Checked for duplicates
Yes - I've already checked
🐛 Describe the bug
The per-product-type breakdown in the harvest run summary log (e.g. Product_Bundle, Product_Collection, Product_Document, Product_SPICE_Kernel) does not always match the actual number of successfully-loaded labels of that type, even though the overall Label Success total is correct.
Observed during review of #137: a test harvest run logged Product_Collection: 4, but only 3 LABEL SUCCESS lines for collection LIDVIDs actually appeared in that same run's log output. All products are expected to be counted regardless of type, and the per-type counts should reconcile with the individual LABEL SUCCESS entries.
🕵️ Expected behavior
Each per-product-type count in the summary should equal the number of LABEL SUCCESS entries actually logged for that type in the same run, and the per-type counts should sum to the overall Label Success total.
📜 To Reproduce
- Run harvest against a small dataset with a known, hand-countable mix of product types (bundle, collection, document, etc.)
- Review the
SUMMARY per-product-type breakdown at the end of the run
- Manually count the
LABEL SUCCESS log lines by product type
- Compare — counts diverge for at least one product type while the overall
Label Success total remains correct
🖥 Environment Info
📚 Version of Software Used
1.3.0-SNAPSHOT
🩺 Test Data / Additional context
Follow-on from #137 review discussion (2026-08-27). Reporter counted 3 bundles, 3 collections, 1 document, and 13 kernels in a harvest run, but the summary reported Product_Collection: 4. Split out per maintainer request, to be addressed via the test harness rather than ad-hoc.
🦄 Related requirements
N/A
For Internal Dev Team To Complete
⚙️ Engineering Details
🎉 Integration & Test
Checked for duplicates
Yes - I've already checked
🐛 Describe the bug
The per-product-type breakdown in the harvest run summary log (e.g.
Product_Bundle,Product_Collection,Product_Document,Product_SPICE_Kernel) does not always match the actual number of successfully-loaded labels of that type, even though the overallLabel Successtotal is correct.Observed during review of #137: a test harvest run logged
Product_Collection: 4, but only 3LABEL SUCCESSlines for collection LIDVIDs actually appeared in that same run's log output. All products are expected to be counted regardless of type, and the per-type counts should reconcile with the individualLABEL SUCCESSentries.🕵️ Expected behavior
Each per-product-type count in the summary should equal the number of
LABEL SUCCESSentries actually logged for that type in the same run, and the per-type counts should sum to the overallLabel Successtotal.📜 To Reproduce
SUMMARYper-product-type breakdown at the end of the runLABEL SUCCESSlog lines by product typeLabel Successtotal remains correct🖥 Environment Info
📚 Version of Software Used
1.3.0-SNAPSHOT
🩺 Test Data / Additional context
Follow-on from #137 review discussion (2026-08-27). Reporter counted 3 bundles, 3 collections, 1 document, and 13 kernels in a harvest run, but the summary reported
Product_Collection: 4. Split out per maintainer request, to be addressed via the test harness rather than ad-hoc.🦄 Related requirements
N/A
For Internal Dev Team To Complete
⚙️ Engineering Details
🎉 Integration & Test