Create script to find failed split jobs

[ShanaLMoore]

Eventually we will need to find all of the works that failed to split PDFs, and re run them.

We will need a script/query to accomplish this.

For each model that must split PDFs; when the work has an attached PDF file set and does not have child works, then we need to resubmit.

The query looks something like: Select all works that have one or more file_sets with mime_type application/pdf and has no child works. It may be that we can apply a quicker filter for all works that don't have children.

We won't run this script until all of the records have been ingested.

Related:

• scientist-softserv/adventist_knapsack#224

[jeremyf]

Option 1:

Find all documents that have the mime_type_ssi field with value “application/pdf” and who’s id field’s value is not found in the split_from_pdf_id_tsi field of any document in Solr.

In discussion with @laritakr this will not work because the field is a late addition

Option 2:

Find all FileSets that have mime_type_ssi of "application/pdf" OR have a label_ssi that ends with ".pdf" (consider the archival versus reader paradigm). We then submit a job for each FileSet.

The job is then responsible for determining if it should re-split.

• This job will delete all IiifPrint::PendingRelationship records for the file set's parent work.
• When the file_set's parent work has child works; return early
• When the file_set does not have a mime_type, we likely failed to attach the file and need to perform a re-ingest of the files/work
  • see scientist-softserv/adventist_knapsack#214
• When the file_set does have a mime_type, we likely have a file and need to submit the IiifPrint::Jobs::RequestSplitPdfJob

[jeremyf]

We can't really "QA" this without unleashing it. It's a bit of a one-time (but let's be honest we'll use it often) task.