[FEAT] Add support for OllamaEmbeddings in Chonkie

[chenzf11]

I need to call the local ollama model, and that's how I'm using chonkie now

from chonkie import SemanticChunker
from chonkie import OpenAIEmbeddings
embeddings = OpenAIEmbeddings(
    url="http://x.x.x.x:11434/v1",
    model="bge-m3:latest",
    api_key="ollama",
)

# Basic initialization with default parameters
chunker = SemanticChunker(
    embedding_model=embeddings,                     # Default model
    threshold=0.5,                               # Similarity threshold (0-1) or (1-100) or "auto"
    chunk_size=500,                              # Maximum tokens per chunk
    min_sentences=1                              # Initial sentences per chunk
)

and in thechonkie/src/chonkie/embeddings/openai.py, add url parameter to use

self.client = OpenAI(
            base_url=url,
            api_key=api_key or os.getenv("OPENAI_API_KEY"),
            organization=organization,
            timeout=timeout,
            max_retries=max_retries,
        )

[bhavnicksm]

Hey @chenzf11! 😄

Thanks for opening a feature request! We have actually thought about this before and added it to our ROADMAP (#123).

Happy to accept PRs for this if you're working on it! Though we would probably try to support it as it's own independent Embeddings implementation inheriting from the BaseEmbeddings class.

Thanks! 😊