mirror of
https://github.com/usips/trufflepig.git
synced 2026-10-26 04:57:58 -04:00
Joshua MoonandCodex
6c50a74dfe
Add bounded retrieval, latency, memory, and real-agent evaluation harnesses. Record numerical CUDA admission and failed default-enable retrieval/latency gates. Keep semantic retrieval opt-in and disclose command failures that confound agent comparison. Co-authored-by: Codex <[email protected]>
44 lines
896 B
Python
44 lines
896 B
Python
"""Frozen development navigation evaluation helpers.
|
|
|
|
The package keeps task labels and snapshot fingerprints on the grader side.
|
|
Use :func:`solver_task` to construct the task view given to an agent.
|
|
"""
|
|
|
|
from .schema import (
|
|
ARMS,
|
|
CORPORA,
|
|
MAX_OUTPUT_TOKENS_PER_CALL,
|
|
MAX_EMITTED_TOKENS,
|
|
MAX_TOOL_CALLS,
|
|
SCHEMA_VERSION,
|
|
SPLIT,
|
|
TASK_TIMEOUT_SECONDS,
|
|
WORKFLOW_VERSION,
|
|
corpus_roots,
|
|
grade_task,
|
|
load_manifest,
|
|
solver_task,
|
|
solver_tasks,
|
|
validate_manifest,
|
|
verify_snapshot,
|
|
)
|
|
|
|
__all__ = [
|
|
"ARMS",
|
|
"CORPORA",
|
|
"MAX_OUTPUT_TOKENS_PER_CALL",
|
|
"MAX_EMITTED_TOKENS",
|
|
"MAX_TOOL_CALLS",
|
|
"SCHEMA_VERSION",
|
|
"SPLIT",
|
|
"TASK_TIMEOUT_SECONDS",
|
|
"WORKFLOW_VERSION",
|
|
"corpus_roots",
|
|
"grade_task",
|
|
"load_manifest",
|
|
"solver_task",
|
|
"solver_tasks",
|
|
"validate_manifest",
|
|
"verify_snapshot",
|
|
]
|