Error in grafting ocr'd text back into PDF

[duckduckgrayduck]

What I did:
I'm adding the ability to optionally force OCR a document after uploading them via URL in the Python wrapper.
urls= ["https://www.chicago.gov/content/dam/city/depts/dcd/tif/24reports/T_063_CanalCongressAR24.pdf", "https://www.chicago.gov/content/dam/city/depts/dcd/tif/24reports/T_072_24thMichiganAR24.pdf"]
uploaded_docs = client.documents.upload_urls(urls, force_ocr=True, ocr_engine="textract")

One document succeeded just fine, the other encountered this error:
https://muckrock.sentry.io/issues/6895388692/?alert_rule_id=1010155&alert_timestamp=1758569111535&alert_type=email&notification_uuid=b002be3c-4651-4b4e-98b1-3e929498be0d&project=2873549

https://www.documentcloud.org/documents/26105901-t_063_canalcongressar24/

This left the document in a failed state, which could probably be handled better (present a working version of the PDF)

[allanlasser]

If we track processing jobs in the database, we could fall back to the result of the last successful job (ex. OCR job fails, fall back to upload job result).

We'll need to report these errors more clearly to users, so they know what went wrong and what steps they need to take next to get the result they want.

Right now, we show the error state but don't do any user-friendly fixes for the error message.