anatScRNAseqQuery: emit Dataset cell as [label](short_form) markdown

Reviewer report on v2-dev: the Dataset column on "Single cell
transcriptomics data for adult antennal lobe projection neuron"
(and every anatScRNAseqQuery result) renders as a raw Python-dict
dump:

  {symbol=, iri=http://flybase.org/reports/FBlc0006090,
   types=[Entity, Individual, DataSet, hasScRNAseq,
          scRNAseq_DataSet],
   short_form=FBlc0006090, unique_facets=[DataSet],
   label=scRNAseq_2018_Davie}

Root cause: the RETURN clause inside get_anatomy_scrnaseq's
main_query passed the `dataset` map straight through:

  RETURN
      cluster.short_form AS id,
      apoc.text.format("[%s](%s)", [cluster.label, cluster.short_form]) AS name,
      apoc.text.join(cluster.unique_facets, '|') AS tags,
      dataset,                  <-- unwrapped Cypher map
      pubs

VFBqueryJsonProcessor's generic-cell formatter has no handler for
arbitrary maps -- it has dedicated handlers for `pubs` (nested
core.label) and image markdown, but anything else falls through to
Map.toString(), which Java prints as the brace-equals shape above.

Fix mirrors how every other clickable-label cell in VFBquery is
emitted (`name`, `parent`, `template`, etc.):

  apoc.text.format("[%s](%s)", [dataset.label, dataset.short_form])
      AS dataset

Plus two coupled touchups:

- columns_to_encode picks up 'dataset' so percent-encoding runs
  on the dataset label if it ever contains URL-unsafe chars
  (consistent with `name`).
- Schema header `dataset.type` flips metadata -> markdown so the
  Java side routes it through stripMarkdownLink for display and
  through the existing markdown-id-extraction path for chip
  navigation. Matches the v2 prod "Dataset" column behaviour
  where the cell is a clickable label routing to the dataset's
  individual page.

Verified locally against pdb via the same probe Robbie reported
the bug on (FBbt_00067123, adult antennal lobe projection neuron):
both adPN and lPN cluster rows now show `[scRNAseq_2018_Davie](FBlc0006090)`
in the wire response, which the Java processor renders as the bare
label "scRNAseq_2018_Davie" with a clickable id slot.

Patch bump 1.14.7 -> 1.14.8. No cache bucket renames; the
@with_solr_cache decorator on the wrapper invalidates entries
implicitly via the function source-hash change.

Related but out-of-scope for this commit: clusterExpression /
GenesExpressed has the same dict-dump pattern in its `anatomy`
field plus a wider column-name regression (Name vs Gene, missing
Cell type / Function). Tracking separately -- "more problematic".

Author: Robbie1977

setup.py

@@ -3,7 +3,7 @@
 
 here = path.abspath(path.dirname(__file__))
 
-__version__ = "1.14.7"
+__version__ = "1.14.8"
 
 # Get the long description from the README file
 with open(path.join(here, 'README.md')) as f:

src/vfbquery/vfb_queries.py

@@ -4310,7 +4310,7 @@ def get_anatomy_scrnaseq(anatomy_short_form: str, return_dataframe=True, limit:
             cluster.short_form AS id,
             apoc.text.format("[%s](%s)", [cluster.label, cluster.short_form]) AS name,
             apoc.text.join(cluster.unique_facets, '|') AS tags,
-            dataset,
+            apoc.text.format("[%s](%s)", [dataset.label, dataset.short_form]) AS dataset,
             pubs
         ORDER BY cluster.label
     """
@@ -4324,7 +4324,7 @@ def get_anatomy_scrnaseq(anatomy_short_form: str, return_dataframe=True, limit:
 
     # Encode markdown links
     if not df.empty:
-        columns_to_encode = ['name']
+        columns_to_encode = ['name', 'dataset']
         df = encode_markdown_links(df, columns_to_encode)
 
     if return_dataframe:
@@ -4335,7 +4335,7 @@ def get_anatomy_scrnaseq(anatomy_short_form: str, return_dataframe=True, limit:
                 "id": {"title": "ID", "type": "selection_id", "order": -1},
                 "name": {"title": "Cluster", "type": "markdown", "order": 0},
                 "tags": {"title": "Tags", "type": "tags", "order": 1},
-                "dataset": {"title": "Dataset", "type": "metadata", "order": 2},
+                "dataset": {"title": "Dataset", "type": "markdown", "order": 2},
                 "pubs": {"title": "Publications", "type": "metadata", "order": 3}
             },
             "rows": [