`ucb_tind` Harvesting error at `map_page` step: `Type 'NoneType' cannot be serialized.`

[aturner]

Running a harvest of UCB Library collection #27084, and getting the error at the map_page step:

TypeError: Type 'NoneType' cannot be serialized.
[2026-03-12, 16:36:53 UTC] {taskinstance.py:1206} INFO - Marking task as FAILED. dag_id=harvest_collection, task_id=mapping.map_page, run_id=manual__2026-03-12T16:34:25+00:00, map_index=0, execution_date=20260312T163425, start_date=20260312T163652, end_date=20260312T163653
[2026-03-12, 16:36:53 UTC] {logging_mixin.py:188} INFO - Message sent to SNS with Message ID: 54918448-65f3-5ee8-8b89-e66488167a34
[2026-03-12, 16:36:53 UTC] {standard_task_runner.py:110} ERROR - Failed to execute job 808705 for task mapping.map_page (Type 'NoneType' cannot be serialized.; 39218)
[2026-03-12, 16:36:53 UTC] {local_task_job_runner.py:240} INFO - Task exited with return code 1

Dag run ID: https://7a8067cb-3b99-477e-a883-7e311175a9b4.c3.us-west-2.airflow.amazonaws.com/dags/harvest_collection/grid?num_runs=100&dag_run_id=manual__2026-03-12T16%3A34%3A25%2B00%3A00&task_id=mapping.map_page&tab=mapped_tasks&map_index=0

[barbarahui]

@aturner Collection 27084 has a fetcher type of OAI-PMH, with a metadataPrefix of marcxml. Our oai fetcher currently says to use either oai_qdc or oai_dc only. This means that we're fetching a dc feed for this collection. Do you want me to add marcxml as an OAI feed format?

[aturner]

@barbarahui thanks for looking into this! We've been using the marcxml metadataPrefix for all of the UCB TIND collections (e.g., https://registry.cdlib.org/admin/library_collection/collection/19333).

But I see an initial problem: collection 27084 had a typo in the "Harvest Extra Data", missing the initial m:

etadataPrefix=marcxml&set=environtourismyosemite

Fixed and re-running it again...

[barbarahui]

@aturner Ah whoops ok, I see what happened!