4baf4269ff
Every GLM slug that missed a DEFAULT_CONTEXT_LENGTHS key fell to the "glm" 202,752 catch-all, so compression fired at ~20% of the real window and the aux feasibility check auto-lowered the session threshold to that number (the 202,752 in the Coatue compression report). - Catalog: GLM entries from Nous + OpenRouter /v1/models (2026-09-09): 5.3 / 5.3-flash 1,310,720 (:batch/:US 1,048,576), 5.2 1,048,576, 5 / 5.1 / 4.7 / 4.6 204,800, *-turbo / 4.7-flash 202,752. - _catalog_key_matches: version separators normalised on both sides so relay slugs like z-ai-glm-5-3 hit glm-5.3 (#97398; approach from PR #97412 by @shellybotmoyer, re-implemented on current main). - get_custom_provider_context_length: entry-level custom_providers[].context_length backs every model the entry serves when no per-model override exists, so /model switch stops dropping it (#98387; from PR #98396 by @liuhao1024, re-implemented on the config_providers sibling). - check_compression_model_feasibility: when aux compression is the main model on the main route, reuse the main model's resolved window instead of re-resolving without its pin (#89500 mechanism 2, #45519). Closes #97398, #98387, #89500, #87825, #97595, #97820 (suffix strip already on main; catalog values now match).
43 lines
2.7 KiB
Python
43 lines
2.7 KiB
Python
"""GLM context windows: hyphenated relay slugs, provider-level custom overrides, and the aux feasibility
|
|
check inherit the right window instead of the ``glm`` 202,752 catch-all (#97398, #98387, #89500)."""
|
|
from types import SimpleNamespace
|
|
|
|
from agent import conversation_compression as cc
|
|
from agent.model_metadata import DEFAULT_CONTEXT_LENGTHS, get_model_context_length
|
|
from hermes_cli.config_providers import get_custom_provider_context_length
|
|
|
|
|
|
def test_hyphenated_relay_slug_resolves_to_specific_catalog_entry():
|
|
"""``z-ai-glm-5-3`` must hit ``glm-5.3``, not the shorter ``glm`` family entry."""
|
|
assert get_model_context_length("z-ai-glm-5-3", provider="custom") == DEFAULT_CONTEXT_LENGTHS["glm-5.3"]
|
|
assert DEFAULT_CONTEXT_LENGTHS["glm-5.3"] > DEFAULT_CONTEXT_LENGTHS["glm"]
|
|
|
|
|
|
def test_entry_level_custom_provider_context_length_backs_models_without_per_model_override():
|
|
providers = [{"name": "relay", "base_url": "https://relay.example/v1", "context_length": 1_000_000,
|
|
"models": {"pinned": {"context_length": 65536}}}]
|
|
assert get_custom_provider_context_length("pinned", "https://relay.example/v1", providers) == 65536
|
|
assert get_custom_provider_context_length("z-ai-glm-5-3", "https://relay.example/v1", providers) == 1_000_000
|
|
assert get_custom_provider_context_length("x", "https://other.example/v1", providers) is None
|
|
|
|
|
|
def test_feasibility_check_inherits_main_window_when_aux_is_the_main_model(monkeypatch):
|
|
"""Main pinned at 1M via model.context_length; aux compression = same model on the same route → the
|
|
threshold must NOT be auto-lowered to a re-resolved catalog value."""
|
|
lowered = []
|
|
monkeypatch.setattr(cc, "_lower_threshold_to_aux_context", lambda *a, **k: lowered.append(k))
|
|
import agent.auxiliary_client as aux
|
|
monkeypatch.setattr(aux, "_resolve_task_provider_model", lambda task: ("auto", None, None, None, None))
|
|
client = SimpleNamespace(base_url="https://relay.example/v1/", api_key="k")
|
|
monkeypatch.setattr(aux, "get_text_auxiliary_client", lambda task, main_runtime=None: (client, "uncatalogued-relay-model"))
|
|
import agent.model_metadata as mm
|
|
monkeypatch.setattr(mm, "get_model_context_length", lambda *a, **k: 128_000)
|
|
agent = SimpleNamespace(
|
|
compression_enabled=True, model="uncatalogued-relay-model", base_url="https://relay.example/v1", provider="custom",
|
|
_custom_providers=None, _aux_compression_context_length_config=None,
|
|
_current_main_runtime=lambda: None, status_callback=None, _emit_status=lambda m: None,
|
|
context_compressor=SimpleNamespace(context_length=1_000_000, threshold_tokens=500_000),
|
|
)
|
|
cc.check_compression_model_feasibility(agent)
|
|
assert lowered == []
|