type: task id: GRAPH_LIBRARY-ADD_TEMPORAL_AXIS_LABEL_SEMANTICS_FOR_POINT_VS_INTERVAL_POSITIONING title: Add temporal axis label semantics for point vs interval positioning description: Capture follow-on work for temporal axes so labels and ticks can distinguish point-in-time dates from interval or boundary-style labels, including between-tick centering when appropriate. milestone: m1-ft-analytics-analyst-pilot owner: data-viz-designer-engineer status: not_started priority: p2

Add temporal axis label semantics for point vs interval positioning

Superseded (2026-03-27). This narrowly scoped follow-on was consolidated into Consolidate semantic chart defaults and rendering control-surface follow-ons so the active backlog carries one M2 umbrella task instead of many M1 leaf items. Keep the notes below as reference detail, not as a separately scheduled task.

Problem

Capture follow-on work for temporal axes so labels and ticks can distinguish point-in-time dates from interval or boundary-style labels, including between-tick centering when appropriate.

Context

During M1 structure-lab work, we noticed that temporal axes are currently treated too generically. There is an important semantic difference between:

point-in-time marks that should sit directly on their ticks
interval or bucket labels that represent a span, where the label should feel centered between boundaries rather than pinned to a single boundary tick

Examples:

if the data point is simply "May", it can be acceptable for the mark to sit on the corresponding tick
if the axis is labeling actual dates such as "January 15", the label should not imply that the tick itself means "January"; it should feel centered in the month span between the relevant boundaries

Today Dataface provides only generic temporal posture defaults plus raw per-chart axis overrides. That is not enough to express the semantic distinction above at a systems level.

Relevant files:

dataface/core/render/chart/presentation.py
dataface/core/render/chart/standard_renderer.py
dataface/core/compile/chart_structures/m1-candidate.yml
dataface/core/render/chart/DESIGN.md
examples/playground/faces/labs/m1-structure-bars-lines-lab.yml

Note: - chart_structures/m1-candidate.yml and m1-structure-bars-lines-lab.yml were captured from the parallel codex/m1-structure-lab branch during the March 26, 2026 design session. If they are missing on main, treat them as historical repro references and recreate the scenario in a fresh lab surface.

Possible Solutions

Keep handling temporal-axis positioning with per-chart settings.x_axis overrides. This is flexible for one-off experiments, but it is manual and does not create a coherent system behavior.
Add a few more static structure presets for temporal axes. This may help for fixed postures, but it still does not solve the semantic problem of deciding when labels/ticks should be point-centered vs interval-centered.
Recommended: add semantic temporal-axis logic in the renderer/presentation layer. Teach Dataface to distinguish point-style temporal encodings from interval/boundary encodings, then choose the appropriate axis positioning defaults automatically while still allowing authored overrides.

Plan

Audit the current temporal-axis behavior in get_smart_x_axis_config().
Define the intended semantic modes for temporal axes: - point-centered - interval/boundary-centered
Decide how Dataface should infer or receive that intent.
Apply the resulting axis config in the shared presentation layer rather than relying on per-chart hacks.
Add regression coverage with representative temporal charts.

Implementation Progress

Not started.

March 26, 2026 notes:

Confirmed that current controls are low-level axis properties such as labelAlign, labelBaseline, labelAngle, tickCount, and authored axis overrides.
Confirmed that shared structure posture is too coarse to encode the semantics of point-vs-interval temporal labeling.
Decided not to patch around this in the lab because the goal is system behavior, not chart-local hacks.

QA Exploration

N/A for now. This task has not been implemented yet.

[ ] QA exploration completed (or N/A for non-UI tasks)

Review Feedback