[verified] feat(nous): drive model picker from Portal recommended-models endpoint

Replace the hardcoded _PROVIDER_MODELS["nous"] catalog (~29 entries that
had to be updated manually on every Portal model release) with a live
fetch from /api/nous/recommended-models, keyed off the user's free/paid
tier. The Portal is now the single source of truth — adding or removing
a Nous model no longer requires a Hermes release.

## What changes

hermes_cli/models.py
  - Remove the hardcoded "nous": [...] list from _PROVIDER_MODELS.
  - Add get_nous_recommended_catalog(): reuses the existing 10-minute
    TTL cache in fetch_nous_recommended_models() (no extra HTTP per
    call; shares the cache with the aux/vision model helper). Selects
    freeRecommendedModels vs paidRecommendedModels based on
    check_nous_free_tier(), preserves server-specified ordering
    (the endpoint already orders each array by "position"), and does
    case-insensitive dedup.
  - Add _nous_catalog(): exception-safe wrapper returning [] on any
    failure, so callers treat Portal unavailability as "no catalog"
    rather than a crash.
  - Rewire provider_model_ids("nous"): Portal recommended-models is
    now primary; the inference /models endpoint stays as a secondary
    live fallback for offline/misconfigured-portal resilience.
  - get_default_model_for_provider("nous") and detect_provider_for_model()
    now route through _nous_catalog() instead of the removed dict key.

hermes_cli/auth.py
  - _login_nous() swapped _PROVIDER_MODELS.get("nous", []) → _nous_catalog().

hermes_cli/main.py
  - /model command nous branch: same swap.

## Design notes

- Free-tier detection happens inside the helper, so single call sites
  don't have to plumb the tier bool around.
- On tier-detection exception, defaults to paid — matches the existing
  convention (never block paying users).
- The _AGGREGATORS gate in detect_provider_for_model()'s cross-provider
  match loop already skipped "nous" when it was in _PROVIDER_MODELS,
  so removing the key changes nothing in that loop.
- partition_nous_models_by_tier() is kept in the call sites; it becomes
  mostly a no-op on the server-tier-filtered list but preserves the
  "upgrade at {portal}" messaging for free-tier users with no free
  models available.

## Tests

tests/hermes_cli/test_nous_recommended_models.py (new, 22 tests):
  - Server-order preservation
  - Free vs paid routing, auto-detection + exception-defaults-to-paid
  - Empty / missing-field / malformed entries → []
  - Case-insensitive dedup preserving first-seen casing
  - provider_model_ids("nous") rewiring: Portal-first, inference fallback,
    force_refresh propagation, exception-falls-through
  - _PROVIDER_MODELS["nous"] is absent
  - get_default_model_for_provider("nous") Portal-driven,
    non-nous providers unaffected
  - detect_provider_for_model() bare-name + current-provider paths
  - _nous_catalog() swallows exceptions → []
  - curated_models_for_provider("nous") routes through Portal

tests/hermes_cli/test_auth_nous_provider.py:
  - Add get_nous_recommended_catalog stub to _patch_login_internals so
    the login flow has models to present without a live network call.

## Verification

scripts/run_tests.sh tests/hermes_cli/ tests/test_empty_model_fallback.py
  tests/acp/test_server.py tests/test_tui_gateway_server.py
  tests/agent/test_bedrock_integration.py
  → 2752 passed. The 4 remaining failures + 1 collection race are
    pre-existing on main (unrelated — skills filtering, tip length,
    Linux stdlib ssl quirk, xdist race), confirmed via git-stash diff.
This commit is contained in:
Ben 2026-04-24 15:29:56 +10:00
parent 6fdbf2f2d7
commit 71db091868
5 changed files with 412 additions and 41 deletions

View file

@ -3414,10 +3414,11 @@ def _login_nous(args, pconfig: ProviderConfig) -> None:
) )
from hermes_cli.models import ( from hermes_cli.models import (
_PROVIDER_MODELS, get_pricing_for_provider, _nous_catalog, get_pricing_for_provider,
check_nous_free_tier, partition_nous_models_by_tier, check_nous_free_tier, partition_nous_models_by_tier,
) )
model_ids = _PROVIDER_MODELS.get("nous", []) # Portal-driven catalog — already tier-filtered server-side.
model_ids = _nous_catalog()
print() print()
unavailable_models: list = [] unavailable_models: list = []

