fix(model_switch): dedup /model picker rows when custom provider endpoint matches a built-in (#16970) (#17511)

When a user authenticates a built-in provider via env var (e.g. DASHSCOPE_API_KEY
triggers the built-in 'alibaba' row) AND defines a custom_providers entry
pointing at the same endpoint, the picker previously emitted two rows for one
endpoint. The built-in row already carries the canonical slug, curated model
list, and correct auth wiring, so the shadow custom entry is redundant.

Adds a _builtin_endpoints set populated as sections 1/2/2b emit rows. Each
entry is the provider's effective base URL (env override via base_url_env_var
wins over the static inference_base_url, so DASHSCOPE_BASE_URL-overridden
endpoints dedup correctly). Section 4 skips any grouped custom entry whose
base_url matches.

Intentionally does NOT repurpose model_catalog.enabled as a 'hide built-ins'
flag. That config controls the remote curated-manifest fetch (documented on
the model-catalog reference page) and overloading it would silently change
behavior for users who disable it for network/privacy reasons.

Three new tests:
- shadow dedup fires when endpoint matches static inference_base_url
- dedup does NOT hide custom entries on genuinely distinct endpoints
- dedup honors the base_url_env_var override path
This commit is contained in:
Teknium 2026-04-29 08:11:05 -07:00 committed by GitHub
parent fa3338c171
commit e120cd5941
No known key found for this signature in database
GPG key ID: B5690EEEBB952194
2 changed files with 179 additions and 0 deletions

View file

@ -1018,6 +1018,37 @@ def list_authenticated_providers(
results: List[dict] = []
seen_slugs: set = set() # lowercase-normalized to catch case variants (#9545)
seen_mdev_ids: set = set() # prevent duplicate entries for aliases (e.g. kimi-coding + kimi-coding-cn)
# Effective base URLs of every built-in row we emit (normalized lower+rstrip).
# Section 4 uses this to hide ``custom_providers`` entries that point at the
# same endpoint as a built-in (e.g. a user-defined "my-dashscope" on
# https://coding-intl.dashscope.aliyuncs.com/v1 collides with the built-in
# alibaba-coding-plan row when DASHSCOPE_API_KEY is present). Fixes #16970.
_builtin_endpoints: set = set()
def _norm_url(url: str) -> str:
return str(url or "").strip().rstrip("/").lower()
def _record_builtin_endpoint(slug: str) -> None:
"""Record the effective base URL for a built-in provider row.
Prefers the live env-override (e.g. DASHSCOPE_BASE_URL) over the
static inference_base_url so the dedup matches what a user typing
that URL into custom_providers would actually hit."""
try:
from hermes_cli.auth import PROVIDER_REGISTRY as _reg
except Exception:
return
pcfg = _reg.get(slug)
if not pcfg:
return
url = ""
if getattr(pcfg, "base_url_env_var", ""):
url = os.environ.get(pcfg.base_url_env_var, "") or ""
if not url:
url = getattr(pcfg, "inference_base_url", "") or ""
normed = _norm_url(url)
if normed:
_builtin_endpoints.add(normed)
data = fetch_models_dev()
@ -1124,6 +1155,7 @@ def list_authenticated_providers(
})
seen_slugs.add(slug.lower())
seen_mdev_ids.add(mdev_id)
_record_builtin_endpoint(slug)
# --- 2. Check Hermes-only providers (nous, openai-codex, copilot, opencode-go) ---
from hermes_cli.providers import HERMES_OVERLAYS
@ -1238,6 +1270,7 @@ def list_authenticated_providers(
})
seen_slugs.add(pid.lower())
seen_slugs.add(hermes_slug.lower())
_record_builtin_endpoint(hermes_slug)
# --- 2b. Cross-check canonical provider list ---
# Catches providers that are in CANONICAL_PROVIDERS but weren't found
@ -1317,6 +1350,7 @@ def list_authenticated_providers(
"source": "canonical",
})
seen_slugs.add(_cp.slug.lower())
_record_builtin_endpoint(_cp.slug)
# --- 3. User-defined endpoints from config ---
# Track (name, base_url) of what section 3 emits so section 4 can skip
@ -1526,6 +1560,15 @@ def list_authenticated_providers(
)
if _pair_key[0] and _pair_key[1] and _pair_key in _section3_emitted_pairs:
continue
# Skip if a built-in row (sections 1/2/2b) already represents this
# endpoint. Fixes #16970: a user-defined "my-dashscope" pointing at
# https://coding-intl.dashscope.aliyuncs.com/v1 duplicates the
# built-in alibaba-coding-plan row whenever DASHSCOPE_API_KEY is
# set. The built-in row carries the curated model list, correct
# auth wiring, and canonical slug — keep it and hide the shadow.
_grp_url_norm = _pair_key[1]
if _grp_url_norm and _grp_url_norm in _builtin_endpoints:
continue
results.append({
"slug": slug,
"name": grp["name"],

