Fix: templates with no painted domains resolve to null (VFB_00050000)

[Robbie1977]

Problem

get_term_info?id=VFB_00050000 resolves to null (HTTP 200), even though the term is a valid, indexed template — L1 larval CNS ssTEM - Cardona/Janelia ("Seymour"), the L1EM larval template.

The server log blames SOLR:

Finished '.../solr/vfb_json/select/?q=id%3AVFB_00050000&wt=json' ... status 200
No results found for ID 'VFB_00050000'
Error details: list index out of range
Error accessing SOLR server!

…but SOLR is fine. q=id:VFB_00050000 on vfb_json returns numFound=1 with a well-formed term_info. The doc is found; parsing crashes.

Root cause

term_info_parse_object iterates template_domains and runs int(image.index[0]) on every entry unconditionally (vfb_queries.py). Templates with no painted anatomical domains still emit a single empty domain from the pipeline (index: [], null anatomical_individual, empty center). image.index[0] on the empty list raises IndexError: list index out of range.

That exception propagates to get_term_info's broad except IndexError, which mis-reports it as "Error accessing SOLR server!" and returns None → the endpoint serves null for the whole term.

Adult templates (e.g. VFB_00101567 JRC2018Unisex, 47 domains) were unaffected because their domains are all well-formed — which is why only this larval template fails.

Note: the parallel template_channel block just above already guards this exact pattern with if len(image.index) > 0; the guard was simply never copied to the template_domains loop.

Fix

Skip degenerate/empty domains in the template_domains loop, mirroring the existing template_channel guard. One targeted change plus a live regression test.

Verification

• Before (unpatched HEAD): get_term_info('VFB_00050000') → None, reproducing the exact log (list index out of range → "Error accessing SOLR server!").
• After: resolves with IsTemplate: True, Id: VFB_00050000, the empty domain dropped (Domains: 0), template image retained.
• No regression: VFB_00101567 still parses with all 47 painted domains.
• New test src/test/test_template_no_painted_domains.py passes (runs live with VFBQUERY_CACHE_ENABLED=false to force the synchronous parse path).

Follow-up (not in this PR)

The except IndexError in get_term_info masks any parse-time IndexError as a SOLR access error, which is what made this hard to diagnose. Worth narrowing it (or logging a traceback) so future parse bugs surface honestly. Separately, the pipeline could emit template_domains: [] rather than a single empty placeholder domain.

🤖 Generated with Claude Code