View file

@ -2177,17 +2177,16 @@ def _model_flow_nous(config, current_model="", args=None):
# login_nous already handles model selection + config update # login_nous already handles model selection + config update
return return
# Already logged in — use curated model list (same as OpenRouter defaults). # Already logged in — use curated model list fetched live from the
# The live /models endpoint returns hundreds of models; the curated list # Portal's recommended-models endpoint (tier-filtered server-side).
# shows only agentic models users recognize from OpenRouter.
from hermes_cli.models import ( from hermes_cli.models import (
_PROVIDER_MODELS, _nous_catalog,
get_pricing_for_provider, get_pricing_for_provider,
check_nous_free_tier, check_nous_free_tier,
partition_nous_models_by_tier, partition_nous_models_by_tier,
) )
model_ids = _PROVIDER_MODELS.get("nous", []) model_ids = _nous_catalog()
if not model_ids: if not model_ids:
print("No curated models available for Nous Portal.") print("No curated models available for Nous Portal.")
return return

View file

@ -107,37 +107,12 @@ def _codex_curated_models() -> list[str]:
_PROVIDER_MODELS: dict[str, list[str]] = { _PROVIDER_MODELS: dict[str, list[str]] = {
"nous": [ # The "nous" catalog is NOT hardcoded. It is fetched live from the Portal's
"moonshotai/kimi-k2.6", # /api/nous/recommended-models endpoint via get_nous_recommended_catalog()
"xiaomi/mimo-v2.5-pro", # and resolved through _nous_catalog() wherever callers look up a nous
"xiaomi/mimo-v2.5", # catalog. This keeps the Hermes model picker in sync with the Portal's
"anthropic/claude-opus-4.7", # curated list (paid vs free) without requiring a Hermes release every
"anthropic/claude-opus-4.6", # time the Portal adds/removes a model.
"anthropic/claude-sonnet-4.6",
"anthropic/claude-sonnet-4.5",
"anthropic/claude-haiku-4.5",
"openai/gpt-5.4",
"openai/gpt-5.4-mini",
"openai/gpt-5.3-codex",
"google/gemini-3-pro-preview",
"google/gemini-3-flash-preview",
"google/gemini-3.1-pro-preview",
"google/gemini-3.1-flash-lite-preview",
"qwen/qwen3.5-plus-02-15",
"qwen/qwen3.5-35b-a3b",
"stepfun/step-3.5-flash",
"minimax/minimax-m2.7",
"minimax/minimax-m2.5",
"minimax/minimax-m2.5:free",
"z-ai/glm-5.1",
"z-ai/glm-5v-turbo",
"z-ai/glm-5-turbo",
"x-ai/grok-4.20-beta",
"nvidia/nemotron-3-super-120b-a12b",
"arcee-ai/trinity-large-thinking",
"openai/gpt-5.4-pro",
"openai/gpt-5.4-nano",
],
"openai-codex": _codex_curated_models(), "openai-codex": _codex_curated_models(),
"copilot-acp": [ "copilot-acp": [
"copilot-acp", "copilot-acp",
@ -673,6 +648,81 @@ def get_nous_recommended_aux_model(
return None return None
def get_nous_recommended_catalog(
*,
free_tier: Optional[bool] = None,
portal_base_url: str = "",
force_refresh: bool = False,
) -> list[str]:
"""Return the Portal's recommended model catalog for the user's tier.
Picks ``freeRecommendedModels`` or ``paidRecommendedModels`` from the
Portal payload depending on the tier and returns the list of
``modelName`` strings **in the order specified by the server**
(the endpoint already orders each array by ``position``).
When ``free_tier`` is ``None`` (default) the user's tier is auto-detected
via :func:`check_nous_free_tier`. Pass an explicit bool to bypass the
detection useful for tests or when the caller already knows the tier.
Reuses :func:`fetch_nous_recommended_models` so repeated callers share
the 10-minute process-wide cache no extra network per lookup.
Returns ``[]`` on any failure (network, parse, empty payload). Callers
should fall back to their own default catalog.
"""
base = portal_base_url or _resolve_nous_portal_url()
payload = fetch_nous_recommended_models(base, force_refresh=force_refresh)
if not payload:
return []
if free_tier is None:
try:
free_tier = check_nous_free_tier()
except Exception:
# On any detection error, assume paid — matches get_nous_recommended_aux_model.
free_tier = False
key = "freeRecommendedModels" if free_tier else "paidRecommendedModels"
entries = payload.get(key)
if not isinstance(entries, list):
return []
# Preserve server-specified ordering; skip malformed / blank entries.
# De-dup case-insensitively while preserving original casing of the
# first occurrence — defensive against any future duplicates in the
# feed.
seen_lower: set[str] = set()
result: list[str] = []
for entry in entries:
name = _extract_model_name(entry)
if not name:
continue
key_lower = name.lower()
if key_lower in seen_lower:
continue
seen_lower.add(key_lower)
result.append(name)
return result
def _nous_catalog() -> list[str]:
"""Safely fetch the Portal-driven Nous model catalog.
