submission_missing_labels leaks request-scoped state after retry/failure and contaminates later ontology processing

[alexskr]

Summary:

When ontology submissions are processed in sequence, a retry or failure inside GenerateMissingLabels can leave behind request-scoped processing state. The next ontology processed on the same thread can then enter an invalid state and behave as if it inherited context from the previous ontology.

Observed behavior:

• Processing a known set of ontologies consistently triggers the retry path in missing-label generation.
• After that happens for one ontology, the next ontology in the batch may behave incorrectly.
• The leaked state is not cleared when the method exits on an error path.

I, [2026-04-06T23:18:00.337601 #1981751]  INFO -- : ["Output size 126498 in `/srv/ncbo/repository/DCTERMS/1/owlapi.xrdf`"]
I, [2026-04-06T23:18:00.579098 #1981751]  INFO -- : ["Triples /srv/ncbo/repository/DCTERMS/1/owlapi.xrdf appended in <http://data.bioontology.org/ontologies/DCTERMS/submissions/1>"]
I, [2026-04-06T23:18:00.714773 #1981751]  INFO -- : ["Extracting metadata from the ontology submission http://data.bioontology.org/ontologies/DCTERMS/submissions/1"]
E, [2026-04-06T23:18:00.788414 #1981751] ERROR -- : ["Empty page encountered. Retrying 1 times..."]
E, [2026-04-06T23:18:02.791694 #1981751] ERROR -- : ["Empty page encountered. Retrying 2 times..."]
E, [2026-04-06T23:18:04.795095 #1981751] ERROR -- : ["Empty page encountered. Retrying 3 times..."]
E, [2026-04-06T23:18:06.799265 #1981751] ERROR -- : ["Empty page encountered. Retrying 4 times..."]
E, [2026-04-06T23:18:08.803101 #1981751] ERROR -- : ["Empty page encountered. Retrying 5 times..."]
E, [2026-04-06T23:18:10.806497 #1981751] ERROR -- : ["Empty page encountered. Retrying 6 times..."]
E, [2026-04-06T23:18:12.810732 #1981751] ERROR -- : ["Empty page encountered. Retrying 7 times..."]
E, [2026-04-06T23:18:14.814104 #1981751] ERROR -- : ["Empty page encountered. Retrying 8 times..."]
E, [2026-04-06T23:18:16.817508 #1981751] ERROR -- : ["Empty page encountered. Retrying 9 times..."]
E, [2026-04-06T23:18:18.820967 #1981751] ERROR -- : ["Empty page encountered. Retrying 10 times..."]
E, [2026-04-06T23:18:20.824379 #1981751] ERROR -- : ["Empty page 1 of 1 persisted after retrying 10 times. Missing Labels Generation of 1 aborted..."]
E, [2026-04-06T23:18:20.850927 #1981751] ERROR -- : Failed, exception: RuntimeError: Empty page 1 of 1 persisted after retrying 10 times. Missing Labels Generation of 1 aborted...
/opt/ontoportal/ncbo_cron/vendor/bundle/ruby/3.2.0/bundler/gems/ontologies_linked_data-a4bbd57fcaa7/lib/ontologies_linked_data/services/submission_process/operations/submission_missing_labels.rb:121:in `block in loop_classes'
/usr/local/rbenv/versions/3.2.10/lib/ruby/3.2.0/benchmark.rb:311:in `realtime'
/opt/ontoportal/ncbo_cron/vendor/bundle/ruby/3.2.0/bundler/gems/ontologies_linked_data-a4bbd57fcaa7/lib/ontologies_linked_data/services/submission_process/operations/submission_missing_labels.rb:70:in `loop_classes'
/opt/ontoportal/ncbo_cron/vendor/bundle/ruby/3.2.0/bundler/gems/ontologies_linked_data-a4bbd57fcaa7/lib/ontologies_linked_data/services/submission_process/operations/submission_missing_labels.rb:31:in `handle_missing_labels'
...
I, [2026-04-06T23:18:20.851030 #1981751]  INFO -- : There are 1598 ontologies left to process.
I, [2026-04-06T23:18:21.091113 #1981751]  INFO -- : Logging parsing output to /srv/ncbo/repository/DDANAT/24/parsing.log
I, [2026-04-06T23:18:21.093405 #1981751]  INFO -- : ["Starting to process http://data.bioontology.org/ontologies/DDANAT/submissions/24"]
I, [2026-04-06T23:18:21.110506 #1981751]  INFO -- : ["Starting to process DDANAT/submissions/24"]
E, [2026-04-06T23:18:21.138812 #1981751] ERROR -- : Failed, exception: Goo::Base::NotValidException: Object is not valid. Errors: {:version=>{:safe_text_10240=>"Attribute `version` with the value `{\"@none\"=>\"1.19\"}` must be safe text (no control or invisible Unicode characters, newlines, or disallowed punctuation) and must not exceed 10240 characters"}}
...
I, [2026-04-06T23:18:21.360526 #1981751]  INFO -- : ["Starting to process http://data.bioontology.org/ontologies/DDI/submissions/2"]
I, [2026-04-06T23:18:21.375814 #1981751]  INFO -- : ["Starting to process DDI/submissions/2"]
E, [2026-04-06T23:18:21.402941 #1981751] ERROR -- : Failed, exception: Goo::Base::NotValidException: Object is not valid. Errors: {:version=>{:safe_text_10240=>"Attribute `version` with the value `{\"@none\"=>\"$Revision$\"}` must be safe text (no control or invisible Unicode characters, newlines, or disallowed punctuation) and must not exceed 10240 characters"}, :missingImports=>{:list=>"Attribute `missingImports` with the value `{\"@none\"=>[\"http://purl.obolibrary.org/obo/iao/dev/ontology-metadata.owl\", \"http://purl.obolibrary.org/obo/iao/dev/external.owl\", \"http://purl.obolibrary.org/obo/iao/dev/externalDerived.owl\"]}` must be `Array`"}}
/opt/ontoportal/ncbo_cron/vendor/bundle/ruby/3.2.0/bundler/gems/goo-fa850f46c94e/lib/goo/base/resource.rb:250:in `save'

DCTERMS is a known trigger because it is handled as a SKOS ontology but lacks required SKOS constructs such as skos:Concept. As a result, no classes are extracted, and missing-label generation encounters an empty iteration path. In this case, the observed failure is not due to flaky 4store behavior; rather, the ontology content exposes a bug in how submission_missing_labels handles empty/failure paths and cleans up request-scoped state.