fix(models): query v1 models for llama-server endpoints (#3380)

* fix(models): query v1 models for llama-server endpoints

* test(models): accept owner kwargs in llama-server regression

routes/model_routes.py

@@ -857,7 +857,7 @@ def _ping_endpoint(base_url: str, api_key: str = None, timeout: float = 1.5) ->
             and 400 <= result["status_code"] < 500
             and result["status_code"] not in (401, 403)
         ):
-            models_url = base.rstrip("/") + "/models"
+            models_url = build_models_url(base)
             try:
                 r2 = httpx.get(models_url, headers=headers, timeout=timeout, verify=llm_verify())
                 result2 = _result_from_response(r2)

src/endpoint_resolver.py

@@ -184,7 +184,7 @@ def build_chat_url(base: str) -> str:
 
 def build_models_url(base: str) -> Optional[str]:
     """Return the provider-specific model-list endpoint URL for a base."""
-    base = resolve_url(base)
+    base = normalize_base(resolve_url(base))
     provider = _detect_provider(base)
     if provider == "anthropic":
         return _anthropic_api_root(base) + "/v1/models"

src/llm_core.py

@@ -1042,7 +1042,9 @@ def list_model_ids(
         if provider == "ollama":
             models_url = _ollama_api_root(base_chat_url) + "/tags"
         else:
-            models_url = base_chat_url.replace("/chat/completions", "/models")
+            from src.endpoint_resolver import build_models_url
+
+            models_url = build_models_url(base_chat_url)
         r = httpx.get(models_url, headers=h, timeout=timeout)
         r.raise_for_status()
         data = r.json()

src/model_context.py

@@ -297,7 +297,9 @@ def _query_context_length(endpoint_url: str, model: str) -> int:
             logger.info(f"Using known context window for {model}: {known}")
         return known or DEFAULT_CONTEXT
 
-    models_url = endpoint_url.replace("/chat/completions", "/models")
+    from src.endpoint_resolver import build_models_url
+
+    models_url = build_models_url(endpoint_url)
     try:
         r = httpx.get(models_url, timeout=REQUEST_TIMEOUT)
         if r.is_success:

tests/test_llama_server_models_url.py

@@ -0,0 +1,58 @@
+"""Regression coverage for llama-server style /v1 model-list endpoints (#3330)."""
+
+import httpx
+
+from src import endpoint_resolver, llm_core, model_context
+
+
+def test_build_models_url_accepts_v1_base_and_chat_url(monkeypatch):
+    monkeypatch.setattr(endpoint_resolver, "resolve_url", lambda url: url)
+
+    assert (
+        endpoint_resolver.build_models_url("http://127.0.0.1:8080/v1")
+        == "http://127.0.0.1:8080/v1/models"
+    )
+    assert (
+        endpoint_resolver.build_models_url("http://127.0.0.1:8080/v1/chat/completions")
+        == "http://127.0.0.1:8080/v1/models"
+    )
+
+
+def test_llm_core_list_model_ids_queries_models_for_v1_base(monkeypatch):
+    monkeypatch.setattr(endpoint_resolver, "resolve_url", lambda url: url)
+    monkeypatch.setattr(llm_core, "_configured_cached_model_ids", lambda url, **kwargs: [])
+    seen = []
+
+    def fake_get(url, headers=None, timeout=None):
+        seen.append(url)
+        request = httpx.Request("GET", url)
+        return httpx.Response(200, json={"data": [{"id": "qwen3"}]}, request=request)
+
+    monkeypatch.setattr(llm_core.httpx, "get", fake_get)
+
+    assert llm_core.list_model_ids("http://127.0.0.1:8080/v1", timeout=1) == ["qwen3"]
+    assert seen == ["http://127.0.0.1:8080/v1/models"]
+
+
+def test_model_context_queries_models_for_v1_base(monkeypatch):
+    monkeypatch.setattr(endpoint_resolver, "resolve_url", lambda url: url)
+    seen = []
+
+    def fake_get(url, timeout=None):
+        seen.append(url)
+        request = httpx.Request("GET", url)
+        if url.endswith("/slots"):
+            return httpx.Response(404, request=request)
+        return httpx.Response(
+            200,
+            json={"data": [{"id": "qwen3", "context_length": 32768}]},
+            request=request,
+        )
+
+    monkeypatch.setattr(model_context.httpx, "get", fake_get)
+
+    assert model_context._query_context_length("http://127.0.0.1:8080/v1", "qwen3") == 32768
+    assert seen == [
+        "http://127.0.0.1:8080/slots",
+        "http://127.0.0.1:8080/v1/models",
+    ]