Add variable to disable removing sql source files during ingestion

[madewithkode]

Fixes

Fixes #3847 by @AetherUnbound

Description

Added an Airflow Variable AIRFLOW_VAR_SQL_RM_SOURCE_DATA_AFTER_INGESTION which would be used to control the removal or retention of source files used for ingestion.

Testing Instructions

Checklist

• My pull request has a descriptive title (not a vague title likeUpdate index.md).
• My pull request targets the default branch of the repository (main) or a parent feature branch.
• My commit messages follow best practices.
• My code follows the established code style of the repository.
• I added or updated tests for the changes I made (if applicable).
• I added or updated documentation (if applicable).
• I tried running the project locally and verified that there are no visible errors.
• I ran the DAG documentation generator (if applicable).

Developer Certificate of Origin

Developer Certificate of Origin

Developer Certificate of Origin
Version 1.1

Copyright (C) 2004, 2006 The Linux Foundation and its contributors.
1 Letterman Drive
Suite D4700
San Francisco, CA, 94129

Everyone is permitted to copy and distribute verbatim copies of this
license document, but changing it is not allowed.


Developer's Certificate of Origin 1.1

By making a contribution to this project, I certify that:

(a) The contribution was created in whole or in part by me and I
    have the right to submit it under the open source license
    indicated in the file; or

(b) The contribution is based upon previous work that, to the best
    of my knowledge, is covered under an appropriate open source
    license and I have the right under that license to submit that
    work with modifications, whether created in whole or in part
    by me, under the same open source license (unless I am
    permitted to submit under a different license), as indicated
    in the file; or

(c) The contribution was provided directly to me by some other
    person who certified (a), (b) or (c) and I have not modified
    it.

(d) I understand and agree that this project and the contribution
    are public and that a record of the contribution (including all
    personal information I submit with it, including my sign-off) is
    maintained indefinitely and may be redistributed consistent with
    this project or the open source license(s) involved.

[AetherUnbound]

This is a great start @madewithkode! You asked about testing, you should actually be able to test this by enabling the inaturalist_workflow locally and letting it run! It will have to download a large resource as part of the execution, so just noting that it may take some time. If everything is set up appropriately (e.g. the Variable is correctly set to false), then the last two steps of the postingestion_tasks workflow should skip and subsequent runs of the DAG locally should not need to redownload the dataset which was downloaded on the first run!

This is also a note for maintainers: there does not seem to be a way to default a Variable retrieval from within a template, and so the task will fail if the Variable is not defined. We will need to add SQL_RM_SOURCE_DATA_AFTER_INGESTION to the production Variables with the default of true, I'll go ahead and do that now since it should have no affect on anything else.

[AetherUnbound]

One spacing nit, along with a fix to the variable name (AIRFLOW_VAR is only prepended for the environment variable, it gets removed when used within Airflow).

Suggested change

	"{{ var.json.AIRFLOW_VAR_SQL_RM_SOURCE_DATA_AFTER_INGESTION}}",
	"{{ var.json.SQL_RM_SOURCE_DATA_AFTER_INGESTION }}",

[madewithkode]

you should actually be able to test this by enabling the inaturalist_workflow locally and letting it run!

Hi @AetherUnbound, Is it possible to outline actual steps to do this? Or maybe point to me to a section of the doc(if any). In the meantime, I have updated the PR with the suggested changes.

[AetherUnbound]

Ahh, apologies! You should be able to take the following steps:

1. Run just c to start the catalog stack
2. Navigate to http://localhost:9090
3. Use airflow/airflow as the username and password respectively to log in
4. Scroll down to the inaturalist_workflow and click on it
5. Click the "unpause DAG" button in the top right, a run will get kicked off
   

For more info about how Airflow works in general, check out their documentation on the UI.

[madewithkode]

Thank you so much, this makes more sense.

[AetherUnbound]

I think this could be simplified:

Suggested change

	python_callable=(lambda x, y: False if not x or not y else True),
	python_callable=(lambda *x: any(x)),

[AetherUnbound]

Oh, it looks like we'll also want to add that condition here too:

openverse/catalog/dags/providers/provider_api_scripts/inaturalist.py

Line 331 in c0105eb

"remove_api_files": "{{params.sql_rm_source_data_after_ingesting}}"

Apologies for missing that! Using the Variable there as well will prevent the dataset from being redownloaded every time.

[AetherUnbound]

I don't think we can call any here because (for complicated Airflow reasons) the any gets evaluated when the DAG is parsed, but not when it's run. When the DAG is parsed, the values passed into any will be the strings we see here. They're only converted into the actual values when the DAG is actually run. I think this should work:

Suggested change

	"remove_api_files": any(
	"{{ params.sql_rm_source_data_after_ingesting }}",
	"{{ var.json.SQL_RM_SOURCE_DATA_AFTER_INGESTION}}",
	)
	"remove_api_files": "{{ params.sql_rm_source_data_after_ingesting and var.json.SQL_RM_SOURCE_DATA_AFTER_INGESTION }}",

[madewithkode]

@AetherUnbound all suggested code changes have been made.
P.S: I am unable to verify the intended behavior on my end due to repeated failures while trying to download the inaturalist_workflow source files(they seem quite large and I have a not-so-reliable internet). I'm hoping there's a way for you to verify and okay this from your end.

[AetherUnbound]

@madewithkode thanks for making those changes! I was able to test this locally, and opted to make two other changes:

• Set the default of the parameter to False, since we now have the Variable which can act as a default/fallback
• Change the condition for removing the files to or, so only one of them has to be specified as true in order to ensure deletion

This has the effect making it so that either the parameter or the Variable being set will remove the files, but by default locally, they won't be removed. In production, since we have the Variable set to true, it will always be removed.

I was able to verify this behavior by:

• Run the DAG once (so the file was downloaded) and check that the cleanup tasks were skipped
• Mark the DAG as failed after it is completed (so a follow-up run doesn't skip during the "check previous success" step)
• Run the DAG again and verify that the load_catalog_of_life_names tasks doesn't have to redownload the COL files.

All that to say, this looks good to me and can be taken out of draft when you're ready!

[madewithkode]

@AetherUnbound All these makes sense. Proceeding to unmark the PR as draft now. Thank you for the extra effort!

[stacimc]

LGTM! I ran into some issues locally with the iNaturalist pull and load TaskGroups, but I confirmed that I see this on main as well so it's probably something in my local. I was able to just manually pass those steps and only test the changes made here. Works great for me.

[AetherUnbound]

Hope it's alright to approve this one too 😁 Thanks again @madewithkode!