vectorvein 0.1.59__tar.gz → 0.1.61__tar.gz

{vectorvein-0.1.59 → vectorvein-0.1.61}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: vectorvein
-Version: 0.1.59
+Version: 0.1.61
 Summary: VectorVein python SDK
 Author-Email: Anderson <andersonby@163.com>
 License: MIT
@@ -8,7 +8,7 @@ Requires-Python: >=3.10
 Requires-Dist: openai>=1.37.1
 Requires-Dist: tiktoken>=0.7.0
 Requires-Dist: httpx>=0.27.0
-Requires-Dist: anthropic[vertex]>=0.31.2
+Requires-Dist: anthropic[bedrock,vertex]>=0.31.2
 Requires-Dist: pydantic>=2.8.2
 Requires-Dist: Pillow>=10.4.0
 Requires-Dist: deepseek-tokenizer>=0.1.0

{vectorvein-0.1.59 → vectorvein-0.1.61}/pyproject.toml

@@ -6,7 +6,7 @@ dependencies = [
     "openai>=1.37.1",
     "tiktoken>=0.7.0",
     "httpx>=0.27.0",
-    "anthropic[vertex]>=0.31.2",
+    "anthropic[vertex,bedrock]>=0.31.2",
     "pydantic>=2.8.2",
     "Pillow>=10.4.0",
     "deepseek-tokenizer>=0.1.0",
@@ -17,7 +17,7 @@ description = "VectorVein python SDK"
 name = "vectorvein"
 readme = "README.md"
 requires-python = ">=3.10"
-version = "0.1.59"
+version = "0.1.61"
 
 [project.license]
 text = "MIT"

{vectorvein-0.1.59 → vectorvein-0.1.61}/src/vectorvein/chat_clients/anthropic_client.py

@@ -8,7 +8,14 @@ from typing import overload, Generator, AsyncGenerator, Any, Literal, Iterable
 import httpx
 from openai._types import NotGiven as OpenAINotGiven
 from openai._types import NOT_GIVEN as OPENAI_NOT_GIVEN
-from anthropic import Anthropic, AnthropicVertex, AsyncAnthropic, AsyncAnthropicVertex
+from anthropic import (
+    Anthropic,
+    AnthropicVertex,
+    AsyncAnthropic,
+    AsyncAnthropicVertex,
+    AnthropicBedrock,
+    AsyncAnthropicBedrock,
+)
 from anthropic._types import NOT_GIVEN
 from anthropic.types import (
     TextBlock,
@@ -219,6 +226,16 @@ class AnthropicChatClient(BaseChatClient):
                 access_token=self.creds.token,
                 http_client=self.http_client,
             )
+        elif self.endpoint.is_bedrock:
+            if self.endpoint.credentials is None:
+                raise ValueError("Anthropic Bedrock endpoint requires credentials")
+            return AnthropicBedrock(
+                aws_access_key=self.endpoint.credentials.get("access_key"),
+                aws_secret_key=self.endpoint.credentials.get("secret_key"),
+                aws_region=self.endpoint.region,
+                base_url=self.endpoint.api_base,
+                http_client=self.http_client,
+            )
         elif self.endpoint.api_schema_type == "default":
             return Anthropic(
                 api_key=self.endpoint.api_key,
@@ -384,7 +401,7 @@ class AnthropicChatClient(BaseChatClient):
                     **kwargs,
                 )
 
-        assert isinstance(self.raw_client, Anthropic | AnthropicVertex)
+        assert isinstance(self.raw_client, Anthropic | AnthropicVertex | AnthropicBedrock)
 
         if isinstance(tools, OpenAINotGiven):
             tools = NOT_GIVEN
@@ -600,6 +617,15 @@ class AsyncAnthropicChatClient(BaseAsyncChatClient):
                 access_token=self.creds.token,
                 http_client=self.http_client,
             )
+        elif self.endpoint.is_bedrock:
+            if self.endpoint.credentials is None:
+                raise ValueError("Anthropic Bedrock endpoint requires credentials")
+            return AsyncAnthropicBedrock(
+                aws_access_key=self.endpoint.credentials.get("aws_access_key"),
+                aws_secret_key=self.endpoint.credentials.get("aws_secret_key"),
+                aws_region=self.endpoint.region,
+                http_client=self.http_client,
+            )
         elif self.endpoint.api_schema_type == "default":
             return AsyncAnthropic(
                 api_key=self.endpoint.api_key,
@@ -767,7 +793,7 @@ class AsyncAnthropicChatClient(BaseAsyncChatClient):
                     **kwargs,
                 )
 
