Basic multimodal embedding support

[jamesbraza]

This PR:

• Upstreams some utils from Image reader and image support in gather_evidence paper-qa#1046 to LMI
• Adds testing of VertexAI multimodal embeddings

[Copilot]

Pull Request Overview

This PR adds multimodal embedding support to LMI by upstreaming utilities from paper-qa and implementing VertexAI multimodal embedding testing. The changes include new byte/string conversion utilities, image validation functions, token estimation for multimodal content, and comprehensive test coverage for multimodal embeddings.

• Adds utility functions for handling binary data and image validation
• Implements multimodal token estimation for rate limiting
• Adds comprehensive testing for VertexAI multimodal embeddings with GCS integration

Reviewed Changes

Copilot reviewed 8 out of 10 changed files in this pull request and generated 3 comments.

Show a summary per file

File	Description
packages/lmi/src/lmi/utils.py	Added utility functions for byte/string conversion, image validation, and image encoding
packages/lmi/src/lmi/embeddings.py	Enhanced token estimation to support multimodal content and updated rate limiting
packages/lmi/tests/test_utils.py	Added comprehensive tests for byte/string conversion utilities
packages/lmi/tests/test_embeddings.py	Added multimodal embedding tests and token estimation tests
packages/lmi/tests/conftest.py	Added fixtures for PNG image testing and GCS integration
packages/lmi/pyproject.toml	Added image optional dependency with pillow requirement
pyproject.toml	Added mypy ignore for google.cloud.storage module
.github/workflows/tests.yml	Added GCP authentication for CI testing

---

_{Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.}

[jabra]

LGTM