24545418eb
- hermes_cli/providers.host_mandated_api_mode: add exact-hostname clause for api.meta.ai → codex_responses (measured 0% cache on /chat/completions vs 93-99% on /responses with retention); update docstring. - hermes_cli/runtime_provider._detect_api_mode_for_url: mirror clause for api.meta.ai (exact hostname, #32243) to keep runtime resolver in lockstep. - agent/agent_init: call host_mandated_api_mode early in api_mode cascade (after explicit api_mode wins, before provider-name specials) via lazy import; single source of truth, preserves user override. - agent/transports/codex._default_prompt_cache_retention_for_request: return 24h for api.meta.ai unconditionally; build_kwargs setdefault preserves override; Bedrock branch untouched. - cli-config.yaml.example: add commented providers.meta example (api_mode auto-detected). - website/docs/developer-guide/adding-providers.md: list Meta alongside Codex/xAI as codex_responses native provider with retention note. - tests: add hermetic behavior-contract suites for mandate, retention, content-addressed prompt_cache_key, reasoning passthrough, AIAgent init, usage cache reporting, model-switch override, and config roundtrip; extend test_model_switch_openai_api_mode with meta cases.
47 lines
1.6 KiB
Python
47 lines
1.6 KiB
Python
"""Usage cache reporting for Meta Responses path."""
|
|
|
|
from types import SimpleNamespace
|
|
|
|
from agent.usage_pricing import normalize_usage
|
|
|
|
|
|
def test_meta_cached_tokens_flows_to_cache_read():
|
|
usage = SimpleNamespace(
|
|
input_tokens=4000,
|
|
output_tokens=100,
|
|
input_tokens_details=SimpleNamespace(cached_tokens=3920, cache_creation_tokens=0),
|
|
output_tokens_details=None,
|
|
prompt_tokens=None,
|
|
completion_tokens=None,
|
|
)
|
|
cu = normalize_usage(usage, api_mode="codex_responses")
|
|
assert cu.cache_read_tokens == 3920
|
|
assert cu.input_tokens == 80 # 4000 - 3920
|
|
# rendered cache line would be cache=3920/4000 (98%)
|
|
total = cu.prompt_tokens
|
|
assert total == 4000
|
|
pct = (cu.cache_read_tokens / total * 100) if total else 0
|
|
assert 97 <= pct <= 99
|
|
rendered = f"cache={cu.cache_read_tokens}/{total} ({pct:.0f}%)"
|
|
assert rendered == "cache=3920/4000 (98%)"
|
|
|
|
|
|
def test_meta_cache_write_tokens():
|
|
first = SimpleNamespace(
|
|
input_tokens=4000,
|
|
output_tokens=100,
|
|
input_tokens_details=SimpleNamespace(cached_tokens=0, cache_creation_tokens=4000),
|
|
)
|
|
cu1 = normalize_usage(first, api_mode="codex_responses")
|
|
assert cu1.cache_write_tokens == 4000
|
|
assert cu1.cache_read_tokens == 0
|
|
|
|
second = SimpleNamespace(
|
|
input_tokens=4000,
|
|
output_tokens=100,
|
|
input_tokens_details=SimpleNamespace(cached_tokens=3920, cache_creation_tokens=0),
|
|
)
|
|
cu2 = normalize_usage(second, api_mode="codex_responses")
|
|
assert cu2.cache_read_tokens == 3920
|
|
assert cu2.cache_write_tokens == 0
|