Chunx

Chunx is an Elixir library for splitting text into meaningful chunks using various strategies. It's particularly useful for processing large texts for LLMs, semantic search, and other NLP tasks.

Credit

This library is based on chonkie-ai/chonkie

Features

• Multiple chunking strategies:

  • Token-based chunking
  • Word-based chunking
  • Sentence-based chunking
  • Semantic chunking with embeddings

• Configurable options for each strategy

• Support for overlapping chunks

• Token count tracking

• Embedding support

Installation

Add chunx to your list of dependencies in mix.exs:

def deps do
  [
    {:chunx, github: "preciz/chunx"}
  ]
end

Usage

Token-based Chunking

{:ok, tokenizer} = Tokenizers.Tokenizer.from_pretrained("gpt2")
{:ok, chunks} = Chunx.Chunker.Token.chunk("Your text here", tokenizer, chunk_size: 512)

Word-based Chunking

{:ok, tokenizer} = Tokenizers.Tokenizer.from_pretrained("gpt2")
{:ok, chunks} = Chunx.Chunker.Word.chunk("Your text here", tokenizer, chunk_size: 512)

Sentence-based Chunking

{:ok, tokenizer} = Tokenizers.Tokenizer.from_pretrained("gpt2")
{:ok, chunks} = Chunx.Chunker.Sentence.chunk("Your text here", tokenizer)

Semantic Chunking

{:ok, tokenizer} = Tokenizers.Tokenizer.from_pretrained("gpt2")

# The embedding function must return a list of Nx.Tensor.t()
embedding_fn = fn texts ->
  # Your embedding function here
end

{:ok, chunks} = Chunx.Chunker.Semantic.chunk("Your text here", tokenizer, embedding_fn)

Configuration

Each chunking strategy accepts various options to customize the chunking behavior:

• chunk_size: Maximum number of tokens per chunk
• chunk_overlap: Number of tokens or percentage to overlap between chunks
• min_sentences: Minimum number of sentences per chunk (for sentence-based)
• threshold: Similarity threshold for semantic chunking
• And more...

See the documentation for each chunker module for detailed configuration options.

Testing

# Run the test suite
mix test

License

MIT License