Use the batched_update DAG with stored CSVs to update Catalog URLs

[obulat]

Problem

We have generated some CSVs with identifier and another column that we need to use to update the Catalog media table, but we don't have a way to efficiently run the media table updates.

Description

The batched update DAG is reusable DAG which can be used to perform an arbitrary batched update on a Catalog media table, while handling deadlocking and timeout concerns.
During the cleanup process in data refresh, we generate the CSVs that contain the item identifier and the cleaned up version of another column (title, url, foreign_landing_url, creator_url and tags). We need a DAG that is similar to the batched update DAG, but can use a CSV table for selecting the items that need to be updated.

It is important that this work does not delete any tags. The tag column, while present in the CSVs, should not be used.

Additional context

The CSV files are saved in the docker container of the ingestion server when we run data refresh.

[sarayourfriend]

Blocking this based on the discussion in #4417 (comment), on the foundation that we should not delete data from the catalog, and this would do so by deleting any information that is filtered during the data refresh cleanup (namely, tags).

We could unblock this by excluding tags from this work, as long as that's the only place data would be deleted as part of this, and handling tags a different way, but that's complicated by one of the big questions in that IP yet to be resolved (where/when to filter tags not meant for search). Any discussion about that should go into that IP.

[zackkrida]

This was discussed in today's priorities meeting. We decided to move forward with this PR but only apply updates to URLs. We do not want to delete any data (tags in this case).

The accuracy threshold filtering of existing tags will now be preserved and iterated on as part of the Removal of the ingestion server #3925 project.