Use HEAD request in Freesound ingester for file size request

[AetherUnbound]

Description

Dependent on #1579 and WordPress/openverse-catalog#746

We should replace the requests.head call with the DelayedRequester::head call described in #1579:

https://github.com/WordPress/openverse-catalog/blob/c332465db5dfdff41f9f89cd6792c01732a570be/openverse_catalog/dags/providers/provider_api_scripts/freesound.py#L172

Implementation

• 🙋 I would be interested in implementing this feature.

[AetherUnbound]

I've bumped this to high because we're now seeing failures of the DAG occurring at this request. Specifically, if the URL 404s, this raises a KeyError for content-length (example Airflow logs). Using the machinery of the DelayedRequester::head method might help us address this.

[AetherUnbound]

Here's the failing sound in question: https://freesound.org/people/dl-jones/sounds/653078/

It looks like missing preview for a sound is indeed an issue with the Freesound data that we'll just need to account for. In these cases would we just want to skip the record entirely, since we can't preview it on our end? If that's the case I should probably make a new issue. CC @WordPress/openverse-catalog

[zackkrida]

In these cases would we just want to skip the record entirely, since we can't preview it on our end?

This is an interesting question, and I think it's okay to skip these records. The audio player is broken in FreeSound's frontend as well, which is a terrible user experience.

[AetherUnbound]

It's odd though, I just logged in and you can download the audio file itself 🤔 I'm going to make a new issue for this though since it seems separate from the HEAD request issue, we can continue discussion there.

[AetherUnbound]

I'm setting this priority to low since #1321 is the issue that's actually affecting the DAG's ability to run successfully.