Address PR #1938 review comments: migrate to google-genai SDK, add Python 3.14, simplify to Google AI only

- Migrate from google-generativeai to google-genai SDK (>=0.2,<1)
- Add Python 3.14 classifier support
- Simplify to Google AI only (remove VertexAI settings and tests)
- Remove google-cloud-aiplatform dependency
- Update README to reflect Google AI focus and google-genai SDK
- Add comprehensive GAP_ANALYSIS.md with 4-phase implementation plan
- Add PR1938_CHECKLIST.md tracking review feedback and next steps
- Verify implementation against official Google migration guide
- All tests passing (5 tests, 100% coverage)

Addresses feedback from @eavanvalkenburg:
1. Added Python 3.14 support
2. Switched to google-genai (GA release, Google-recommended)
3. Removed VertexAI to simplify first iteration

Author: Jeyaram Jeyaraj

python/packages/google/README.md

@@ -1,34 +1,33 @@
 # Get Started with Microsoft Agent Framework Google
 
-> **Note**: This package is currently under active development. The chat client implementations for Google AI and Vertex AI are coming soon. This initial release provides the foundational settings and configuration classes.
+> **Note**: This package is currently under active development. The chat client implementation for Google AI is coming soon. This initial release provides the foundational settings and configuration classes.
 
 Please install this package via pip:
 
 ```bash
 pip install agent-framework-google --pre
 ```
 
-## Google AI & Vertex AI Integration
+## Google AI (Gemini API) Integration
 
-This package will provide integration with Google's generative AI platforms:
+This package provides integration with Google's Gemini API for Agent Framework:
 
 - **Google AI (Gemini API)**: Direct access to Google's Gemini models with API key authentication
-- **Vertex AI**: Enterprise-grade access via Google Cloud Platform with advanced features like grounding and code execution
+
+> **Note**: This package uses the new `google-genai` SDK as recommended by Google. See the [migration guide](https://ai.google.dev/gemini-api/docs/migrate) for more information.
 
 ### Current Status
 
 **Available Now:**
 - `GoogleAISettings`: Configuration class for Google AI (Gemini API) authentication and settings
-- `VertexAISettings`: Configuration class for Vertex AI authentication and settings
 
 **Coming Soon:**
 - `GoogleAIChatClient`: Chat client for Google AI with streaming, function calling, and multi-modal support
-- `VertexAIChatClient`: Enterprise chat client with grounding (Google Search) and code execution capabilities
 - Integration tests and usage samples
 
 ### Configuration
 
-You can configure the settings classes now, which will be used by the chat clients in the next release:
+You can configure the settings class now, which will be used by the chat client in the next release:
 
 #### Google AI Settings
 
@@ -48,34 +47,9 @@ settings = GoogleAISettings(
 )
 ```
 
-#### Vertex AI Settings
-
-```python
-from agent_framework_google import VertexAISettings
-
-# Configure via environment variables
-# VERTEX_AI_PROJECT_ID=your-project-id
-# VERTEX_AI_LOCATION=us-central1
-# VERTEX_AI_CHAT_MODEL_ID=gemini-1.5-pro
-# GOOGLE_APPLICATION_CREDENTIALS=/path/to/credentials.json
-
-settings = VertexAISettings()
-
-# Or pass parameters directly
-settings = VertexAISettings(
-    project_id="your-project-id",
-    location="us-central1",
-    chat_model_id="gemini-1.5-pro"
-)
-```
-
 ### Future Usage (Coming Soon)
 
-Once the chat clients are released, usage will look like this:
-
-#### Google AI (Gemini API)
-
-Use Google AI for straightforward access to Gemini models with API key authentication.
+Once the chat client is released, usage will look like this:
 
