wehub-resource-sync
44 lines
1.2 KiB
Python
44 lines
1.2 KiB
Python
"""Regression tests for admin model-settings UI gates."""
|
|
|
|
from pathlib import Path
|
|
|
|
|
|
def _model_settings_template() -> str:
|
|
root = Path(__file__).resolve().parents[1]
|
|
return (
|
|
root / "omlx/admin/templates/dashboard/_modal_model_settings.html"
|
|
).read_text()
|
|
|
|
|
|
def _section(html: str, start_marker: str, end_marker: str) -> str:
|
|
return html.split(start_marker, 1)[1].split(end_marker, 1)[0]
|
|
|
|
|
|
def test_lightning_mtp_and_turboquant_are_not_ui_mutexed():
|
|
html = _model_settings_template()
|
|
|
|
turboquant = _section(
|
|
html,
|
|
"<!-- TurboQuant KV Cache -->",
|
|
"<!-- IndexCache (DSA models only) -->",
|
|
)
|
|
lightning_mtp = _section(
|
|
html,
|
|
"<!-- Lightning MTP (built-in MTP head speculative decoding) -->",
|
|
"<!-- Experimental Features -->",
|
|
)
|
|
|
|
assert "modelSettings.mtp_enabled" not in turboquant
|
|
assert "modelSettings.turboquant_kv_enabled" not in lightning_mtp
|
|
|
|
|
|
def test_vlm_mtp_still_conflicts_with_turboquant():
|
|
html = _model_settings_template()
|
|
vlm_mtp = _section(
|
|
html,
|
|
"<!-- VLM MTP",
|
|
"<!-- Performance",
|
|
)
|
|
|
|
assert "modelSettings.turboquant_kv_enabled" in vlm_mtp
|