task_graph__node_link__edge_between_nodes_label¶
Program Contract¶
Program: label(edge_text(edge_between(source_node, target_node))); scene=node_link; scope=edge_between_nodes_label
Candidate set: the visible graph, tree, network, route, matrix, table, node, edge, label, weight, path, and option elements inside the edge_between_nodes_label objective scope.
Operands: visible scene state and prompt-bound operands named by edge_text, edge_between, source_node, target_node, node_link, edge_between_nodes_label.
Operation: evaluate label over the candidate set using the visible graph structure, labels, weights, directions, reachability, paths, connectivity, or option-selection constraints encoded in the program expression; generation enforces a unique final answer.
Output binding: answer uses the string schema; generation binds a unique final answer.
Annotation witnesses: annotation uses the bbox schema; the prompt/annotation contract defines the minimal visual witnesses.
Query ids: single.
Reasoning Operations¶
Families: topology
Summary¶
- Domain:
graph - Scene id:
node_link - Scene:
node_link - Task id:
task_graph__node_link__edge_between_nodes_label - Objective: return the visible text label on the edge between two named nodes.
Query IDs¶
edge_between_nodes_label|directed_edge_between_nodes_label- Query ids are internal replay metadata; public sampling is at the task-id level.
Answer And Annotation¶
- Answer type:
string. - Annotation type:
bbox. - Annotation marks the pixel-space box around the queried visible edge-label text.
- Count tasks require
answer_gt.value == len(annotation_gt.value)unless the annotation schema is keyed or sequence based.
Rendering Contract¶
- The scene uses the graph-domain renderer for
node_link. - Visual style, fonts, panel treatment, layout jitter, and post-render noise are non-semantic and must be recorded in trace metadata.
- Annotation projection is computed after final layout and style placement.
- Visible edge labels are sampled from the shared label manifest with per-instance support size
16. - Edge labels are lowercase text of
3..5characters and are filtered so they do not duplicate any visible node label. - Instances cap visible labeled edges at
12; edge-label boxes must be collision-free with node boxes and other edge-label boxes. - Edge-label task rendering uses a
960x720canvas and keeps node labels at20pxwhile drawing edge-label text at22px.
Prompt Contract¶
- Prompt text comes from graph prompt templates and scene config, not hardcoded user-facing text.
- Answer mode emits
{"answer": ...}. - Answer-and-annotation mode emits
{"annotation": ..., "answer": ...}with annotation matching the schema above.