 ```python
 from agent_framework.google import GoogleAIChatClient
@@ -94,31 +68,6 @@ response = await agent.run("Hello!")
 print(response.text)
 ```
 
-#### Vertex AI (Coming Soon)
-
-Use Vertex AI for enterprise features including grounding with Google Search and code execution.
-
-```python
-from agent_framework.google import VertexAIChatClient
-from agent_framework import HostedWebSearchTool
-
-# Configure via environment variables
-# VERTEX_AI_PROJECT_ID=your-project-id
-# VERTEX_AI_LOCATION=us-central1
-# VERTEX_AI_CHAT_MODEL_ID=gemini-1.5-pro
-# GOOGLE_APPLICATION_CREDENTIALS=/path/to/credentials.json
-
-client = VertexAIChatClient()
-agent = client.create_agent(
-    name="Assistant",
-    instructions="You are a helpful assistant",
-    tools=[HostedWebSearchTool()]  # Vertex AI exclusive
-)
-
-response = await agent.run("What's the latest news?")
-print(response.text)
-```
-
 ## Configuration
 
 ### Environment Variables
@@ -127,12 +76,6 @@ print(response.text)
 - `GOOGLE_AI_API_KEY`: Your Google AI API key ([Get one here](https://ai.google.dev/))
 - `GOOGLE_AI_CHAT_MODEL_ID`: Model to use (e.g., `gemini-1.5-pro`, `gemini-1.5-flash`)
 
-**Vertex AI:**
-- `VERTEX_AI_PROJECT_ID`: Your GCP project ID
-- `VERTEX_AI_LOCATION`: GCP region (e.g., `us-central1`)
-- `VERTEX_AI_CHAT_MODEL_ID`: Model to use (e.g., `gemini-1.5-pro`)
-- `GOOGLE_APPLICATION_CREDENTIALS`: Path to service account JSON file
-
 ### Supported Models
 
 - `gemini-1.5-pro`: Most capable model
@@ -141,39 +84,32 @@ print(response.text)
 
 ## Features
 
-### Common Features (Both Google AI & Vertex AI)
+### Planned Features
 - ✅ Chat completion (streaming and non-streaming)
 - ✅ Function/tool calling
 - ✅ Multi-modal support (text, images, video, audio)
 - ✅ System instructions
 - ✅ Conversation history management
 
-### Vertex AI Exclusive Features (Coming Soon)
-- Grounding with Google Search
-- Grounding with Vertex AI Search
-- Code execution tool
-- Enterprise security and compliance
-- VPC-SC support
-
 ## Development Roadmap
 
 This package is being developed incrementally:
 
 - ✅ **Phase 1 (Current)**: Package structure and settings classes
 - 🚧 **Phase 2 (Next)**: Google AI chat client with streaming and function calling
 - 🚧 **Phase 3**: Google AI integration tests and samples
-- 🚧 **Phase 4**: Vertex AI chat client with enterprise features
-- 🚧 **Phase 5**: Vertex AI integration tests and samples
-- 🚧 **Phase 6**: Advanced features (context caching, safety settings, etc.)
+- 🚧 **Phase 4**: Advanced features (context caching, safety settings, structured output)
+
+> **Note**: Vertex AI support may be added in a future iteration based on user demand.
 
 ## Examples
 
-Examples will be available once the chat clients are implemented. Check back soon or watch the [repository](https://github.com/microsoft/agent-framework) for updates.
+Examples will be available once the chat client is implemented. Check back soon or watch the [repository](https://github.com/microsoft/agent-framework) for updates.
 
 ## Documentation
 
 For more information:
 - [Google AI Documentation](https://ai.google.dev/docs)
-- [Vertex AI Documentation](https://cloud.google.com/vertex-ai/docs)
+- [Google Gemini API Migration Guide](https://ai.google.dev/gemini-api/docs/migrate)
 - [Agent Framework Documentation](https://aka.ms/agent-framework)
 - [Agent Framework Repository](https://github.com/microsoft/agent-framework)

python/packages/google/agent_framework_google/__init__.py

@@ -2,16 +2,15 @@
 
 import importlib.metadata
 
-# NOTE: Client classes will be imported here in PR #2 and PR #4
-# from ._chat_client import GoogleAIChatClient, VertexAIChatClient
+# NOTE: Client class will be imported here in a future PR
+# from ._chat_client import GoogleAIChatClient
 
 try:
     __version__ = importlib.metadata.version(__name__)
 except importlib.metadata.PackageNotFoundError:
     __version__ = "0.0.0"  # Fallback for development mode
 
 __all__ = [
-    # "GoogleAIChatClient",  # Will be added in PR #2
