fix: improve import error messages for LLM client dependencies (#4605)

python/packages/autogen-ext/src/autogen_ext/models/anthropic/__init__.py

@@ -1,16 +1,24 @@
-from ._anthropic_client import (
-    AnthropicBedrockChatCompletionClient,
-    AnthropicChatCompletionClient,
-    BaseAnthropicChatCompletionClient,
-)
-from .config import (
-    AnthropicBedrockClientConfiguration,
-    AnthropicBedrockClientConfigurationConfigModel,
-    AnthropicClientConfiguration,
-    AnthropicClientConfigurationConfigModel,
-    BedrockInfo,
-    CreateArgumentsConfigModel,
-)
+try:
+    from ._anthropic_client import (
+        AnthropicBedrockChatCompletionClient,
+        AnthropicChatCompletionClient,
+        BaseAnthropicChatCompletionClient,
+    )
+    from .config import (
+        AnthropicBedrockClientConfiguration,
+        AnthropicBedrockClientConfigurationConfigModel,
+        AnthropicClientConfiguration,
+        AnthropicClientConfigurationConfigModel,
+        BedrockInfo,
+        CreateArgumentsConfigModel,
+    )
+except ImportError as e:
+    raise ImportError(
+        "Dependencies for Anthropic client not found. "
+        "Please install the Anthropic package: "
+        "pip install autogen-ext[anthropic]\n"
+        f"Original error: {e}"
+    ) from e
 
 __all__ = [
     "AnthropicChatCompletionClient",

python/packages/autogen-ext/src/autogen_ext/models/azure/__init__.py

@@ -1,4 +1,12 @@
-from ._azure_ai_client import AzureAIChatCompletionClient
-from .config import AzureAIChatCompletionClientConfig
+try:
+    from ._azure_ai_client import AzureAIChatCompletionClient
+    from .config import AzureAIChatCompletionClientConfig
+except ImportError as e:
+    raise ImportError(
+        "Dependencies for Azure AI client not found. "
+        "Please install the Azure AI Inference package: "
+        "pip install autogen-ext[azure]\n"
+        f"Original error: {e}"
+    ) from e
 
 __all__ = ["AzureAIChatCompletionClient", "AzureAIChatCompletionClientConfig"]

python/packages/autogen-ext/src/autogen_ext/models/ollama/__init__.py

@@ -1,8 +1,16 @@
-from ._ollama_client import OllamaChatCompletionClient
-from .config import (
-    BaseOllamaClientConfigurationConfigModel,
-    CreateArgumentsConfigModel,
-)
+try:
+    from ._ollama_client import OllamaChatCompletionClient
+    from .config import (
+        BaseOllamaClientConfigurationConfigModel,
+        CreateArgumentsConfigModel,
+    )
+except ImportError as e:
+    raise ImportError(
+        "Dependencies for Ollama client not found. "
+        "Please install the Ollama package: "
+        "pip install autogen-ext[ollama]\n"
+        f"Original error: {e}"
+    ) from e
 
 __all__ = [
     "OllamaChatCompletionClient",

python/packages/autogen-ext/src/autogen_ext/models/openai/__init__.py

@@ -1,16 +1,24 @@
-from . import _message_transform
-from ._openai_client import (
-    AZURE_OPENAI_USER_AGENT,
-    AzureOpenAIChatCompletionClient,
-    BaseOpenAIChatCompletionClient,
-    OpenAIChatCompletionClient,
-)
-from .config import (
-    AzureOpenAIClientConfigurationConfigModel,
-    BaseOpenAIClientConfigurationConfigModel,
-    CreateArgumentsConfigModel,
-    OpenAIClientConfigurationConfigModel,
-)
+try:
+    from . import _message_transform
+    from ._openai_client import (
+        AZURE_OPENAI_USER_AGENT,
+        AzureOpenAIChatCompletionClient,
+        BaseOpenAIChatCompletionClient,
+        OpenAIChatCompletionClient,
+    )
+    from .config import (
+        AzureOpenAIClientConfigurationConfigModel,
+        BaseOpenAIClientConfigurationConfigModel,
+        CreateArgumentsConfigModel,
+        OpenAIClientConfigurationConfigModel,
+    )
+except ImportError as e:
+    raise ImportError(
+        "Dependencies for OpenAI client not found. "
+        "Please install the OpenAI package: "
+        "pip install autogen-ext[openai]\n"
+        f"Original error: {e}"
+    ) from e
 
 __all__ = [
     "OpenAIChatCompletionClient",