1. Learn
    2. /
  2. Courses
    3. /
  3. Developing LLM Applications with LangChain
Connected

Exercise

Splitting by character

A key process in implementing Retrieval Augmented Generation (RAG) is splitting documents into chunks for storage in a vector database.

There are several splitting strategies available in LangChain, some with more complex routines than others. In this exercise, you'll implement a character text splitter, which splits documents based on characters and measures the chunk length by the number of characters.

Remember that there is no ideal splitting strategy, you may need to experiment with a few to find the right one for your use case.

Instructions

100 XP
  • Import the CharacterTextSplitter class from langchain_text_splitters.
  • Create a CharacterTextSplitter instance with separator="\n", chunk_size=24, and chunk_overlap=10.
  • Use the .split_text() method to split the quote and print the chunks and chunk lengths.