-    # "VertexAIChatClient",  # Will be added in PR #4
+    # "GoogleAIChatClient",  # Will be added in a future PR
     "__version__",
 ]

python/packages/google/agent_framework_google/_chat_client.py

@@ -47,53 +47,5 @@ class GoogleAISettings(AFBaseSettings):
     chat_model_id: str | None = None
 
 
-class VertexAISettings(AFBaseSettings):
-    """Vertex AI settings for Google Cloud access.
-
-    The settings are first loaded from environment variables with the prefix 'VERTEX_AI_'.
-    If the environment variables are not found, the settings can be loaded from a .env file
-    with the encoding 'utf-8'. If the settings are not found in the .env file, the settings
-    are ignored; however, validation will fail alerting that the settings are missing.
-
-    Keyword Args:
-        project_id: The Google Cloud project ID.
-        location: The Google Cloud region (e.g., us-central1).
-        chat_model_id: The Vertex AI chat model ID (e.g., gemini-1.5-pro).
-        credentials_path: Optional path to service account JSON file.
-        env_file_path: If provided, the .env settings are read from this file path location.
-        env_file_encoding: The encoding of the .env file, defaults to 'utf-8'.
-
-    Examples:
-        .. code-block:: python
-
-            from agent_framework.google import VertexAISettings
-
-            # Using environment variables
-            # Set VERTEX_AI_PROJECT_ID=your-project-id
-            # VERTEX_AI_LOCATION=us-central1
-            # VERTEX_AI_CHAT_MODEL_ID=gemini-1.5-pro
-            # GOOGLE_APPLICATION_CREDENTIALS=/path/to/credentials.json
-
-            settings = VertexAISettings()
-
-            # Or passing parameters directly
-            settings = VertexAISettings(
-                project_id="your-project-id",
-                location="us-central1",
-                chat_model_id="gemini-1.5-pro"
-            )
-
-            # Or loading from a .env file
-            settings = VertexAISettings(env_file_path="path/to/.env")
-    """
-
-    env_prefix: ClassVar[str] = "VERTEX_AI_"
-
-    project_id: str | None = None
-    location: str | None = None
-    chat_model_id: str | None = None
-    credentials_path: str | None = None
-
-
-# NOTE: Client implementations will be added in PR #2 and PR #4
-# For PR #1, we're only setting up the package structure and settings
+# NOTE: Client implementations will be added in a future PR
+# For now, we're only setting up the package structure and settings

python/packages/google/pyproject.toml

@@ -1,6 +1,6 @@
 [project]
 name = "agent-framework-google"
-description = "Google AI and Vertex AI integration for Microsoft Agent Framework."
+description = "Google AI (Gemini API) integration for Microsoft Agent Framework."
 authors = [{ name = "Microsoft", email = "[email protected]"}]
 readme = "README.md"
 requires-python = ">=3.10"
@@ -19,12 +19,12 @@ classifiers = [
   "Programming Language :: Python :: 3.11",
   "Programming Language :: Python :: 3.12",
   "Programming Language :: Python :: 3.13",
+  "Programming Language :: Python :: 3.14",
   "Typing :: Typed",
 ]
 dependencies = [
     "agent-framework-core",
-    "google-generativeai>=0.8,<1",
-    "google-cloud-aiplatform>=1.60,<2",
+    "google-genai>=0.2,<1",
 ]
 
 [tool.uv]

python/packages/google/tests/conftest.py

@@ -39,29 +39,3 @@ def google_ai_unit_test_env(monkeypatch, exclude_list, override_env_param_dict):
         monkeypatch.setenv(key, value)  # type: ignore
 
     return env_vars
-
-
-@fixture
-def vertex_ai_unit_test_env(monkeypatch, exclude_list, override_env_param_dict):  # type: ignore
-    """Fixture to set environment variables for VertexAISettings."""
-    if exclude_list is None:
-        exclude_list = []
-
-    if override_env_param_dict is None:
-        override_env_param_dict = {}
-
-    env_vars = {
-        "VERTEX_AI_PROJECT_ID": "test-project-id",
-        "VERTEX_AI_LOCATION": "us-central1",
-        "VERTEX_AI_CHAT_MODEL_ID": "gemini-1.5-pro",
-    }
-
-    env_vars.update(override_env_param_dict)  # type: ignore
