[ML] Add AnthropicModelHandler for direct Claude inference

[bvolpato]

Adds a ModelHandler for Anthropic Claude models using the Messages API, enabling direct inference in Beam pipelines without requiring Vertex AI or GCP setup.

This follows the same pattern as the existing GeminiModelHandler, which provides direct Google AI Studio access alongside the Vertex AI-based handler. The motivation is the same: users should be able to use Claude models with just an API key, without needing GCP project setup, service accounts, or IAM configuration.

Changes

• anthropic_inference.py - AnthropicModelHandler extending RemoteModelHandler, with message_from_string and message_from_conversation request functions, and retry logic for 429/5xx errors
• Added support for System Prompts: Set global pipeline behaviors via system= to configure the model persona or strict instruction sets.
• Added support for Structured JSON Outputs: Pass output_config= following Anthropic's GA JSON Schema format to force structured formatting out of Claude endpoints (enforces anthropic>=0.86.0).
• anthropic_inference_test.py - Unit tests covering retry logic, request functions, model handler, output formatting logic, system prompts, and a full pipeline E2E test with picklable fakes.
• anthropic_inference_it_test.py - Integration tests for generating structured outputs (e.g. Fizz rules) and multi-turn conversations (requires ANTHROPIC_API_KEY).
• anthropic_tests_requirements.txt - Testing dependency declaration enforcing anthropic>=0.86.0.
• pytest.ini - Registration of the anthropic_postcommit marker.

CI note

The integration tests are not wired into inferencePostCommitIT or any Dataflow gradle task since the OSS project will not have Anthropic API keys configured. The IT tests exist for contributors to run locally:

ANTHROPIC_API_KEY="..." python -m pytest apache_beam/ml/inference/anthropic_inference_it_test.py -v

Unit tests (mock-based, no API key needed):

python -m pytest apache_beam/ml/inference/anthropic_inference_test.py -v

---

Thank you for your contribution! Follow this checklist to help us incorporate your contribution quickly and easily:

• Mention the appropriate issue in your description (for example: addresses #123), if applicable. This will automatically add a link to the pull request in the issue. If you would like the issue to automatically close on merging the pull request, comment fixes #<ISSUE NUMBER> instead.
• Update CHANGES.md with noteworthy changes.
• If this contribution is large, please file an Apache Individual Contributor License Agreement.

See the Contributor Guide for more tips on how to make review process smoother.

To check the build health, please visit https://github.com/apache/beam/blob/master/.test-infra/BUILD_STATUS.md

GitHub Actions Tests Status (on master branch)

See CI.md for more information about GitHub Actions CI or the workflows README to see a list of phrases to trigger workflows.

[gemini-code-assist]

Summary of Changes

Hello, I'm Gemini Code Assist¹! I'm currently reviewing this pull request and will post my feedback shortly. In the meantime, here's a summary to help you and other reviewers quickly get up to speed!

This pull request introduces a new AnthropicModelHandler to Apache Beam's RunInference transform, enabling direct integration with Anthropic Claude models via their Messages API. This allows users to perform batch inference with Claude models using only an API key, bypassing the need for Vertex AI or GCP setup, similar to the existing Gemini integration. The change expands Beam's ML inference capabilities by adding support for another major LLM provider.

Highlights

• New AnthropicModelHandler: Added a dedicated AnthropicModelHandler to enable direct inference with Anthropic Claude models within Apache Beam pipelines, bypassing the need for Vertex AI or GCP setup.
• Request Functions: Implemented message_from_string and message_from_conversation functions to handle different input formats for Claude's Messages API, supporting both single-turn prompts and multi-turn conversations.
• Robust Error Handling: Included retry logic for Anthropic API errors, specifically targeting rate limiting (429) and server-side errors (5xx) to improve pipeline resilience.
• Comprehensive Testing: Developed extensive unit tests using mocks and integration tests requiring an API key, ensuring the reliability and correctness of the Anthropic integration.
• Build System Integration: Configured the build system to include Anthropic integration tests in post-commit checks, registering a new Pytest marker and a Gradle task.

🧠 New Feature in Public Preview: You can now enable Memory to help Gemini Code Assist learn from your team's feedback. This makes future code reviews more consistent and personalized to your project's style. Click here to enable Memory in your admin console.

Using Gemini Code Assist

The full guide for Gemini Code Assist can be found on our documentation page, here are some quick tips.

Invoking Gemini

You can request assistance from Gemini at any point by creating a comment using either /gemini <command> or @gemini-code-assist <command>. Below is a summary of the supported commands on the current page.

Feature	Command	Description
Code Review	/gemini review	Performs a code review for the current pull request in its current state.
Pull Request Summary	/gemini summary	Provides a summary of the current pull request in its current state.
Comment	@gemini-code-assist	Responds in comments when explicitly tagged, both in pull request comments and review comments.
Help	/gemini help	Displays a list of available commands.

Customization

To customize Gemini Code Assist for GitHub experience, repository maintainers can create a configuration file and/or provide a custom code review style guide (such as PEP-8 for Python) by creating and adding files to a .gemini/ folder in the base of the repository. Detailed instructions can be found here.

Limitations & Feedback

Gemini Code Assist may make mistakes. Please leave feedback on any instances where its feedback is incorrect or counter productive. You can react with 👍 and 👎 on @gemini-code-assist comments. If you're interested in giving your feedback about your experience with Gemini Code Assist for GitHub and other Google products, sign up here.

Footnotes

1. Review the Privacy Notices, Generative AI Prohibited Use Policy, Terms of Service, and learn how to configure Gemini Code Assist in GitHub here. Gemini can make mistakes, so double check it and use code with caution. ↩

[bvolpato]

Not too sure about this, I've based on the geminiInferenceTest, but is ANTHROPIC_API_KEY is not defined at runtime, it will be a noop.

[jrmccluskey]

You should define the task but not add it to the inferencePostCommitIT task set yet since the repo doesn't have an anthropic API key secret. The OpenAI handler is in the same boat. At some point I need to see if Apache has resources around for these

[github-actions]

Assigning reviewers:

R: @jrmccluskey for label python.

Note: If you would like to opt out of this review, comment assign to next reviewer.

Available commands:

• stop reviewer notifications - opt out of the automated review tooling
• remind me after tests pass - tag the comment author after tests pass
• waiting on author - shift the attention set back to the author (any comment or push by the author will return the attention set to the reviewers)

The PR bot will only process comments in the main thread (not review comments).