unique_toolkit 0.8.20__tar.gz → 0.8.22__tar.gz

{unique_toolkit-0.8.20 → unique_toolkit-0.8.22}/CHANGELOG.md

@@ -5,9 +5,19 @@ All notable changes to this project will be documented in this file.
 The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/), 
 and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 
+
+## [0.8.22] - 2025-08-25
+- Add DeepSeek-R1, DeepSeek-V3.1, Qwen3-235B-A22B and Qwen3-235B-A22B-Thinking-2507 to supported model list
+
+## [0.8.21] - 2025-08-26
+- Fixed old (not used) function "create_async" in language_model.functions : The function always returns "Unauthorized" --> Added "user_id" argument to fix this.
+
 ## [0.8.20] - 2025-08-24
 - Fixed forced-tool-calls
 
+## [0.8.20] - 2025-08-05
+- Bump SDK version to support the latest features.
+
 ## [0.8.19] - 2025-08-24
 - Enforce usage of ruff using pipeline
 

{unique_toolkit-0.8.20 → unique_toolkit-0.8.22}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.1
 Name: unique_toolkit
-Version: 0.8.20
+Version: 0.8.22
 Summary: 
 License: Proprietary
 Author: Martin Fadler
@@ -114,9 +114,19 @@ All notable changes to this project will be documented in this file.
 The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.0.0/), 
 and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 
+
+## [0.8.22] - 2025-08-25
+- Add DeepSeek-R1, DeepSeek-V3.1, Qwen3-235B-A22B and Qwen3-235B-A22B-Thinking-2507 to supported model list
+
+## [0.8.21] - 2025-08-26
+- Fixed old (not used) function "create_async" in language_model.functions : The function always returns "Unauthorized" --> Added "user_id" argument to fix this.
+
 ## [0.8.20] - 2025-08-24
 - Fixed forced-tool-calls
 
+## [0.8.20] - 2025-08-05
+- Bump SDK version to support the latest features.
+
 ## [0.8.19] - 2025-08-24
 - Enforce usage of ruff using pipeline
 

{unique_toolkit-0.8.20 → unique_toolkit-0.8.22}/pyproject.toml

@@ -1,6 +1,6 @@
 [tool.poetry]
 name = "unique_toolkit"
-version = "0.8.20"
+version = "0.8.22"
 description = ""
 authors = [
     "Martin Fadler <martin.fadler@unique.ch>",

{unique_toolkit-0.8.20 → unique_toolkit-0.8.22}/unique_toolkit/language_model/functions.py

@@ -94,6 +94,7 @@ def complete(
 
 async def complete_async(
     company_id: str,
+    user_id: str,
     messages: LanguageModelMessages | list[ChatCompletionMessageParam],
     model_name: LanguageModelName | str,
     temperature: float = DEFAULT_COMPLETE_TEMPERATURE,
@@ -142,6 +143,7 @@ async def complete_async(
     try:
         response = await unique_sdk.ChatCompletion.create_async(
             company_id=company_id,
+            user_id=user_id,
             model=model,
             messages=cast(
                 "list[unique_sdk.Integrated.ChatCompletionRequestMessage]",
@@ -419,6 +421,7 @@ def complete_with_references(
 
 async def complete_with_references_async(
     company_id: str,
+    user_id: str,
     messages: LanguageModelMessages,
     model_name: LanguageModelName | str,
     content_chunks: list[ContentChunk] | None = None,
@@ -432,6 +435,7 @@ async def complete_with_references_async(
     # Use toolkit language model functions for chat completion
     response = await complete_async(
         company_id=company_id,
+        user_id=user_id,
         model_name=model_name,
         messages=messages,
         temperature=temperature,

{unique_toolkit-0.8.20 → unique_toolkit-0.8.22}/unique_toolkit/language_model/infos.py

