Stream: troubleshooting

Topic: public test or demo instance with s3 direct upload support


view this post on Zulip Mattias de Hollander (Mar 03 2025 at 15:49):

I am trying out python-dvuploader for uploading large datasets. Is there a public instance available for testing that support direct uploads? The ones I have tested fall back to the native api. @Jan Range , do you have suggestions?

view this post on Zulip Jan Range (Mar 03 2025 at 15:54):

@Mattias de Hollander thanks for reaching out! Unfortunately, there is no public direct upload Dataverse instance to my knowledge. If it's solely for testing purposes, I recommend using Docker to host it locally. Would this be an option for you? Happy to help setting it up :smile:

view this post on Zulip Mattias de Hollander (Mar 03 2025 at 16:06):

Ok, than I don't look any further. And thanks for the fast reply and developing this client/library. Running it locally with docker is possible for sure. I give that a try.

view this post on Zulip Jan Range (Mar 03 2025 at 16:07):

Thanks a lot! Let me know if you need any help :smile:

view this post on Zulip Philip Durbin ๐Ÿš€ (Mar 03 2025 at 17:09):

@Mattias de Hollander the default Dataverse dev environment includes LocalStack with direct upload enabled: https://guides.dataverse.org/en/6.5/developers/dev-environment.html#quickstart

This line: https://github.com/IQSS/dataverse/blob/v6.5/docker-compose-dev.yml#L40

view this post on Zulip Jan Range (Mar 03 2025 at 17:14):

Also a note, when using LocalStack, set the DVUPLOADER_TESTING environment variable to True. Otherwise, the received URLs internally won't work, since these are aliased by Docker. The URL http://localstack/... will then be converted to http://localhost/....

view this post on Zulip Jan Range (Mar 03 2025 at 17:14):

https://github.com/gdcc/python-dvuploader?tab=readme-ov-file#development

view this post on Zulip Philip Durbin ๐Ÿš€ (Mar 03 2025 at 17:23):

@Mattias de Hollander so, I just learned that direct upload is enabled on https://beta.dataverse.org (on the default store) so you could try your test there.

view this post on Zulip Mattias de Hollander (Mar 04 2025 at 09:44):

Good to hear direct upload is enabled on the beta dataverse website. Has been a while since I worked on dataverse and was not aware of this instance. However, I can register, but I think I need an elevated role to be able to add data, or create a new dataverse first. Can someone do that? Thanks.

In the mean time I give spinning up a docker instance a try.

view this post on Zulip Philip Durbin ๐Ÿš€ (Mar 04 2025 at 13:22):

Are you able to create a dataverse (we're starting to call these "collections" by the way).

view this post on Zulip Mattias de Hollander (Mar 04 2025 at 13:40):

Not at beta.dataverse.org. Good naming update :)

view this post on Zulip Philip Durbin ๐Ÿš€ (Mar 04 2025 at 14:11):

@Mattias de Hollander it says, "Anyone with a Dataverse account can add sub dataverses":

Screenshot 2025-03-04 at 9.10.55โ€ฏAM.png

view this post on Zulip Philip Durbin ๐Ÿš€ (Mar 04 2025 at 14:12):

You don't see an "Add Data" button?

view this post on Zulip Mattias de Hollander (Mar 04 2025 at 14:26):

I do see that button. Here is my view:
image.png

And the response for both options is this:
image.png

view this post on Zulip Philip Durbin ๐Ÿš€ (Mar 04 2025 at 14:27):

500 error?!? D'oh! :doh:

view this post on Zulip Philip Durbin ๐Ÿš€ (Mar 04 2025 at 14:28):

org.postgresql.util.PSQLException: ERROR: column "displayoncreate" does not exist

view this post on Zulip Philip Durbin ๐Ÿš€ (Mar 04 2025 at 14:29):

Shoot. We're working a fix for this. Please disregard what I said. Please don't use the beta server until this is fixed. I'll let you know!

view this post on Zulip Mattias de Hollander (Mar 04 2025 at 14:30):

No problem. Happy when it is there, otherwise local options take more time but also work out. Thanks anyways.

view this post on Zulip Philip Durbin ๐Ÿš€ (Mar 04 2025 at 14:37):

If you create a collection on https://demo.dataverse.org I can point it to store with S3 direct upload configured. Lemme know.

view this post on Zulip Mattias de Hollander (Mar 04 2025 at 15:37):

That would be great. Here is one: https://demo.dataverse.org/dataverse/mhtest1

view this post on Zulip Philip Durbin ๐Ÿš€ (Mar 04 2025 at 16:02):

Ok, please try it now.

view this post on Zulip Mattias de Hollander (Mar 04 2025 at 16:26):

Thanks! I give it a try tomorrow and let you know.

view this post on Zulip Mattias de Hollander (Mar 05 2025 at 08:26):

Super, it works now with direct upload support. I will start exploring it. Thanks a lot for the support.

view this post on Zulip Philip Durbin ๐Ÿš€ (Mar 05 2025 at 12:09):

Yep, sure.

view this post on Zulip Mattias de Hollander (Mar 11 2025 at 14:31):

I've started testing the S3 direct upload support, thanks for making this possible. If you have any ideas regarding my follow-up question about setting parameters for direct upload with python-dvuploader (https://dataverse.zulipchat.com/#narrow/channel/432390-large-data/topic/setting.20parameters.20for.20direct.20upload.20with.20python-dvuploader/with/504607881), I'd appreciate your input.


Last updated: Oct 30 2025 at 06:21 UTC