task_charts__dashboard__category_total_extremum_label

Contract

1. Domain: charts
2. Scene id: dashboard
3. Source implementation domain/group: charts/dashboard
4. Query ids: largest_category_total_label, smallest_category_total_label
5. Semantic query details are recorded in query_id and trace params.
6. Controlled-unanswerable sampling may omit one category from one dashboard panel.

Implementation

1. Registered class: trace_tasks.tasks.charts.dashboard.category_total_extremum_label.ChartsDashboardCategoryTotalExtremumLabelTask
2. Prompt lookup domain/group: charts/dashboard
3. Generation is deterministic from instance_seed, explicit params, prompt bundle, renderer config, and code versions.
4. Answers and annotation are produced from the same metadata execution trace.

Annotation Contract

1. Answer schema: string_label or the literal string unanswerable.
2. Annotation schema: point_set.
3. Answerable instances annotate the answer category mark in every dashboard panel.
4. Unanswerable instances use an empty annotation array because at least one category is not shown in every dashboard panel.
5. Renderer context such as legends, axes, decorative labels, titles, and distractor text is metadata unless the task explicitly asks for it as annotation.

Program Contract

Program: if every category is shown in every dashboard panel: arg_extreme(category, sum(value(category,panel) for panel in dashboard_panels), direction={largest,smallest}); else unanswerable; output=string_label|unanswerable; annotation=point_set(answer_category_marks_across_panels)|empty_set; scene=dashboard; scope=category_total_extremum_label

Candidate set: the visible dashboard panels, linked marks, and panel labels inside the category_total_extremum_label objective scope. Operands: prompt-bound labels, categories, series names, thresholds, intervals, references, and encoded chart values, plus the active query id's comparator, direction, target role, or extremum focus when present. Operation: evaluate if every category is shown in every dashboard panel: arg_extreme over the candidate set using the filters, comparisons, aggregations, rankings, projections, or counterfactual edits named in the program expression; generation enforces a unique final answer. Output binding: answer is the string_label value bound by string_label|unanswerable. Annotation witnesses: point_set witnesses bound by point_set(answer_category_marks_across_panels)|empty_set. Answerable instances annotate the answer category mark in every dashboard panel. Unanswerable instances use an empty annotation array because at least one category is not shown in every dashboard panel. Renderer context such as legends, axes, decorative labels, titles, and distractor text is metadata unless the task explicitly asks for it as annotation. Query ids: largest_category_total_label, smallest_category_total_label.

Reasoning Operations

Families: ranking, aggregation

Query Details

Query id	Program signature	Answer schema	Annotation schema
largest_category_total_label	selection.category_total_extremum_label	string_label_or_unanswerable	point_set
smallest_category_total_label	selection.category_total_extremum_label	string_label_or_unanswerable	point_set