[Atlas] Live tau-biology pipeline-status dashboard (hypothesis to notebook to result) done

← Live Dashboard Artifact Framework
Kanban-style live view of every tau hypothesis through hypothesis/notebook/run/figure/wiki-citation pipeline state with coverage_pct.

Completion Notes

Auto-completed by supervisor after successful deploy to main

Git Commits (2)

Squash merge: orchestra/task/a4c450f7-biomni-analysis-parity-port-15-use-cases (87 commits) (#717)2026-04-27
Squash merge: orchestra/task/ef6dfe00-live-tau-biology-pipeline-status-dashboa (2 commits) (#694)2026-04-27
Spec File

Goal

Tau biology is one of the densest hypothesis clusters in SciDEX. Researchers
need a single live view that traces every tau hypothesis through its current
pipeline state: which have computational notebooks, which notebooks have
been executed, which produced figures, which were cited back into the wiki,
and which open questions remain unanswered. Build this as a dashboard
artifact powered entirely by the view_spec_json DSL.

Acceptance Criteria

☐ Dashboard slug tau-biology-pipeline-status registered via
register_dashboard.
☐ view_spec_json with five data_sources, all safe-query-validated:
1. tau_hypotheseshypotheses filtered to mechanism_tag includes
tau/MAPT/tauopathy.
2. notebooks_for_tau — left-joins artifacts WHERE artifact_type
IN ('notebook','analysis')
linked via artifact_links to those
hypotheses.
3. notebook_runs — checks notebook_runs (or
processing_steps) for the most recent execution status.
4. derived_figuresartifacts WHERE artifact_type='figure'
linked to the notebooks.
5. tau_open_questions — Q-OPENQ field=tau biology top 10.
render.template = pipeline_kanban.html (new) with columns
[Hypothesis only] [Notebook drafted] [Executed] [Figured]
[Cited back]
, rows are hypotheses; cell shows artifact id +
timestamp.
☐ Materialized state row count: dashboard reports
coverage_pct = executed_count / hypothesis_count in the page header.
☐ Refresh interval 1800s; snapshot on demand via existing
POST /api/dashboard/{id}/snapshot.
☐ Pytest with synthetic hypotheses+notebooks+runs covers the kanban
bucketing and the coverage pct math.
☐ Operator check: dashboard is reachable at
/dashboard/tau-biology-pipeline-status and lists ≥1 hypothesis when
run against production DB.

Approach

  • Audit which mechanism_tag strings exist in hypotheses so the regex
  • filter is comprehensive (SELECT DISTINCT mechanism_tag FROM hypotheses
    WHERE mechanism_tag ILIKE '%tau%'
    ).
  • Build the notebook-link query off artifact_links (link_type
  • analyzes / derived_from depending on convention).
  • Add pipeline_kanban.html Jinja template to the templates dir consumed
  • by dashboard_engine._TEMPLATES.
  • Register via register_dashboard in a one-shot setup script
  • scripts/register_dashboard_tau_pipeline.py.

    Dependencies

    • e352460b-2d76 — view_spec_json DSL
    • 41620b88-115d — seed-dashboard pattern
    • Q-OPENQ tournament — supplies open-question source

    Work Log

    Sibling Tasks in Quest (Live Dashboard Artifact Framework) ↗