sorry state of Zenodo integration

132 views
Skip to first unread message

Dima Pasechnik

unread,
Jun 14, 2023, 1:49:14 PM6/14/23
to sage-devel, sagemath-admins
The backup repo
https://github.com/sagemath/sage-archive-2023-02-01
apparently threw Zenodo integration off - now
our main repo, sage, is not associated with Zenodo any more, instead
that backup repo is the one associated with Zenodo.
Probably one had to get rid of .zenodo.* f and CITATION.cff files there
first, I really don't know.

Anyhow, I'm exchanging emails with Zenodo support, trying to sort this out.

I also don't understand why sage-archive-2023-02-01 repo has releases.
Was this repo created by some kind of a GitHub-specific clonig process?

Dima

Matthias Koeppe

unread,
Jun 14, 2023, 1:52:10 PM6/14/23
to sage-devel
On Wednesday, June 14, 2023 at 10:49:14 AM UTC-7 Dima Pasechnik wrote:
I also don't understand why sage-archive-2023-02-01 repo has releases.
Was this repo created by some kind of a GitHub-specific clonig process?

As it says in its "About" box: "This repository used to be the user-facing mirror of the Sage source tree. As Sage development migrated on 2023-02-01 from https://trac.sagemath.org/ to our new repository on GitHub, https://github.com/sagemath/sage, we have renamed and archived this repository."



 

Dima Pasechnik

unread,
Jun 14, 2023, 1:59:57 PM6/14/23
to sage-devel
Was it a github repo? Some other repo? Did the latter actually have any github releases?

AFAIK, the user-facing mirror was sagetrac-mirror, still there, no?





 

--
You received this message because you are subscribed to the Google Groups "sage-devel" group.
To unsubscribe from this group and stop receiving emails from it, send an email to sage-devel+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/sage-devel/2411ed3f-7687-4816-84f8-ae097fd4021cn%40googlegroups.com.

Matthias Koeppe

unread,
Jun 14, 2023, 2:11:05 PM6/14/23
to sage-devel
We did have sagemath/sage with branches develop and master. 
And you made releases there.

Dima Pasechnik

unread,
Jun 14, 2023, 8:08:11 PM6/14/23
to sage-...@googlegroups.com
On Wed, Jun 14, 2023 at 7:11 PM Matthias Koeppe
<matthia...@gmail.com> wrote:
>
> We did have sagemath/sage with branches develop and master.
> And you made releases there.

well, it's to understand how exactly Zenodo integration broke.
Apparently for them renaming of a repo is not a reason to turn the
integration off,
so it wasn't a good idea to rename/duplicate it.
I still don't understand why https://github.com/sagemath/sage has all
the releases
as in https://github.com/sagemath/sage-archive-2023-02-01

One problem is that the integration wasn't set up by me in the 1st
place, I can't unlink
https://github.com/sagemath/sage-archive-2023-02-01 from Zenodo. Our
1st Zenodo DOI is for Sage 6.6,
https://github.com/sagemath/sage-archive-2023-02-01/releases/tag/6.6
release tagged by Harald.

Harald, was it you who set up Zenodo DOIs for Sage?

Dima
> To view this discussion on the web visit https://groups.google.com/d/msgid/sage-devel/637c1733-ea8d-4d41-a329-fba061c75b2an%40googlegroups.com.

Matthias Koeppe

unread,
Jun 14, 2023, 8:15:15 PM6/14/23
to sage-devel
On Wednesday, June 14, 2023 at 5:08:11 PM UTC-7 Dima Pasechnik wrote:
On Wed, Jun 14, 2023 at 7:11 PM Matthias Koeppe
<matthia...@gmail.com> wrote:
>
> We did have sagemath/sage with branches develop and master.
> And you made releases there.
 
so it wasn't a good idea to rename/duplicate it.

Well, it was the only way to import the migration archive, per GitHub support team.
 
I still don't understand why https://github.com/sagemath/sage has all
the releases
as in https://github.com/sagemath/sage-archive-2023-02-01

The new repository was created from the migration archive, and part of its data was the release information.

Harald Schilly

unread,
Jun 15, 2023, 4:09:51 AM6/15/23
to sagemat...@googlegroups.com, sage-devel
Hi, I was indeed the one who set this up, but I vaguely remember someone else added files to the sources back then.

In any case, I accessed zenodo and removed the link with zenodo from the archived version, and added the sagemath/sage repository.

Let's see what happens, i.e. it says it is "syncing" now. ... OK, nothing happened. Seems like only new tagged releases will trigger a registration. 

I also found this, kind of a work around to trigger zenodo:

-- Harald




--

---
You received this message because you are subscribed to the Google Groups "sagemath-admins" group.
To unsubscribe from this group and stop receiving emails from it, send an email to sagemath-admi...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/sagemath-admins/CAAWYfq2KTrsMLm8W4Y2-g-gTFGw_aGxFZ8sQZQtneOALjKJC9A%40mail.gmail.com.

Dima Pasechnik

unread,
Jun 15, 2023, 5:51:12 AM6/15/23
to sagemat...@googlegroups.com, sage-devel
Hi Harald,

On Thu, Jun 15, 2023 at 9:09 AM Harald Schilly <harald....@gmail.com> wrote:
>
> Hi, I was indeed the one who set this up, but I vaguely remember someone else added files to the sources back then.

files are a different story, no files will do anything if the
integration is off, I suppose.

>
> In any case, I accessed zenodo and removed the link with zenodo from the archived version, and added the sagemath/sage repository.
>
> Let's see what happens, i.e. it says it is "syncing" now. ... OK, nothing happened. Seems like only new tagged releases will trigger a registration.
>
> I also found this, kind of a work around to trigger zenodo:
> https://github.com/zenodo/zenodo/issues/1463#issuecomment-1468932469

OK, but we probably should not be running this on "all" releases, only
on the missing on Zenodo ones.
Also, it's you who has to run this script, most probably I can't (as
it's somehow "owned" by you).

Dima


>
> -- Harald
>
>
>
>
> On Wed, Jun 14, 2023 at 7:49 PM Dima Pasechnik <dim...@gmail.com> wrote:
>>
>> The backup repo
>> https://github.com/sagemath/sage-archive-2023-02-01
>> apparently threw Zenodo integration off - now
>> our main repo, sage, is not associated with Zenodo any more, instead
>> that backup repo is the one associated with Zenodo.
>> Probably one had to get rid of .zenodo.* f and CITATION.cff files there
>> first, I really don't know.
>>
>> Anyhow, I'm exchanging emails with Zenodo support, trying to sort this out.
>>
>> I also don't understand why sage-archive-2023-02-01 repo has releases.
>> Was this repo created by some kind of a GitHub-specific clonig process?
>>
>> Dima
>>
>> --
>>
>> ---
>> You received this message because you are subscribed to the Google Groups "sagemath-admins" group.
>> To unsubscribe from this group and stop receiving emails from it, send an email to sagemath-admi...@googlegroups.com.
>> To view this discussion on the web visit https://groups.google.com/d/msgid/sagemath-admins/CAAWYfq2KTrsMLm8W4Y2-g-gTFGw_aGxFZ8sQZQtneOALjKJC9A%40mail.gmail.com.
>
> --
>
> ---
> You received this message because you are subscribed to the Google Groups "sagemath-admins" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to sagemath-admi...@googlegroups.com.
> To view this discussion on the web visit https://groups.google.com/d/msgid/sagemath-admins/CAGG4CB4LQ771ud0FtKiqoAXX9jZMKkA8RJEQfG4A0skM8yAUXw%40mail.gmail.com.

Harald Schilly

unread,
Jun 15, 2023, 5:56:07 AM6/15/23
to sagemat...@googlegroups.com, sage-devel
On Thu, Jun 15, 2023 at 11:51 AM Dima Pasechnik <dim...@gmail.com> wrote:
OK, but we probably should not be running this on "all" releases, only
on the missing on Zenodo ones.
Also, it's you who has to run this script, most probably I can't (as
it's somehow "owned" by you).

Well, I'll try if I can trigger version 10.0. I also noticed there are two webhooks in the github repo. Not sure what to do, I'll leave it as it is for now.

-- Harald



 

Dima Pasechnik

unread,
Jun 15, 2023, 5:59:08 AM6/15/23
to sagemath-admins, sage-devel
What webhooks?


-- Harald



 

--

---
You received this message because you are subscribed to the Google Groups "sagemath-admins" group.
To unsubscribe from this group and stop receiving emails from it, send an email to sagemath-admi...@googlegroups.com.

Harald Schilly

unread,
Jun 15, 2023, 6:09:21 AM6/15/23
to sagemat...@googlegroups.com, sage-devel
On Thu, Jun 15, 2023 at 11:59 AM Dima Pasechnik <dim...@gmail.com> wrote:
Well, I'll try if I can trigger version 10.0. I also noticed there are two webhooks in the github repo. Not sure what to do, I'll leave it as it is for now.

What webhooks?


So, anyone who has access to the github repo settings can trigger zenodo.
The URL of the webhook has a token, at the end as a query param.
Zenodo added these webhooks, they're here:

The *second* one is the one added by me, supposedly.
When you click on it and open the tab "Recent Deliveries", it just says "ping".
The other one has more activity. Maybe that's the broken one, IDK.

In any case, I coped the token from the second webhook, and ran this:

import requests
from pprint import pprint
repo = "sagemath/sage"
token = "SECRET RANDOM STRING"
headers = {"Accept": "application/vnd.github.v3+json"}
repo_response = requests.get(f"https://api.github.com/repos/{repo}", headers=headers)
release_response = requests.get(f"https://api.github.com/repos/{repo}/releases", headers=headers)
releases = release_response.json()
# only the latest one
release = releases[0]
pprint(release)
payload = {"action": "published", "release": release, "repository": repo_response.json()}
# this should say something with "200"
print(requests.post(
    f"https://zenodo.org/api/hooks/receivers/github/events/?access_token={token}",
    json=payload
))


I.e. that POST request above must be what github would do, when something happens with releases. So, this just mimics what github would do, that's all.

Long story short: as a result, we have this now: https://zenodo.org/record/8042261

-- Harald

 
Reply all
Reply to author
Forward
0 new messages