task_graph__pipe_network__bridge_count¶
Summary¶
- Domain:
graph - Scene id:
pipe_network - Task id:
task_graph__pipe_network__bridge_count - Objective: count open pipe segments whose removal disconnects part of the open-pipe network.
- Implementation:
src/trace_tasks/tasks/graph/pipe_network/bridge_count.py.
Query IDs¶
- Supported
query_idvalues:single. - Internal prompt key:
pipe_bridge_count. - Public sampling is at the task-id level.
Taxonomy Contract¶
- Program contract: compute all bridge edges in the open-pipe graph while ignoring blocked pipes.
- Answer schema:
integer. - Annotation schema:
segment_set. - Target bridge count, node count, grid shape, label style, color, font, background, context text, and noise are generation/render metadata, not public query branches.
Program Contract¶
Program: count(open_pipe edge where removing edge disconnects open_pipe_graph); output=integer; annotation=segment_set(bridge_pipe_endpoint_center_segments); scene=pipe_network; scope=bridge_count
Candidate set: the visible graph, tree, network, route, matrix, table, node, edge, label, weight, path, and option elements inside the bridge_count objective scope.
Operands: visible scene state and prompt-bound operands named by open_pipe, edge, where, removing, disconnects, open_pipe_graph, bridge_pipe_endpoint_center_segments, pipe_network, bridge_count.
Operation: evaluate count 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 integer schema; generation binds a unique final answer.
Annotation witnesses: annotation uses the segment_set schema; the prompt/annotation contract defines the minimal visual witnesses.
Query ids: single.
Reasoning Operations¶
Families: filtering, counting, topology
Answer And Annotation¶
- Answer type:
integer. - Annotation type:
segment_set. - Annotation marks one
[[x0,y0],[x1,y1]]node-center segment per bridge pipe. - Segment endpoint order is not semantic.
- The answer equals the number of annotation segments, and both come from the same finalized open-pipe graph.
Rendering Contract¶
- The scene shows labeled junction fittings connected by open pipes and visible blocked pipes marked with a red X.
- Blocked pipes are visible distractors and are ignored by the bridge computation.
- Visual style, fonts, panel treatment, layout jitter, context text, and post-render noise are non-semantic and recorded in trace metadata.
- Annotation projection is computed after final layout and style placement.
Prompt Contract¶
- Prompt text comes from
src/trace_tasks/resources/prompts/graph/pipe_network/graph_pipe_network_v1.json. scene_key:pipe_network.task_key:pipe_network_query.query_key:pipe_bridge_count.- Answer mode emits
{"answer": ...}. - Answer-and-annotation mode emits
{"annotation": ..., "answer": ...}withannotationmatching thesegment_setschema.