Skip to main content

llama-index packs node_parser integration

Project description

Semantic Chunking Llama Pack

This LlamaPack implements the semantic chunking algorithm first proposed by Greg Kamradt in his Five Levels of Text Splitting tutorial.

How it works:

  • Split text into sentences.
  • For each sentence, generate an embedding.
  • Measure cosine distance between each pair of consecutive sentences.
  • Get the 95% percentile cosine distance, set that as the threshold.
  • Create a new chunk if the cosine distance of a sentence compared to prev. exceeds that threshold.

CLI Usage

You can download llamapacks directly using llamaindex-cli, which comes installed with the llama-index python package:

llamaindex-cli download-llamapack LLMCompilerAgentPack --download-dir ./llm_compiler_agent_pack

You can then inspect the files at ./llm_compiler_agent_pack and use them as a template for your own project!

Code Usage

You can download the pack to a directory. NOTE: You must specify skip_load=True - the pack contains multiple files, which makes it hard to load directly.

We will show you how to import the agent from these files!

Project details


Download files

Download the file for your platform. If you're not sure which to choose, learn more about installing packages.

Source Distribution

llama_index_packs_node_parser-0.1.2.tar.gz (4.5 kB view details)

Uploaded Source

Built Distribution

File details

Details for the file llama_index_packs_node_parser-0.1.2.tar.gz.

File metadata

File hashes

Hashes for llama_index_packs_node_parser-0.1.2.tar.gz
Algorithm Hash digest
SHA256 09a40616b05f56ac952cc209a75f7ecfc19c5208d8c2ce7b6012347909a5026b
MD5 ad63a07110505700ba4eeaf0aa96a2e9
BLAKE2b-256 97b28e99d14646e5c84a94368515e72995f652166f765577a35ef5d0db0d29f6

See more details on using hashes here.

File details

Details for the file llama_index_packs_node_parser-0.1.2-py3-none-any.whl.

File metadata

File hashes

Hashes for llama_index_packs_node_parser-0.1.2-py3-none-any.whl
Algorithm Hash digest
SHA256 625feb89f0df8811e57cad24945282a396d28ba2cdd0ffa8bd9cd8c9518cf77b
MD5 afd8d31045161f68b1a3ea22fbbf6db6
BLAKE2b-256 0c6dc39c2dc9d15b50c42d2c0ccf09c8dd7d765ce102337d8b2b9c43b53d5a91

See more details on using hashes here.

Supported by

AWS AWS Cloud computing and Security Sponsor Datadog Datadog Monitoring Fastly Fastly CDN Google Google Download Analytics Microsoft Microsoft PSF Sponsor Pingdom Pingdom Monitoring Sentry Sentry Error logging StatusPage StatusPage Status page