@@ -48,6 +48,10 @@ class LanguageModelName(StrEnum):
     LITELLM_OPENAI_GPT_5_MINI = "litellm:openai-gpt-5-mini"
     LITELLM_OPENAI_GPT_5_NANO = "litellm:openai-gpt-5-nano"
     LITELLM_OPENAI_GPT_5_CHAT = "litellm:openai-gpt-5-chat"
+    LITELLM_DEEPSEEK_R1 = "litellm:deepseek-r1"
+    LITELLM_DEEPSEEK_V3 = "litellm:deepseek-v3-1"
+    LITELLM_QWEN_3 = "litellm:qwen-3-235B-A22B"
+    LITELLM_QWEN_3_THINKING = "litellm:qwen-3-235B-A22B-thinking"
 
 
 class EncoderName(StrEnum):
@@ -875,6 +879,69 @@ class LanguageModelInfo(BaseModel):
                     deprecated_at=date(2026, 8, 7),
                     retirement_at=date(2026, 8, 7),
                 )
+            case LanguageModelName.LITELLM_DEEPSEEK_R1:
+                return cls(
+                    name=model_name,
+                    provider=LanguageModelProvider.LITELLM,
+                    version="deepseek-r1",
+                    capabilities=[
+                        ModelCapabilities.FUNCTION_CALLING,
+                        ModelCapabilities.STRUCTURED_OUTPUT,
+                        ModelCapabilities.STREAMING,
+                        ModelCapabilities.REASONING,
+                    ],
+                    token_limits=LanguageModelTokenLimits(
+                        token_limit_input=64_000, token_limit_output=4_000
+                    ),
+                    published_at=date(2025, 1, 25),
+                )
+            case LanguageModelName.LITELLM_DEEPSEEK_V3:
+                return cls(
+                    name=model_name,
+                    provider=LanguageModelProvider.LITELLM,
+                    version="deepseek-v3-1",
+                    capabilities=[
+                        ModelCapabilities.FUNCTION_CALLING,
+                        ModelCapabilities.STRUCTURED_OUTPUT,
+                        ModelCapabilities.REASONING,
+                    ],
+                    token_limits=LanguageModelTokenLimits(
+                        token_limit_input=128_000, token_limit_output=4_000
+                    ),
+                    published_at=date(2025, 8, 1),
+                )
+            case LanguageModelName.LITELLM_QWEN_3:
+                return cls(
+                    name=model_name,
+                    provider=LanguageModelProvider.LITELLM,
+                    version="qwen-3",
+                    capabilities=[
+                        ModelCapabilities.FUNCTION_CALLING,
+                        ModelCapabilities.STRUCTURED_OUTPUT,
+                        ModelCapabilities.STREAMING,
+                        ModelCapabilities.REASONING,
+                    ],
+                    published_at=date(2025, 4, 29),
+                    token_limits=LanguageModelTokenLimits(
+                        token_limit_input=256_000, token_limit_output=32_768
+                    ),
+                )
+            case LanguageModelName.LITELLM_QWEN_3_THINKING:
+                return cls(
+                    name=model_name,
+                    provider=LanguageModelProvider.LITELLM,
+                    version="qwen-3-thinking",
+                    capabilities=[
+                        ModelCapabilities.FUNCTION_CALLING,
+                        ModelCapabilities.STRUCTURED_OUTPUT,
+                        ModelCapabilities.STREAMING,
+                        ModelCapabilities.REASONING,
+                    ],
+                    token_limits=LanguageModelTokenLimits(
+                        token_limit_input=256_000, token_limit_output=32_768
+                    ),
+                    published_at=date(2025, 7, 25),
+                )
 
             case _:
                 if isinstance(model_name, LanguageModelName):

{unique_toolkit-0.8.20 → unique_toolkit-0.8.22}/unique_toolkit/language_model/service.py

@@ -261,6 +261,7 @@ class LanguageModelService:
 
         return await complete_async(
             company_id=self._company_id,
+            user_id=self._user_id,
             messages=messages,
             model_name=model_name,
             temperature=temperature,
@@ -276,6 +277,7 @@ class LanguageModelService:
     async def complete_async_util(
         cls,
         company_id: str,
+        user_id: str,
         messages: LanguageModelMessages,
         model_name: LanguageModelName | str,
         temperature: float = DEFAULT_COMPLETE_TEMPERATURE,
@@ -291,6 +293,7 @@ class LanguageModelService:
 
         return await complete_async(
             company_id=company_id,
+            user_id=user_id,
             messages=messages,
             model_name=model_name,
             temperature=temperature,