Thin exception-safe wrapper around :func:`get_nous_recommended_catalog`
used by call-sites that previously read ``_PROVIDER_MODELS["nous"]``
directly. Returns ``[]`` on any failure so callers can treat it as a
missing/empty catalog.
Hits the shared 10-minute cache in :func:`fetch_nous_recommended_models`,
so repeated calls within the picker / auto-detect flows cost nothing.
"""
try:
return get_nous_recommended_catalog()
except Exception:
return []
# --------------------------------------------------------------------------- # ---------------------------------------------------------------------------
# Canonical provider list — single source of truth for provider identity. # Canonical provider list — single source of truth for provider identity.
# Every code path that lists, displays, or iterates providers derives from # Every code path that lists, displays, or iterates providers derives from
@ -801,6 +851,12 @@ def get_default_model_for_provider(provider: str) -> str:
selected a model (e.g. ``hermes auth add openai-codex`` without selected a model (e.g. ``hermes auth add openai-codex`` without
``hermes model``). ``hermes model``).
""" """
if provider == "nous":
# Nous catalog lives at the Portal — not in _PROVIDER_MODELS.
nous = _nous_catalog()
if nous:
return nous[0]
return ""
models = _PROVIDER_MODELS.get(provider, []) models = _PROVIDER_MODELS.get(provider, [])
return models[0] if models else "" return models[0] if models else ""
@ -1377,6 +1433,14 @@ def detect_provider_for_model(
name_lower = name.lower() name_lower = name.lower()
# Helper: resolve a provider's model catalog. For "nous" the catalog
# lives at the Portal (not _PROVIDER_MODELS); everything else reads
# the static dict.
def _catalog_for(pid: str) -> list[str]:
if pid == "nous":
return _nous_catalog()
return _PROVIDER_MODELS.get(pid, [])
# --- Step 0: bare provider name typed as model --- # --- Step 0: bare provider name typed as model ---
# If someone types `/model nous` or `/model anthropic`, treat it as a # If someone types `/model nous` or `/model anthropic`, treat it as a
# provider switch and pick the first model from that provider's catalog. # provider switch and pick the first model from that provider's catalog.
@ -1384,7 +1448,7 @@ def detect_provider_for_model(
# openrouter requires an explicit model name to be useful. # openrouter requires an explicit model name to be useful.
resolved_provider = _PROVIDER_ALIASES.get(name_lower, name_lower) resolved_provider = _PROVIDER_ALIASES.get(name_lower, name_lower)
if resolved_provider not in {"custom", "openrouter"}: if resolved_provider not in {"custom", "openrouter"}:
default_models = _PROVIDER_MODELS.get(resolved_provider, []) default_models = _catalog_for(resolved_provider)
if ( if (
resolved_provider in _PROVIDER_LABELS resolved_provider in _PROVIDER_LABELS
and default_models and default_models
@ -1396,7 +1460,7 @@ def detect_provider_for_model(
_AGGREGATORS = {"nous", "openrouter", "ai-gateway", "copilot", "kilocode"} _AGGREGATORS = {"nous", "openrouter", "ai-gateway", "copilot", "kilocode"}
# If the model belongs to the current provider's catalog, don't suggest switching # If the model belongs to the current provider's catalog, don't suggest switching
current_models = _PROVIDER_MODELS.get(current_provider, []) current_models = _catalog_for(current_provider)
if any(name_lower == m.lower() for m in current_models): if any(name_lower == m.lower() for m in current_models):
return None return None
@ -1701,7 +1765,18 @@ def provider_model_ids(provider: Optional[str], *, force_refresh: bool = False)
if normalized == "copilot-acp": if normalized == "copilot-acp":
return list(_PROVIDER_MODELS.get("copilot", [])) return list(_PROVIDER_MODELS.get("copilot", []))
if normalized == "nous": if normalized == "nous":
# Try live Nous Portal /models endpoint # Primary: Portal's curated recommended-models endpoint. Returns
# either the free or paid list depending on the user's tier, in
# the order specified by the Portal (position field). Shares the
# 10-minute process-wide cache with get_nous_recommended_aux_model.
try:
recommended = get_nous_recommended_catalog(force_refresh=force_refresh)
if recommended:
return recommended
except Exception:
pass
# Fallback: live Nous inference /models endpoint (whatever the
# user's API key actually has access to right now).
try: try:
from hermes_cli.auth import fetch_nous_models, resolve_nous_runtime_credentials from hermes_cli.auth import fetch_nous_models, resolve_nous_runtime_credentials
creds = resolve_nous_runtime_credentials() creds = resolve_nous_runtime_credentials()

