feat: add Chinese UI, Responses models, and mirrored Windows setup

Localize the web console and persist custom model and capability settings.
Use OpenAI Responses exclusively and move runtime limits into environment configuration.
Bootstrap uv and the virtual environment on Windows, support optional China mirrors,
and package source distributions without environments, build output, or credentials.

Update regression tests and repository quality checks for the new behavior.

apps/admin_console/services/model_service.py

@@ -12,6 +12,8 @@
 # See the License for the specific language governing permissions and
 # limitations under the License.
 
+from artemis.config.settings import settings
+
 import json
 import logging
 import time
@@ -26,7 +28,7 @@ class ModelService:
     # parse_llm_config re-reads and re-validates the config file on every call;
     # /api/sessions calls this once per session row, so cache (provider, model_id)
     # for a short TTL. Config edits take effect within _LLM_INFO_TTL seconds.
-    _LLM_INFO_TTL = 10.0
+    _LLM_INFO_TTL = settings.ARTEMIS_MODEL_CACHE_TTL_SECONDS
     _llm_info_cache: tuple[float, str, str] | None = None
 
     @classmethod
@@ -35,8 +37,8 @@ class ModelService:
         now = time.monotonic()
         if cached and now - cached[0] < cls._LLM_INFO_TTL:
             return cached[1], cached[2]
-        provider = "google"
-        model_id = "gemini-3.7-flash"
+        provider = "openai"
+        model_id = settings.OPENAI_MODEL
         try:
             from artemis.config import parse_llm_config