Semantic Chunking

Semantic chunking is a method of splitting documents into smaller chunks by analyzing semantic similarity between text segments using embeddings. It uses the chonkie library to identify natural breakpoints where the semantic meaning changes significantly, based on a configurable similarity threshold. This helps preserve context and meaning better than fixed-size chunking by ensuring semantically related content stays together in the same chunk, while splitting occurs at meaningful topic transitions.

from bitca.agent import Agent
from bitca.document.chunking.semantic import SemanticChunking
from bitca.knowledge.pdf import PDFUrlKnowledgeBase
from bitca.vectordb.pgvector import PgVector

db_url = "postgresql+psycopg://ai:ai@localhost:5532/ai"

knowledge_base = PDFUrlKnowledgeBase(
    urls=["https://bitca-public.s3.amazonaws.com/recipes/ThaiRecipes.pdf"],
    vector_db=PgVector(table_name="recipes_semantic_chunking", db_url=db_url),
    chunking_strategy=SemanticChunking(),
)
knowledge_base.load(recreate=False)  # Comment out after first run

agent = Agent(
    knowledge_base=knowledge_base,
    search_knowledge=True,
)

agent.print_response("How to make Thai curry?", markdown=True)

​Params

Parameter	Type	Default	Description
embedder	Embedder	OpenAIEmbedder	The embedder to use for semantic chunking.
chunk_size	int	5000	The maximum size of each chunk.
similarity_threshold	float	0.5	The similarity threshold for determining chunk boundaries.