View file

@ -381,6 +381,16 @@ class TestLoginNousSkipKeepsCurrent:
models_mod, "partition_nous_models_by_tier", models_mod, "partition_nous_models_by_tier",
lambda ids, p, free_tier=False: (ids, []), lambda ids, p, free_tier=False: (ids, []),
) )
# The nous catalog is fetched live from the Portal — stub it so the
# login flow has models to present to the user without a network call.
monkeypatch.setattr(
models_mod, "get_nous_recommended_catalog",
lambda *a, **kw: [
"xiaomi/mimo-v2-pro",
"anthropic/claude-sonnet-4.6",
"openai/gpt-5.4",
],
)
monkeypatch.setattr(ns, "prompt_enable_tool_gateway", lambda cfg: None) monkeypatch.setattr(ns, "prompt_enable_tool_gateway", lambda cfg: None)
def test_skip_keep_current_preserves_provider_and_model(self, tmp_path, monkeypatch): def test_skip_keep_current_preserves_provider_and_model(self, tmp_path, monkeypatch):

View file

@ -0,0 +1,286 @@
"""Focused smoke tests for get_nous_recommended_catalog + provider_model_ids('nous') rewiring."""
from unittest.mock import patch
def test_paid_catalog_preserves_server_order():
from hermes_cli.models import get_nous_recommended_catalog
payload = {
"paidRecommendedModels": [
{"modelName": "minimax/minimax-m2.7", "position": 0},
{"modelName": "google/gemini-3-flash-preview", "position": 1},
{"modelName": "openai/gpt-5.4", "position": 2},
],
"freeRecommendedModels": [{"modelName": "free/one", "position": 0}],
}
with patch("hermes_cli.models.fetch_nous_recommended_models", return_value=payload):
result = get_nous_recommended_catalog(free_tier=False)
assert result == [
"minimax/minimax-m2.7",
"google/gemini-3-flash-preview",
"openai/gpt-5.4",
]
def test_free_catalog_returns_free_list():
from hermes_cli.models import get_nous_recommended_catalog
payload = {
"paidRecommendedModels": [{"modelName": "paid/one"}],
"freeRecommendedModels": [
{"modelName": "free/a"},
{"modelName": "free/b"},
],
}
with patch("hermes_cli.models.fetch_nous_recommended_models", return_value=payload):
result = get_nous_recommended_catalog(free_tier=True)
assert result == ["free/a", "free/b"]
def test_empty_payload_returns_empty_list():
from hermes_cli.models import get_nous_recommended_catalog
with patch("hermes_cli.models.fetch_nous_recommended_models", return_value={}):
assert get_nous_recommended_catalog(free_tier=False) == []
assert get_nous_recommended_catalog(free_tier=True) == []
def test_missing_field_returns_empty_list():
from hermes_cli.models import get_nous_recommended_catalog
with patch(
"hermes_cli.models.fetch_nous_recommended_models",
return_value={"paidRecommendedModels": None},
):
assert get_nous_recommended_catalog(free_tier=False) == []
def test_malformed_entries_skipped():
from hermes_cli.models import get_nous_recommended_catalog
payload = {
"paidRecommendedModels": [
{"modelName": "keep/a"},
{},
"not-a-dict",
{"modelName": ""},
{"modelName": " "},
{"modelName": "keep/b"},
]
}
with patch("hermes_cli.models.fetch_nous_recommended_models", return_value=payload):
assert get_nous_recommended_catalog(free_tier=False) == ["keep/a", "keep/b"]
def test_dedup_case_insensitive_preserves_first_casing():
from hermes_cli.models import get_nous_recommended_catalog
payload = {
"paidRecommendedModels": [
{"modelName": "Foo"},
{"modelName": "foo"},
{"modelName": "FOO"},
{"modelName": "Bar"},
]
}
with patch("hermes_cli.models.fetch_nous_recommended_models", return_value=payload):
assert get_nous_recommended_catalog(free_tier=False) == ["Foo", "Bar"]
