task_graph__node_link__unique_related_node_label¶
Program Contract¶
Program: label(unique(related_nodes(reference_node, relation))); scene=node_link; scope=unique_related_node_label
Candidate set: the visible graph, tree, network, route, matrix, table, node, edge, label, weight, path, and option elements inside the unique_related_node_label objective scope.
Operands: visible scene state and prompt-bound operands named by unique, related_nodes, reference_node, relation, node_link, unique_related_node_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 point 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:
relation - Task id:
task_graph__node_link__unique_related_node_label - Objective: return the unique node label satisfying the requested direct relation.
Query IDs¶
unique_neighbor_label|unique_successor_label|unique_predecessor_label- Query ids are internal replay metadata; public sampling is at the task-id level.
Answer And Annotation¶
- Answer type:
string. - Annotation type:
point. - Annotation marks the node-center pixel point for the unique answer node.
- 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.
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.