usc-chroma-vecs-v1-chunks-v1-s8192-o512-sentence-transformers-static-retrieval-mrl-en-v1

This dataset contains a pre-built ChromaDB database with US Congressional legislation embeddings.

Dataset Structure

This dataset contains the ChromaDB files for legislation chunks with embeddings. The database can be loaded directly using ChromaDB's PersistentClient.

Usage

import chromadb
from huggingface_hub import snapshot_download

# Download the dataset
local_dir = snapshot_download(repo_id="hyperdemocracy/usc-chroma-vecs-v1-chunks-v1-s8192-o512-sentence-transformers-static-retrieval-mrl-en-v1", repo_type="dataset")

# Load the ChromaDB
client = chromadb.PersistentClient(path=local_dir + "/chromadb")
collection = client.get_collection("usc")

# Query the database (embedding function is already configured)
results = collection.query(
    query_texts=["your query here"],
    n_results=10
)

# Get collection metadata to see embedding model info
metadata = collection.metadata
print(f"Embedding model: {metadata['embedding_model']}")
print(f"Chunk size: {metadata['chunk_size']}")

Configuration

This ChromaDB includes the following configuration stored as collection metadata:

• Embedding model: The SentenceTransformer model used for embeddings
• Chunk size: Character count for text chunking
• Chunk overlap: Overlap between adjacent chunks
• Chunking version: Version of the chunking algorithm
• Embedding version: Version of the embedding process

The embedding function is automatically configured when loading the collection, enabling direct text queries without manual embedding.