def test_auto_detect_free_tier_calls_check():
from hermes_cli.models import get_nous_recommended_catalog
payload = {
"paidRecommendedModels": [{"modelName": "paid/x"}],
"freeRecommendedModels": [{"modelName": "free/y"}],
}
with (
patch("hermes_cli.models.fetch_nous_recommended_models", return_value=payload),
patch("hermes_cli.models.check_nous_free_tier", return_value=True),
):
assert get_nous_recommended_catalog() == ["free/y"]
with (
patch("hermes_cli.models.fetch_nous_recommended_models", return_value=payload),
patch("hermes_cli.models.check_nous_free_tier", return_value=False),
):
assert get_nous_recommended_catalog() == ["paid/x"]
def test_tier_detection_exception_defaults_to_paid():
from hermes_cli.models import get_nous_recommended_catalog
payload = {
"paidRecommendedModels": [{"modelName": "paid/x"}],
"freeRecommendedModels": [{"modelName": "free/y"}],
}
with (
patch("hermes_cli.models.fetch_nous_recommended_models", return_value=payload),
patch("hermes_cli.models.check_nous_free_tier", side_effect=RuntimeError("boom")),
):
assert get_nous_recommended_catalog() == ["paid/x"]
def test_provider_model_ids_nous_uses_recommended_first():
from hermes_cli import models as m
with patch(
"hermes_cli.models.get_nous_recommended_catalog",
return_value=["portal/one", "portal/two"],
) as mock_rec:
result = m.provider_model_ids("nous")
assert result == ["portal/one", "portal/two"]
mock_rec.assert_called_once()
def test_provider_model_ids_nous_falls_back_when_recommended_empty():
"""When Portal returns [] and inference API is unreachable, result is [].
The hardcoded ``_PROVIDER_MODELS["nous"]`` catalog has been removed;
Portal is now the sole source of truth (with the inference /models
endpoint as a live fallback). No static fallback exists anymore.
"""
from hermes_cli import models as m
with (
patch("hermes_cli.models.get_nous_recommended_catalog", return_value=[]),
patch("hermes_cli.auth.resolve_nous_runtime_credentials", return_value=None),
):
result = m.provider_model_ids("nous")
assert result == []
def test_provider_model_ids_nous_falls_back_to_inference_models_endpoint():
from hermes_cli import models as m
with (
patch("hermes_cli.models.get_nous_recommended_catalog", return_value=[]),
patch(
"hermes_cli.auth.resolve_nous_runtime_credentials",
return_value={"api_key": "k", "base_url": "https://x"},
),
patch(
"hermes_cli.auth.fetch_nous_models",
return_value=["infer/a", "infer/b"],
),
):
result = m.provider_model_ids("nous")
assert result == ["infer/a", "infer/b"]
def test_provider_model_ids_nous_recommended_exception_falls_through():
from hermes_cli import models as m
with (
patch(
"hermes_cli.models.get_nous_recommended_catalog",
side_effect=RuntimeError("portal down"),
),
patch(
"hermes_cli.auth.resolve_nous_runtime_credentials",
return_value={"api_key": "k", "base_url": "https://x"},
),
patch(
"hermes_cli.auth.fetch_nous_models",
return_value=["infer/a"],
),
):
result = m.provider_model_ids("nous")
assert result == ["infer/a"]
def test_provider_model_ids_force_refresh_propagates():
from hermes_cli import models as m
with patch(
"hermes_cli.models.get_nous_recommended_catalog",
return_value=["x/y"],
) as mock_rec:
m.provider_model_ids("nous", force_refresh=True)
assert mock_rec.call_args.kwargs.get("force_refresh") is True
# ---------------------------------------------------------------------------
# Hardcoded-list removal: everything that used to read _PROVIDER_MODELS["nous"]
# now routes through _nous_catalog() → Portal.
# ---------------------------------------------------------------------------
def test_nous_key_not_in_static_provider_models():