-        assert isinstance(self.raw_client, AsyncAnthropic | AsyncAnthropicVertex)
+        assert isinstance(self.raw_client, AsyncAnthropic | AsyncAnthropicVertex | AsyncAnthropicBedrock)
 
         if isinstance(tools, OpenAINotGiven):
             tools = NOT_GIVEN

{vectorvein-0.1.59 → vectorvein-0.1.61}/src/vectorvein/chat_clients/base_client.py

@@ -6,7 +6,14 @@ from typing import Generator, AsyncGenerator, Any, overload, Literal, Iterable
 
 import httpx
 from openai import OpenAI, AsyncOpenAI, AzureOpenAI, AsyncAzureOpenAI
-from anthropic import Anthropic, AnthropicVertex, AsyncAnthropic, AsyncAnthropicVertex
+from anthropic import (
+    Anthropic,
+    AnthropicVertex,
+    AsyncAnthropic,
+    AsyncAnthropicVertex,
+    AnthropicBedrock,
+    AsyncAnthropicBedrock,
+)
 
 from ..settings import settings
 from ..types import defaults as defs
@@ -57,7 +64,9 @@ class BaseChatClient(ABC):
 
     @cached_property
     @abstractmethod
-    def raw_client(self) -> OpenAI | AzureOpenAI | Anthropic | AnthropicVertex | httpx.Client | None:
+    def raw_client(
+        self,
+    ) -> OpenAI | AzureOpenAI | Anthropic | AnthropicVertex | AnthropicBedrock | httpx.Client | None:
         pass
 
     @overload
@@ -199,7 +208,15 @@ class BaseAsyncChatClient(ABC):
     @abstractmethod
     def raw_client(
         self,
-    ) -> AsyncOpenAI | AsyncAzureOpenAI | AsyncAnthropic | AsyncAnthropicVertex | httpx.AsyncClient | None:
+    ) -> (
+        AsyncOpenAI
+        | AsyncAzureOpenAI
+        | AsyncAnthropic
+        | AsyncAnthropicVertex
+        | AsyncAnthropicBedrock
+        | httpx.AsyncClient
+        | None
+    ):
         pass
 
     @overload

{vectorvein-0.1.59 → vectorvein-0.1.61}/src/vectorvein/types/defaults.py

@@ -231,6 +231,20 @@ QWEN_MODELS: Final[Dict[str, Dict[str, Any]]] = {
         "function_call_available": False,
         "response_format_available": True,
     },
+    "qwen2.5-coder-32b-instruct": {
+        "id": "qwen2.5-coder-32b-instruct",
+        "context_length": 30000,
+        "max_output_tokens": 4096,
+        "function_call_available": False,
+        "response_format_available": False,
+    },
+    "qwq-32b-preview": {
+        "id": "qwq-32b-preview",
+        "context_length": 30000,
+        "max_output_tokens": 4096,
+        "function_call_available": False,
+        "response_format_available": False,
+    },
     "qwen2.5-72b-instruct": {
         "id": "qwen2.5-72b-instruct",
         "context_length": 131072,
@@ -238,6 +252,14 @@ QWEN_MODELS: Final[Dict[str, Dict[str, Any]]] = {
         "function_call_available": False,
         "response_format_available": True,
     },
+    "qwen2-vl-72b-instruct": {
+        "id": "qwen2-vl-72b-instruct",
+        "context_length": 131072,
+        "max_output_tokens": 8192,
+        "function_call_available": False,
+        "response_format_available": False,
+        "native_multimodal": True,
+    },
     "qwen-max": {
         "id": "qwen-max",
         "context_length": 8096,

{vectorvein-0.1.59 → vectorvein-0.1.61}/src/vectorvein/types/llm_parameters.py

@@ -37,6 +37,7 @@ class EndpointSetting(BaseModel):
     credentials: Optional[dict] = Field(None, description="Additional credentials if needed.")
     is_azure: bool = Field(False, description="Indicates if the endpoint is for Azure.")
     is_vertex: bool = Field(False, description="Indicates if the endpoint is for Vertex.")
+    is_bedrock: bool = Field(False, description="Indicates if the endpoint is for Bedrock.")
     rpm: int = Field(description="Requests per minute.", default=defs.ENDPOINT_RPM)
     tpm: int = Field(description="Tokens per minute.", default=defs.ENDPOINT_TPM)
     concurrent_requests: int = Field(