Repository files navigation

Ever found yourself building yet another RAG bot (your 2,342,148th one), only to hit that all-too-familiar wall? You know the one —— where you're stuck choosing between:

• Library X: A behemoth that takes forever to install and probably includes three different kitchen sinks
• Library Y: So bare-bones it might as well be a "Hello World" program
• Writing it yourself? For the 2,342,149th time,sigh

And you think to yourself:

"WHY CAN'T THIS JUST BE SIMPLE?!"
"Why do I need to choose between bloated and bare-bones?"
"Why can't I just install, import, and CHONK?!"

Well, look no further than Chonkie! (chonkie boi is a gud boi 🦛💕)

🚀 Feature-rich: All the CHONKs you'd ever need
✨ Easy to use: Install, Import, CHONK
⚡ Fast: CHONK at the speed of light! zooooom
🌐 Wide support: Supports all your favorite tokenizer CHONKS
🪶 Light-weight: No bloat, just CHONK
🦛 Cute CHONK mascot: psst it's a pygmy hippo btw
❤️Moto Moto's favorite python library

Chonkie is a chunking library that "just works™".

To install chonkie, simply run:

pip install chonkie

Chonkie follows the rule to have minimal default installs, read theDOCS to know the installation for your required chunker, or simply installall if you don't want to think about it (not recommended).

pip install chonkie[all]

Here's a basic example to get you started:

# First import the chunker you want from ChonkiefromchonkieimportTokenChunker# Import your favorite tokenizer library# Also supports AutoTokenizers, TikToken and AutoTikTokenizerfromtokenizersimportTokenizertokenizer=Tokenizer.from_pretrained("gpt2")# Initialize the chunkerchunker=TokenChunker(tokenizer)# Chunk some textchunks=chunker("Woah! Chonkie, the chunking library is so cool! I love the tiny hippo hehe.")# Access chunksforchunkinchunks:print(f"Chunk:{chunk.text}")print(f"Tokens:{chunk.token_count}")

More example usages given inside theDOCS

Chonkie provides several chunkers to help you split your text efficiently for RAG applications. Here's a quick overview of the available chunkers:

• TokenChunker: Splits text into fixed-size token chunks.
• WordChunker: Splits text into chunks based on words.
• SentenceChunker: Splits text into chunks based on sentences.
• RecursiveChunker: Splits text hierarchically using customizable rules to create semantically meaningful chunks.
• SemanticChunker: Splits text into chunks based on semantic similarity.
• SDPMChunker: Splits text using a Semantic Double-Pass Merge approach.
• LateChunker (experimental): Embeds text and then splits it to have better chunk embeddings.

More on these methods and the approaches taken inside theDOCS

"I may be smol hippo, but I pack a punch!" 🦛

Here's a quick peek at how Chonkie performs:

Size📦

• Default Install: 15MB (vs 80-171MB for alternatives)
• With Semantic: Still lighter than the competition!

Speed⚡

• Token Chunking: 33x faster than the slowest alternative
• Sentence Chunking: Almost 2x faster than competitors
• Semantic Chunking: Up to 2.5x faster than others

Check out our detailedbenchmarks to see how Chonkie races past the competition! 🏃‍♂️💨

Want to help make Chonkie even better? Check out ourCONTRIBUTING.md guide! Whether you're fixing bugs, adding features, or improving docs, every contribution helps make Chonkie a better CHONK for everyone.

Remember: No contribution is too small for this tiny hippo! 🦛

Chonkie would like to CHONK its way through a special thanks to all the users and contributors who have helped make this library what it is today! Your feedback, issue reports, and improvements have helped make Chonkie the CHONKIEST it can be.

And of course, special thanks toMoto Moto for endorsing Chonkie with his famous quote:

"I like them big, I like them chonkie."~ Moto Moto

If you use Chonkie in your research, please cite it as follows:

@misc{chonkie2024,author ={Minhas, Bhavnick AND Nigam, Shreyash},title ={Chonkie: A Fast Feature-full Chunking Library for RAG Bots},year ={2024},publisher ={GitHub},howpublished ={\url{https://github.com/bhavnick/chonkie}},}