Close v2 parity gap on 8 image/dataset/expression queries

Live parity sweep against v2 prod (May 2026) flagged that v2-dev
/run_query returned a leaner column set than v2 SOLR-backed
processor for 8 queries, even after v1.10.1 batch fixed 6
image-bearing queries and v1.11.0 fixed get_similar_neurons.
Each gap is a single missing column or group the V2 frontend
already understands via COL_HEADER_MAP in uk.ac.vfb.geppetto
VFBqueryJsonProcessor, so the fix is upstream Cypher + schema
preview_columns.

Owlery-backed (Template_Space + Imaging_Technique)

Extend _owlery_query_to_results to extract template and technique
from the same anatomy_channel_image[0].channel_image SOLR structure
thumbnail already comes from. Bumps headers (both
_get_standard_query_headers and _get_neurons_part_here_headers)
and preview_columns for the three queries explicitly flagged:
NeuronsPresynapticHere, TractsNervesInnervatingHere, LineageClonesIn.

AlignedDatasets / AllDatasets (full v2 column shape)

Both functions previously returned only [id, name, tags]. Replace
with Cypher following prod XMI Datasets-available chain. Returns
pubs (Reference), license, template, technique, thumbnail,
image_count. Each branch wrapped in CALL subquery scoped to ds so
the outer carrier row stays one-per-ds. Shared helpers keep the
pair in lockstep.

PaintedDomains (Definition column)

Cypher already populated description; response builder did not list
it in headers/rows. Add it. COL_HEADER_MAP[description] = Definition.

TransgeneExpressionHere (replace delegate, add image columns)

Function previously delegated to get_expression_overlaps_here so
v2-dev got the lean 4-column output. Replace with proper Cypher
matching prod XMI dataSources.0/queries.7. Returns name (the ep),
pubs (Reference), tags, template, technique, thumbnail.
Expressed_in deferred — TODO comment in function body.

SimilarMorphologyToNBexp (Type column)

Mirrors the PR #42 / v1.11.0 fix for get_similar_neurons: add a
pipe-joined type column from primary INSTANCEOF Class inside a CALL
subquery so the outer row stays one-per-primary.

Files

src/vfbquery/vfb_queries.py
  - _owlery_query_to_results
  - _get_standard_query_headers, _get_neurons_part_here_headers
  - get_aligned_datasets, get_all_datasets (+ 3 shared helpers)
  - get_painted_domains
  - get_transgene_expression_here
  - get_similar_morphology_nb_exp
  - 5 *_to_schema preview_columns updates

Refs: projects/geppetto-vfbquery-migration/V2_V2DEV_PARITY_SWEEP.md

Author: Robbie1977