Skip to content

Fix/empty tags yaml#9

Merged
Robbie1977 merged 18 commits into
Robbie1977:masterfrom
VirtualFlyBrain:fix/empty-tags-yaml
Jun 28, 2026
Merged

Fix/empty tags yaml#9
Robbie1977 merged 18 commits into
Robbie1977:masterfrom
VirtualFlyBrain:fix/empty-tags-yaml

Conversation

@Robbie1977

Copy link
Copy Markdown
Owner

No description provided.

Robbie1977 and others added 18 commits May 11, 2026 09:46
…p-fixes

Anatomy diagrams: post-merge follow-up fixes
Static audit after the previous follow-up still flagged three pairwise
rect overlaps on Schlegel Fig 1:

  AL  x ORN  30x20 px overlap (AL covered the ORN cell drawing).
  AL  x ALLN  15x5 px overlap (ALLN y2 reached into AL y1).
  mPN x LHN   5x20 px overlap (LHN x2 reached into mPN x1).

These were missed by the previous TL/centre/BR hit-test because the
overlap sits in the body of the rect, not at probed corners.

Two changes:

1. Reorder the AL/ALLN/ORN cluster so the cell-type rects come before
   the region rect. AL is a brain region; ORN and ALLN are cell-type
   drawings positioned inside the AL region in the schematic, so their
   rects necessarily overlap the AL rect. Per HTML5, the first matching
   <area> in tree order wins, so putting cell types before the region
   makes the visible cells route to their own term while the rest of
   AL still routes to AL.

2. Trim ALLN.y2 from 260 -> 254 (one below AL.y1=255) so the small
   geometric overlap with AL is gone. Trim LHN.x2 from 85 -> 79 (one
   below mPN.x1=80) for the same reason, and reorder LHN before mPN.

After this commit static audit reports a single residual geometric
overlap (ORN x AL, 30x20 px); it is intentional, resolved by ordering,
and documents the cell-inside-region layout faithfully.
OCR pass with Tesseract on each natural-resolution figure surfaced visible
text labels that sit just below the cell-drawing rects and were therefore
unclickable.

Two changes:

1. Hulse Fig 1 (panel C): the AB(R) / AB(L) text labels are at displayed
   y~736-751 in panel C, well below the existing rects at y=590-625 that
   cover the AB region drawings. Add two new rects at the label
   positions, linking the text labels to the same per-side asymmetrical
   body classes (FBbt_00051240 right, FBbt_00051239 left).

2. Aso 2014a Fig 10 (panels D, E, F): the existing MBON panel rects end
   at y=770, but the panel labels ('MBON-y3 (1)', 'MBON-y3b'1 (1)',
   'MBON-y2a'1 (2)', 'MBON-y1pedc>a/B (1)') sit at displayed y~842-934.
   Extend each bottom-row panel rect's y2 to cover the text label below
   the cell drawing:
     MBON-y3        y2 770 -> 890
     MBON-y3b'1     y2 770 -> 890
     MBON-y2a'1     y2 770 -> 940
     MBON-y1pedc>a/B y2 770 -> 940

Static audit after these changes: zero new overlaps (Hulse 28 areas,
Aso unchanged at 45). The Schlegel ORN x AL geometric overlap from the
previous commit remains intentional and ordered.
…orrect IP language, link to cookies page

Privacy Shield was invalidated by CJEU Schrems II (July 2020) and replaced by the EU-U.S. Data Privacy Framework in July 2023 (Commission Implementing Decision (EU) 2023/1795). The 'IP address (anonymized)' line was incorrect for GA4, which transiently uses IP for geolocation and discards it rather than anonymising in the GA-UA sense. Server-log paragraph added to declare that container logs are ephemeral and not persisted.
…ldwide users

VFB has a worldwide user base of Drosophila researchers; the SensusAccess document conversion service is only available to UoE staff and students and so is unhelpful to most. The AbilityNet 'My Computer My Way' link above already covers generic browser/OS accessibility customisation for any user.
Cite McLachlan et al. (2026), bioRxiv (doi:10.64898/2026.06.16.732577),
and summarise the 30-task benchmark (25/30 MCP vs 14/30 web vs 2/30 bare;
89% vs 11% on quantification tasks). Re-date the post to 2026-06-21 so the
preprint announcement surfaces in the news feed.
A term with an empty Tags list still had its ID-prefix tag appended onto
an empty CSV string, producing "tags: [,VFB]". The leading comma is an
empty flow-sequence node, which go-yaml rejects ("did not find expected
node content"), failing the Hugo build on VFB_00017893_v8.md.

Build the tag list properly and drop empty entries before joining.
@Robbie1977 Robbie1977 merged commit 7113903 into Robbie1977:master Jun 28, 2026
1 check passed
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants