vectorvein 0.1.42__tar.gz → 0.1.43__tar.gz

{vectorvein-0.1.42 → vectorvein-0.1.43}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: vectorvein
-Version: 0.1.42
+Version: 0.1.43
 Summary: Default template for PDM package
 Author-Email: Anderson <andersonby@163.com>
 License: MIT

{vectorvein-0.1.42 → vectorvein-0.1.43}/pyproject.toml

@@ -17,7 +17,7 @@ description = "Default template for PDM package"
 name = "vectorvein"
 readme = "README.md"
 requires-python = ">=3.10"
-version = "0.1.42"
+version = "0.1.43"
 
 [project.license]
 text = "MIT"

{vectorvein-0.1.42 → vectorvein-0.1.43}/src/vectorvein/chat_clients/utils.py

@@ -116,9 +116,14 @@ def convert_type(value, value_type):
 
 def get_token_counts(text: str | dict, model: str = "", use_token_server_first: bool = False) -> int:
     if use_token_server_first and settings.token_server is not None:
+        base_url = (
+            settings.token_server.url
+            if settings.token_server.url is not None
+            else f"http://{settings.token_server.host}:{settings.token_server.port}"
+        )
         _, response = (
             Retry(httpx.post)
-            .args(url=settings.token_server, json={"text": text, "model": model}, timeout=None)
+            .args(url=f"{base_url}/count_tokens", json={"text": text, "model": model}, timeout=None)
             .retry_times(5)
             .sleep_time(1)
             .run()
@@ -199,7 +204,7 @@ def get_token_counts(text: str | dict, model: str = "", use_token_server_first:
         if isinstance(endpoint_id, dict):
             endpoint_id = endpoint_id["endpoint_id"]
         endpoint = settings.get_endpoint(endpoint_id)
-        url = f"{endpoint.api_base}/models/{model_setting.id}:countTokens"
+        base_url = f"{endpoint.api_base}/models/{model_setting.id}:countTokens"
         params = {"key": endpoint.api_key}
         request_body = {
             "contents": {
@@ -211,7 +216,7 @@ def get_token_counts(text: str | dict, model: str = "", use_token_server_first:
         }
         _, response = (
             Retry(httpx.post)
-            .args(url, json=request_body, params=params, timeout=None)
+            .args(base_url, json=request_body, params=params, timeout=None)
             .retry_times(5)
             .sleep_time(10)
             .run()

{vectorvein-0.1.42 → vectorvein-0.1.43}/src/vectorvein/server/token_server.py

@@ -31,16 +31,16 @@ def run_token_server(host: str | None = None, port: int | None = None):
         port (int): 服务器端口。
     """
     if host is None or port is None:
-        token_server_url = settings.token_server
-        if token_server_url is None:
+        if settings.token_server is None:
             raise ValueError("Token server is not enabled.")
 
-        _host, _port = token_server_url.split(":")
+        _host = settings.token_server.host
+        _port = settings.token_server.port
     else:
         _host = host
         _port = port
 
-    uvicorn.run(token_server, host=_host, port=int(_port))
+    uvicorn.run(token_server, host=_host, port=_port)
 
 
 if __name__ == "__main__":

{vectorvein-0.1.42 → vectorvein-0.1.43}/src/vectorvein/settings/__init__.py

@@ -9,11 +9,17 @@ from ..types.enums import BackendType
 from ..types.llm_parameters import BackendSettings, EndpointSetting
 
 
+class Server(BaseModel):
+    host: str
+    port: int
+    url: Optional[str]
+
+
 class Settings(BaseModel):
     endpoints: List[EndpointSetting] = Field(
         default_factory=list, description="Available endpoints for the LLM service."
     )
-    token_server: Optional[str] = Field(default=None, description="Token server address. Format: host:port")
+    token_server: Optional[Server] = Field(default=None, description="Token server address. Format: host:port")
 
     anthropic: BackendSettings = Field(default_factory=BackendSettings, description="Anthropic models settings.")
     deepseek: BackendSettings = Field(default_factory=BackendSettings, description="Deepseek models settings.")