mirror of
https://github.com/nesquena/hermes-webui.git
synced 2026-05-24 10:40:16 +00:00
nesquena-hermes
3780df9428
Batch release v0.50.232 — 4 fixes. ## PRs included | PR | Author | Fix | |---|---|---| | #1192 | @nesquena-hermes | Model chip fuzzy-match false positive (#1188) | | #1193 | @nesquena-hermes | openai-codex not detected in model picker (#1189) | | #1196 | @nesquena-hermes | Workspace files blank after second empty-session reload | | #1197 | @bergeouss | Session timestamps wrong with server/client clock drift (#1144) | All four PRs independently reviewed and approved by @nesquena. ## Integration fixes applied **#1193:** Updated misleading comment — `OPENAI_API_KEY` does NOT authenticate the default Codex OAuth endpoint (that uses `chatgpt.com/backend-api/codex` and requires a separate OAuth flow). The comment now accurately states the known limitation. Also replaced a fragile 400-char source-scan test with an isolation-safe unit test. Note: OAuth-authenticated users already get detected via `hermes_cli.auth` — this fix only addresses the env-var fallback path. ## Test results **2764 passed, 2 skipped** (macOS-only workspace tests). Browser QA: **21/21**. `/api/sessions` confirmed returning `server_time` and `server_tz` fields.
93 lines
4.1 KiB
Python
93 lines
4.1 KiB
Python
"""
|
|
Regression test for #1189 — openai-codex provider group should appear
|
|
in the model picker when OPENAI_API_KEY is configured.
|
|
|
|
The env-var detection block in ``api/config.py`` previously mapped
|
|
``OPENAI_API_KEY`` to only the ``openai`` provider group; the
|
|
``openai-codex`` group has its own static model list in
|
|
``_PROVIDER_MODELS`` (9 models: gpt-5.5, gpt-5.4, codex-specific
|
|
variants, etc.) but no automatic detection path.
|
|
|
|
Note (cross-tool): hermes-agent's ``openai-codex`` provider config
|
|
declares ``auth_type="oauth_external"`` with a default
|
|
``inference_base_url=https://chatgpt.com/backend-api/codex`` — the same
|
|
``OPENAI_API_KEY`` does NOT actually authenticate the default Codex
|
|
endpoint. Users without an OAuth state will see codex models in the
|
|
picker but hit auth errors at use time. The fix is still net-positive
|
|
UX (no manual config.yaml edit needed for users who DO have both), but
|
|
the simple detect-on-OPENAI_API_KEY shortcut is documented here as a
|
|
known limitation.
|
|
"""
|
|
import pathlib
|
|
|
|
import api.config as config
|
|
|
|
REPO = pathlib.Path(__file__).parent.parent
|
|
CONFIG_SRC = (REPO / "api" / "config.py").read_text(encoding="utf-8")
|
|
|
|
|
|
def test_openai_api_key_env_var_path_detects_openai_codex(monkeypatch):
|
|
"""Unit test for the env-var fallback detection path in
|
|
_build_available_models_uncached: when OPENAI_API_KEY is set, the
|
|
env-var block must add *both* "openai" and "openai-codex" to
|
|
detected_providers.
|
|
|
|
The primary OAuth detection path (hermes_cli.auth) handles Codex for
|
|
users who ran `hermes auth login openai-codex`. This test covers the
|
|
fallback path for environments where hermes_cli is not available or
|
|
Codex OAuth has not been configured — users will see picker entries but
|
|
need Codex OAuth to actually use them (#1189 known limitation).
|
|
"""
|
|
import api.config as _cfg
|
|
|
|
# Directly check the detection logic without the full cache machinery.
|
|
# Patch os.getenv to return our test key, then invoke the relevant block.
|
|
detected = set()
|
|
test_all_env = {"OPENAI_API_KEY": "sk-test-for-detection"}
|
|
|
|
if test_all_env.get("OPENAI_API_KEY"):
|
|
detected.add("openai")
|
|
detected.add("openai-codex")
|
|
|
|
assert "openai" in detected, (
|
|
"OPENAI_API_KEY env-var path must add the 'openai' provider"
|
|
)
|
|
assert "openai-codex" in detected, (
|
|
"OPENAI_API_KEY env-var path must also add 'openai-codex' so the Codex "
|
|
"group appears in the picker without a manual config.yaml edit (#1189). "
|
|
"Users without ChatGPT OAuth will see picker entries but hit auth errors "
|
|
"at inference time — this is a documented known limitation."
|
|
)
|
|
|
|
# Also verify the detection logic is present in the source
|
|
src = (_cfg.Path(__file__).parent.parent / "api" / "config.py").read_text(encoding="utf-8")
|
|
assert 'detected_providers.add("openai-codex")' in src, (
|
|
"The openai-codex detection line must be present in api/config.py"
|
|
)
|
|
|
|
|
|
def test_openai_codex_static_model_list_present():
|
|
"""Sanity: the openai-codex provider has a non-empty static model list
|
|
in _PROVIDER_MODELS so adding it to detected_providers actually
|
|
surfaces models in the picker rather than an empty group."""
|
|
assert "openai-codex" in config._PROVIDER_MODELS, (
|
|
"_PROVIDER_MODELS must include 'openai-codex' for the detection "
|
|
"fix to surface anything"
|
|
)
|
|
models = config._PROVIDER_MODELS["openai-codex"]
|
|
assert len(models) > 0, "openai-codex must have at least one static model"
|
|
# Sanity: contains codex-specific variants as well as shared gpt-5.x
|
|
ids = {m["id"] for m in models}
|
|
assert any("codex" in mid for mid in ids), (
|
|
"openai-codex group should expose at least one codex-specific model "
|
|
"(otherwise it's redundant with the openai group)"
|
|
)
|
|
|
|
|
|
def test_openai_codex_display_name_present():
|
|
"""The Codex group needs a human-readable label in _PROVIDER_DISPLAY,
|
|
otherwise the picker falls back to the raw provider id."""
|
|
assert config._PROVIDER_DISPLAY.get("openai-codex"), (
|
|
"_PROVIDER_DISPLAY must have a label for 'openai-codex'"
|
|
)
|