"""_PROVIDER_MODELS must not contain a hardcoded nous entry."""
from hermes_cli.models import _PROVIDER_MODELS
assert "nous" not in _PROVIDER_MODELS
def test_get_default_model_for_provider_nous_uses_portal():
from hermes_cli import models as m
with patch(
"hermes_cli.models._nous_catalog",
return_value=["portal/first", "portal/second"],
) as mock_cat:
assert m.get_default_model_for_provider("nous") == "portal/first"
mock_cat.assert_called_once()
def test_get_default_model_for_provider_nous_empty_returns_empty_string():
from hermes_cli import models as m
with patch("hermes_cli.models._nous_catalog", return_value=[]):
assert m.get_default_model_for_provider("nous") == ""
def test_get_default_model_for_provider_other_providers_unaffected():
"""Non-nous providers must still read from _PROVIDER_MODELS."""
from hermes_cli import models as m
# gemini has a static catalog; first entry should be returned without
# touching the Portal helper.
with patch("hermes_cli.models._nous_catalog") as mock_cat:
result = m.get_default_model_for_provider("gemini")
assert result # non-empty
mock_cat.assert_not_called()
def test_detect_bare_provider_name_nous_uses_portal():
"""`/model nous` typed as a model name → switch to nous + Portal's first model."""
from hermes_cli import models as m
with patch(
"hermes_cli.models._nous_catalog",
return_value=["portal/first", "portal/second"],
):
result = m.detect_provider_for_model("nous", current_provider="anthropic")
assert result == ("nous", "portal/first")
def test_detect_bare_provider_name_nous_empty_no_switch():
"""When Portal returns [], bare `/model nous` should NOT claim a switch."""
from hermes_cli import models as m
with patch("hermes_cli.models._nous_catalog", return_value=[]):
# detect_provider_for_model should not return a spurious (nous, "")
result = m.detect_provider_for_model("nous", current_provider="anthropic")
# Either None or a non-nous match is acceptable; the forbidden outcome
# is (nous, "") or (nous, "nous").
if result is not None:
pid, model = result
assert not (pid == "nous" and model in ("", "nous"))
def test_detect_current_nous_provider_uses_portal_catalog():
"""Model already in nous's Portal catalog → no auto-switch when current=nous."""
from hermes_cli import models as m
with patch(
"hermes_cli.models._nous_catalog",
return_value=["anthropic/claude-sonnet-4.6", "openai/gpt-5.4"],
):
# User is on nous, types a model that IS in nous's Portal catalog.
# detect_provider_for_model should return None (no switch needed).
result = m.detect_provider_for_model(
"anthropic/claude-sonnet-4.6",
current_provider="nous",
)
assert result is None
def test_nous_catalog_wraps_exception_to_empty_list():
"""_nous_catalog swallows exceptions from get_nous_recommended_catalog."""
from hermes_cli import models as m
with patch(
"hermes_cli.models.get_nous_recommended_catalog",
side_effect=RuntimeError("portal down"),
):
assert m._nous_catalog() == []
def test_curated_models_for_provider_nous_routes_through_portal():
"""curated_models_for_provider('nous') → Portal list via provider_model_ids."""
from hermes_cli import models as m
with patch(
"hermes_cli.models.get_nous_recommended_catalog",
return_value=["portal/a", "portal/b"],
):
tuples = m.curated_models_for_provider("nous")
ids = [mid for mid, _ in tuples]
assert ids == ["portal/a", "portal/b"]