View file

@ -453,6 +453,142 @@ def test_list_authenticated_providers_no_duplicate_labels_across_schemas(monkeyp
)
def test_list_authenticated_providers_hides_custom_shadowing_builtin_endpoint(monkeypatch):
"""#16970: a custom_providers entry whose ``base_url`` matches a built-in
provider's endpoint should be hidden. The built-in row already represents
that endpoint with its canonical slug, curated model list, and auth wiring.
Repro: user sets ``DASHSCOPE_API_KEY`` (triggers the built-in ``alibaba``
row pointing at the static ``inference_base_url``) AND defines a
``my-alibaba`` custom provider pointing at the same URL. Before the fix,
the picker showed both rows for one endpoint.
"""
monkeypatch.setenv("DASHSCOPE_API_KEY", "sk-test")
monkeypatch.setattr(
"agent.models_dev.fetch_models_dev",
lambda: {
"alibaba": {
"name": "Alibaba Cloud (DashScope)",
"env": ["DASHSCOPE_API_KEY"],
}
},
)
monkeypatch.setattr("hermes_cli.providers.HERMES_OVERLAYS", {})
custom_providers = [
{
"name": "my-alibaba",
# Matches PROVIDER_REGISTRY['alibaba'].inference_base_url exactly.
"base_url": "https://dashscope-intl.aliyuncs.com/compatible-mode/v1",
"api_key": "sk-sp-test",
"model": "qwen3.6-plus",
"models": {"qwen3.6-plus": {"context_length": 500000}},
}
]
providers = list_authenticated_providers(
current_provider="my-alibaba",
user_providers={},
custom_providers=custom_providers,
max_models=50,
)
slugs = [p["slug"] for p in providers]
# Built-in alibaba row should be present.
assert "alibaba" in slugs, (
f"Expected built-in alibaba row, got slugs: {slugs}"
)
# Custom shadow row should be hidden — its base_url matches the built-in's.
assert not any("my-alibaba" in s for s in slugs), (
f"Custom my-alibaba should have been dedup'd against the built-in "
f"alibaba endpoint, got slugs: {slugs}"
)
def test_list_authenticated_providers_keeps_custom_with_distinct_endpoint(monkeypatch):
"""Dedup must only apply when the endpoint matches a built-in. A custom
provider on a genuinely distinct endpoint stays visible even if a
built-in is also authenticated."""
monkeypatch.setenv("DASHSCOPE_API_KEY", "sk-test")
monkeypatch.setattr(
"agent.models_dev.fetch_models_dev",
lambda: {
"alibaba": {
"name": "Alibaba Cloud (DashScope)",
"env": ["DASHSCOPE_API_KEY"],
}
},
)
monkeypatch.setattr("hermes_cli.providers.HERMES_OVERLAYS", {})
custom_providers = [
{
"name": "my-private-relay",
"base_url": "https://relay.example.internal/v1",
"api_key": "sk-relay-test",
"model": "qwen3.6-plus",
"models": {"qwen3.6-plus": {}},
}
]
providers = list_authenticated_providers(
current_provider="my-private-relay",
user_providers={},
custom_providers=custom_providers,
max_models=50,
)
slugs = [p["slug"] for p in providers]
assert any("my-private-relay" in s for s in slugs), (
f"Custom provider on distinct endpoint must stay visible, got: {slugs}"
)
def test_list_authenticated_providers_dedup_honors_base_url_env_override(monkeypatch):
"""The dedup must track the EFFECTIVE endpoint — if DASHSCOPE_BASE_URL
overrides the static inference_base_url, a custom provider pointing at
the overridden URL (not the static one) should still be recognized as
a duplicate."""
monkeypatch.setenv("DASHSCOPE_API_KEY", "sk-test")
monkeypatch.setenv(
"DASHSCOPE_BASE_URL",
"https://custom-dashscope.example.com/v1",
)
monkeypatch.setattr(
"agent.models_dev.fetch_models_dev",
lambda: {
"alibaba": {
"name": "Alibaba Cloud (DashScope)",
"env": ["DASHSCOPE_API_KEY"],
}
},
)
monkeypatch.setattr("hermes_cli.providers.HERMES_OVERLAYS", {})
custom_providers = [
{
"name": "my-dashscope-override",
# Same URL as DASHSCOPE_BASE_URL env override above.
"base_url": "https://custom-dashscope.example.com/v1",
"api_key": "sk-test",
"model": "qwen3.6-plus",
}
]
providers = list_authenticated_providers(
current_provider="alibaba",
user_providers={},
custom_providers=custom_providers,
max_models=50,
)
slugs = [p["slug"] for p in providers]
assert not any("my-dashscope-override" in s for s in slugs), (
f"Custom entry matching env-overridden built-in endpoint should be "
f"dedup'd, got: {slugs}"
)
# =============================================================================
# Tests for _get_named_custom_provider with providers: dict
# =============================================================================