Clear multipart pending flag without resource_patch

[jguo144]

Description

When finishing or aborting a multipart upload, clearing cloudstorage_multipart_pending no longer goes through resource_patch (and thus avoids the full package update / Solr indexing path). The flag is removed from resource extras with a direct SQLAlchemy update, flag_modified, and Session.commit().

Context

Multipart finish/abort only needs to drop an extra on the resource. Using resource_patch pulls in heavier CKAN action machinery than necessary, sometimes causing timeouts.

Changes

• Add _clear_cloudstorage_multipart_pending(resource_id) in multipart.py.
• Call it from finish_multipart and abort_multipart instead of resource_patch.
• Capture resource_id / upload_name before deleting the multipart upload row so logging and storage calls stay safe.

[orca-security-us]

Orca Security Scan Summary

Status	Check	Issues by priority
Passed	Infrastructure as Code	0   0   0   0	View in Orca
Passed	Secrets	0   0   0   0	View in Orca
Passed	Vulnerabilities	0   0   0   0	View in Orca

[nolanpruett]

LGTM!