-
-    for key, value in env_vars.items():
-        if key in exclude_list:
-            monkeypatch.delenv(key, raising=False)  # type: ignore
-            continue
-        monkeypatch.setenv(key, value)  # type: ignore
-
-    return env_vars

python/packages/google/tests/test_settings.py

@@ -1,7 +1,7 @@
 # Copyright (c) Microsoft. All rights reserved.
 import pytest
 
-from agent_framework_google._chat_client import GoogleAISettings, VertexAISettings
+from agent_framework_google._chat_client import GoogleAISettings
 
 
 # region GoogleAISettings Tests
@@ -55,81 +55,3 @@ def test_google_ai_settings_override_env() -> None:
 
 
 # endregion
-
-
-# region VertexAISettings Tests
-
-
-def test_vertex_ai_settings_from_env(vertex_ai_unit_test_env: dict[str, str]) -> None:
-    """Test VertexAISettings initialization from environment variables."""
-    settings = VertexAISettings()
-    assert settings.project_id == vertex_ai_unit_test_env["VERTEX_AI_PROJECT_ID"]
-    assert settings.location == vertex_ai_unit_test_env["VERTEX_AI_LOCATION"]
-    assert settings.chat_model_id == vertex_ai_unit_test_env["VERTEX_AI_CHAT_MODEL_ID"]
-
-
-def test_vertex_ai_settings_from_params() -> None:
-    """Test VertexAISettings initialization from parameters."""
-    settings = VertexAISettings(
-        project_id="my-project",
-        location="europe-west4",
-        chat_model_id="gemini-1.5-pro",
-        credentials_path="/path/to/creds.json",
-    )
-    assert settings.project_id == "my-project"
-    assert settings.location == "europe-west4"
-    assert settings.chat_model_id == "gemini-1.5-pro"
-    assert settings.credentials_path == "/path/to/creds.json"
-
-
-@pytest.mark.parametrize("exclude_list", [["VERTEX_AI_PROJECT_ID"]], indirect=True)
-def test_vertex_ai_settings_missing_project_id(vertex_ai_unit_test_env: dict[str, str]) -> None:
-    """Test VertexAISettings when project ID is missing."""
-    settings = VertexAISettings()
-    assert settings.project_id is None
-    assert settings.location == vertex_ai_unit_test_env["VERTEX_AI_LOCATION"]
-    assert settings.chat_model_id == vertex_ai_unit_test_env["VERTEX_AI_CHAT_MODEL_ID"]
-
-
-@pytest.mark.parametrize("exclude_list", [["VERTEX_AI_LOCATION"]], indirect=True)
-def test_vertex_ai_settings_missing_location(vertex_ai_unit_test_env: dict[str, str]) -> None:
-    """Test VertexAISettings when location is missing."""
-    settings = VertexAISettings()
-    assert settings.project_id == vertex_ai_unit_test_env["VERTEX_AI_PROJECT_ID"]
-    assert settings.location is None
-    assert settings.chat_model_id == vertex_ai_unit_test_env["VERTEX_AI_CHAT_MODEL_ID"]
-
-
-@pytest.mark.parametrize("exclude_list", [["VERTEX_AI_CHAT_MODEL_ID"]], indirect=True)
-def test_vertex_ai_settings_missing_model_id(vertex_ai_unit_test_env: dict[str, str]) -> None:
-    """Test VertexAISettings when model ID is missing."""
-    settings = VertexAISettings()
-    assert settings.project_id == vertex_ai_unit_test_env["VERTEX_AI_PROJECT_ID"]
-    assert settings.location == vertex_ai_unit_test_env["VERTEX_AI_LOCATION"]
-    assert settings.chat_model_id is None
-
-
-def test_vertex_ai_settings_override_env() -> None:
-    """Test VertexAISettings parameter override of environment variables."""
-    settings = VertexAISettings(
-        project_id="override-project",
-        location="asia-southeast1",
-        chat_model_id="gemini-2.0-flash",
-    )
-    assert settings.project_id == "override-project"
-    assert settings.location == "asia-southeast1"
-    assert settings.chat_model_id == "gemini-2.0-flash"
-
-
-def test_vertex_ai_settings_credentials_path() -> None:
-    """Test VertexAISettings with credentials path."""
-    settings = VertexAISettings(
-        project_id="test-project",
-        location="us-central1",
-        chat_model_id="gemini-1.5-pro",
-        credentials_path="/custom/path/creds.json",
-    )
-    assert settings.credentials_path == "/custom/path/creds.json"
-
-
-# endregion