If you still have the same OSF credential configuration, could you just just datalad osf-credentials
and check that authentication works?
Sorry for me forgetting about this one -- confirming it was credentials. Added new an it worked.
do you think it would be feasible somehow to percolate that information to the user, since otherwise it is just a 400 (not 403) which only is seen.
Mhh, its been a while since I had worked on this, but if I remember correctly, I never saw any authentication errors in the http responses that we could propagate. I assume we could check for credential validity before a clone, but that seems to be a costly operation
feel welcome to close until issue reemerges reported by a user (might never come, I might be special ;-))
You're definitely special :P
I'll close it for now.
Following the docs example at http://docs.datalad.org/projects/osf/en/latest/tutorial/cloneosfdata.html I am trying to clone an existing demo dataset, but unfortunately that does not work for not exactly a clear reason:
What could be the culprit? Direct
git clone
invocation also fails similarly :Also, going to that osf dataset https://osf.io/n6bgd/ shows in description: "this component can be git or datalad cloned from a 'osf://ID' URL, where 'ID' is the OSF node ID that shown in the OSF HTTP URL, e.g. https://osf.io/q8xnk/ can be cloned from osf://q8xnk"
n6bgd
) not some otherq8xnk
, so fast/silly readers like me do not just skip the preamble of the lengthy description and would start assuming that may be the ID is for the same dataset but some different version or smth/
or not does not matter but better to be consistent and either have it in both documentation and description (example) or not (i.e.osf://q8xnk
vsosf://q8xnk/
)DataLad 0.13.2.dev309 WTF (configuration, datalad, dependencies, environment, extensions, git-annex, location, metadata_extractors, python, system)
# WTF ## configuration