Feat/embeddings

[Verdenroz]

Summary

• Adds embedding support via a new EmbeddingAdapter protocol and Chimeric.embed() / Chimeric.aembed() methods
• Implements embeddings for OpenAI, Google (Gemini), and Cohere — single string and batch (list[str]) inputs both supported
• Introduces EmbeddingResponse and EmbeddingUsage types for a unified response shape across providers
• Refactors Chimeric.__init__() to accept explicit timeout, max_retries, and default_headers parameters instead of **kwargs, improving discoverability and IDE support

Breaking Changes

• Chimeric() no longer accepts arbitrary **kwargs — timeout, max_retries, and default_headers must now be passed by name

Test Plan

• Unit tests for OpenAIAdapter embedding request building and response parsing
• Integration tests (cassette-recorded) for OpenAI, Google, and Cohere: single embed, batch embed, async variants, cosine similarity, and dimensions kwarg
• test_doc_examples.py and test_chimeric.py updated to cover new Chimeric() constructor signature

Checklist

• Tests pass (nox -s tests)
• Coverage ≥ 80%
• Docs updated (docs/usage/embeddings.md added, docs/usage/responses.md and docs/usage/configuration.md updated)
• No hardcoded secrets or credentials

[codecov]

Codecov Report

❌ Patch coverage is 72.27723% with 28 lines in your changes missing coverage. Please review.

Files with missing lines	Patch %	Lines
src/chimeric/adapters/google.py	25.00%	12 Missing ⚠️
src/chimeric/adapters/cohere.py	25.00%	9 Missing ⚠️
src/chimeric/http.py	87.80%	3 Missing and 2 partials ⚠️
src/chimeric/adapters/base.py	50.00%	1 Missing ⚠️
src/chimeric/chimeric.py	87.50%	1 Missing ⚠️

📢 Thoughts on this report? Let us know!