Add support to Gemini providers

v0.6.3

Add support to use Gemini vision and embedding models

Version 0.6.3: SplitterMR now supports Gemini as a backend for both embedding and vision models.

To use HuggingFace, Gemini, Claude or Grok models, you must install SplitterMR with the multimodal extra:

pip install "splitter-mr[multimodal]"

Features

• Add GeminiVisionModel class to Vision models.
• Add GeminiEmbedding class to embedding models.
• Apply lazy import strategy to classes which require extras to be installed (e.g., docling, markitdown, etc.).

Documentation

• Update documentation.