Skip to content

task_symbolic__clock__elapsed_time_value

1) Identity

  1. Domain: symbolic
  2. Scene id: clock
  3. Scene: clock
  4. Task id: task_symbolic__clock__elapsed_time_value
  5. Objective: compute the forward elapsed minutes from clock A to clock B.

Program Contract

Program: clock.elapsed_time_value(scene=clock, scope=two_labeled_analog_clocks_with_minute_options, output=option_letter)

Candidate set: the six labeled minute answer cards below the two visible analog clock faces labeled A and B. Operands: the displayed start time on clock A and displayed end time on clock B. Operation: compute the forward elapsed time from A to B around the 12-hour clock. Output binding: answer is the option label whose card shows the elapsed minutes. Annotation witnesses: the scalar bbox of the selected answer card. Query ids: single.

Reasoning Operations

Families: formula_evaluation

2) Scene + Task Contract

  1. Public branch metadata: query_id
  2. Supported public query_id: single
  3. Supported non-semantic visual axes:
  4. scene_variant: classic|minimal|outline
  5. style_variant: accented|marker|studio
  6. accent_color_name: shared symbolic clock colors
  7. answer_gt.type: option_letter
  8. Answer schema: one visible option label from A through F.
  9. annotation_gt.type: bbox
  10. Annotation schema: scalar bbox for the selected answer card.
  11. Scene contract:
  12. two labeled analog clocks are shown side by side,
  13. clock A is the starting time and clock B is the ending time,
  14. six labeled minute answer cards are shown below the clocks,
  15. the prompt explicitly asks for the forward elapsed time around the 12-hour clock,
  16. the configured elapsed-minute support excludes zero and full-cycle ambiguities.

3) Prompt Contract

  1. Bundle: symbolic_clock_v1
  2. scene_key: clock
  3. task_key: clock_elapsed_time_value_query
  4. Query prompt key: elapsed_time_value
  5. Required slots:
  6. scene: object_description
  7. answer output: json_output_contract_answer_only, answer_hint, json_example_answer_only
  8. answer+annotation mode: json_output_contract, annotation_hint, answer_hint, json_example
  9. Modes: answer_only, answer_and_annotation

4) Annotation + Trace Contract

  1. Prompt-facing annotation is the scalar bbox of the selected answer card.
  2. Clock faces, clock labels, and caption text are not prompt-facing annotation.
  3. projected_annotation includes bbox and pixel_bbox.
  4. render_map includes source clock bboxes, option bboxes, and selected option bbox.
  5. execution_trace records both displayed times, raw elapsed minutes, answer label, option values, and resolved visual axes.

5) Determinism + Tests

  1. Deterministic sampling/rendering from instance_seed.
  2. Answers and annotation come from the finalized rendered start/end clocks.
  3. Behavior/trace/prompt tests: tests/test_symbolic_clock_elapsed_sequence_tasks.py
  4. Source-layout contract tests: tests/test_public_source_layout_contracts.py