Set environment vars to set vendor-specific dandi-schema used by the DANDI instance

[candleindark]

See commit message for details.

This PR closes #223.

TODOs (more like blocks)

• dandi-schema vendorization getting "full runtime config" Support dynamic configuration dandi-schema#316
• dandi-archive being able to set the config from DJANGO_DANDI_ variables (PR: TODO)
• Add environment variable needed for vendorization of the publisher for to_datacite, Add vendorization of the Publisher for to_datacite dandi-schema#303

[candleindark]

@aaronkanzer @kabilar Do you know where, which repo, I can modify to apply the same changes to LINC?

[yarikoptic]

@waxlamp please review this PR from your side of infra -- it should not effect running instance but would allow for vendorization on which there is work ongoing in dandi/dandi-schema#294 and from there on.

[kabilar]

@aaronkanzer @kabilar Do you know where, which repo, I can modify to apply the same changes to LINC?

Thanks @candleindark. Here is the LINC terraform repo: https://github.com/lincbrain/linc-archive-infrastructure

[jjnesbitt]

A couple of questions.

1. What is the intended use of these vars? They are specified in the additional_django_vars section, but lack the DJANGO_* prefix on them.
2. We already have a env var DJANGO_DANDI_DOI_API_PREFIX, which specifies the same value as DANDI_DOI_PREFIX. So is this additional redundant?

[jjnesbitt]

Ah, after looking into it further it seems they are passed to the schema. I do wonder if there's a better place to set these, but I imagine not.

[candleindark]

1. What is the intended use of these vars? They are specified in the additional_django_vars section, but lack the DJANGO_* prefix on them.

Just as you noted, they are used to set the configuration in dandischema, https://github.com/dandi/dandi-schema/blob/f2ecf9d9b5cb6cff8ef0a3a1de427044cfb3e59d/dandischema/conf.py#L21-L63. My understanding is that variables defined in additional_django_vars don't have to start with the DJANGO_*. The set configuration in dandischema is be shared among dandi-cli and dandi-archive in order to "vendorize" DANDI, dandi/dandi-schema#294.

2. We already have a env var DJANGO_DANDI_DOI_API_PREFIX, which specifies the same value as DANDI_DOI_PREFIX. So is this additional redundant?

Yes, it is redundant, but since the shared config is not just used in dandi-archive having an env variable starting DJANGO_* may not be appropriate. It may be a good idea to eliminate DJANGO_DANDI_DOI_API_PREFIX in the future though.

[candleindark]

Ah, after looking into it further it seems they are passed to the schema. I do wonder if there's a better place to set these, but I imagine not.

The configuration must be set correctly before dandischema.models is imported. Providing those env vars will achieve that in dandi-archive.

[jjnesbitt]

The configuration must be set correctly before dandischema.models is imported. Providing those env vars will achieve that in dandi-archive.

Can we not use set_instance_config, instead of the env vars?

[yarikoptic]

@jjnesbitt manually how so that it is in effect at dandischema library import?

[jjnesbitt]

Manually via set_instance_config, instead of with environment variables.

[candleindark]

Manually via set_instance_config, instead of with environment variables.

If calling of set_instance_config() is necessitated because of naming of variables. I can introduce aliases to the fields of dandischema.conf.Config so that they can be set via DJANGO_DANDI_DOI_API_PREFIX and DJANGO_DANDI_INSTANCE_NAME.

[yarikoptic]

ok, let's finalize/merge

• Support dynamic configuration dandi-schema#316

and then establish needed changes in dandi-archive for .set_instance_config . @jjnesbitt would you do that or should we try a PR?

[jjnesbitt]

I can make that PR.

[yarikoptic]

@jjnesbitt @waxlamp please chime in on either you want anything else changed. It should prepare our instance for use of vendorized instance of dandischema

[candleindark]

Note
At this point, the instance config of dandischema used by dandi-archive will most likely to be set by calling set_instance_config. As to where/how to call set_instance_config in dandi-archive, please checkout dandi/dandi-archive#2483 (comment).

[candleindark]

@jjnesbitt This PR has been updated and is ready for review. All environment variables used now conform to the DJANGO_DANDI_ prefix format.

[Copilot]

Pull Request Overview

This PR adds vendor-specific configuration for DANDI instances by introducing environment variables for instance name and identifier. The changes support vendorization of the publisher for DataCite integration and provide runtime configuration for different deployment environments.

Key changes:

• Added DJANGO_DANDI_INSTANCE_NAME environment variable for both staging and production
• Added DJANGO_DANDI_INSTANCE_IDENTIFIER environment variable with RRID identifier

Reviewed Changes

Copilot reviewed 2 out of 2 changed files in this pull request and generated no comments.

File	Description
terraform/staging_pipeline.tf	Adds instance name "DANDI-STAGING" and RRID identifier for staging environment
terraform/api.tf	Adds instance name "DANDI" and RRID identifier for production environment

---

_{Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.}