task_symbolic__radial_code_wheel__code_output_label¶
1) Identity¶
- Domain:
symbolic - Scene id:
radial_code_wheel - Task id:
task_symbolic__radial_code_wheel__code_output_label - Objective contract: choose the output-label option reached by a three-symbol radial code.
Program Contract¶
Program: radial_code_wheel.code_to_output_lookup(scene=radial_code_wheel, scope=source_code_and_six_output_options, output=option_label)
Candidate set: the six visible output-label option cards labeled A..F.
Operands: the shown three-symbol source code and the three-ring radial code wheel mapping to terminal output labels.
Operation: follow the source code through the inner, middle, and outer ring symbols to find the terminal output, then select the option containing that output label.
Output binding: answer is the selected option letter.
Annotation witnesses: a point_map with inner_ring_symbol, middle_ring_symbol, and outer_ring_symbol centers for the shown code.
Query ids: single.
Reasoning Operations¶
Families: transformation
2) Scene + task contract¶
- Entities/relations:
- one three-ring radial code wheel with symbols
A|B|C|D; - one visible source code card;
- 64 unique terminal output labels on the wheel;
- six visual output-label options
A..F. - Supported
query_idvalues:single answer_gt.type:string- Default
annotation_gt.type:point_map - Annotation schema:
point_map - Alternate annotation forms: none
- Annotation witness policy:
- minimal object/primitive witnesses: the inner, middle, and outer ring-symbol centers followed by the shown code.
- annotation shape choice:
point_map, because the task binds three ordered semantic ring roles. - map annotation role names:
inner_ring_symbol,middle_ring_symbol,outer_ring_symbol - numeric/readout annotation handling: not applicable.
- answer-option annotation policy: the selected option card is not annotated; the answer field already provides the chosen option label.
- Overlap/touch policy: terminal labels and option cards are rendered from fixed non-overlapping layout slots.
3) Prompt contract¶
prompt_bundle_id:symbolic_radial_code_wheel_v1scene_key:radial_code_wheeltask_key:code_output_label- Optional query-id prompt mapping: none
- Required slots:
- answer output:
json_output_contract_answer_only,answer_hint,json_example_answer_only - answer+annotation mode:
json_output_contract,annotation_hint,answer_hint,json_example - JSON example validity rule: examples use a capital option letter answer and a three-role point-map annotation.
- Variant counts: 5 scene templates, 5 task templates, 5 output-mode templates.
- Output modes:
answer_only,answer_and_annotation
4) Determinism + constraints¶
- Seed namespaces used: task dataset, scene variant, prompt variants, and symbolic background/noise namespaces.
- Unique-answer policy: all 64 terminal labels are unique; six option labels are unique; exactly one option contains the reached output label.
- Reject/resample conditions: invalid code length/symbols, duplicate terminal labels, unsupported option count, or insufficient distractors.
- No-auto-relaxation guarantee: invalid constraints raise and retry with the next deterministic retry seed.
5) Tests¶
- Determinism test:
tests/test_symbolic_radial_code_wheel_tasks.py - Answer/annotation consistency test: same focused test file validates code-to-terminal output and annotation roles.
- Prompt metadata/placeholder test: same focused test file validates prompt bundle keys and required slots.
- Constraint-specific tests: same focused test file validates unique terminal/output option construction.