Generate LlamaDataset Metadata Files
Generate LlamaDataset metadata files (card.json, README.md) for submissions to LlamaHub. Expedites the process of packaging your RAG datasets.
Why it matters
Streamline your LlamaIndex dataset submission process by automatically generating the required `card.json` and `README.md` metadata files. This pack simplifies the creation of documentation for your RAG evaluations.
Outcomes
What it gets done
Create `card.json` and `README.md` for LlamaDataset submissions.
Automate metadata generation based on RAG evaluation results.
Provide templates for dataset documentation.
Integrate with LlamaIndex CLI for easy download and usage.
Install
Add it to your toolbox
Run in your project directory:
curl -fsSL https://spark.entire.vc/get/li-pack-packs-llama-dataset-metadata | bash Steps
Steps in the chain
Use llamaindex-cli to download the pack: llamaindex-cli download-llamapack LlamaDatasetMetadataPack --download-dir ./llama_dataset_metadata_pack
Inspect the files at ./llama_dataset_metadata_pack and use them as a template for your own project
Download the pack to ./llama_dataset_metadata_pack directory using Python: from llama_index.core.llama_pack import download_llama_pack; LlamaDatasetMetadataPack = download_llama_pack('LlamaDatasetMetadataPack', './llama_dataset_metadata_pack')
Create an instance of LlamaDatasetMetadataPack: llama_dataset_metadata_pack = LlamaDatasetMetadataPack()
Call the run method with dataset name, description, rag_dataset, index, benchmark_df, and baseline_name to create and save card.json and README.md to disk
Overview
LlamaDataset Metadata Pack
What it does
This pack creates `card.json` and `README.md` metadata files for `LlamaDataset` submissions to LlamaHub, helping to expedite the packaging process.
How it connects
Use this pack when preparing a `LlamaDataset` submission package for LlamaHub. It is intended for use after a RAG evaluation has been performed, as it requires outputs like `index`, `rag_dataset`, and `benchmark_df` from such an evaluation.
Source README
Description pending for li-pack-packs-llama-dataset-metadata.
Step 1: Download LlamaDatasetMetadataPack using CLI
Use llamaindex-cli to download the pack: llamaindex-cli download-llamapack LlamaDatasetMetadataPack --download-dir ./llama_dataset_metadata_pack
Step 2: Inspect downloaded files
Inspect the files at ./llama_dataset_metadata_pack and use them as a template for your own project
Step 3: Download pack via Python code
Download the pack to ./llama_dataset_metadata_pack directory using Python: from llama_index.core.llama_pack import download_llama_pack; LlamaDatasetMetadataPack = download_llama_pack('LlamaDatasetMetadataPack', './llama_dataset_metadata_pack')Step 4: Construct LlamaDatasetMetadataPack instance
Create an instance of LlamaDatasetMetadataPack: llama_dataset_metadata_pack = LlamaDatasetMetadataPack()
Step 5: Run pack to generate metadata files
Call the run method with dataset name, description, rag_dataset, index, benchmark_df, and baseline_name to create and save card.json and README.md to disk
Discussion
Questions & comments · 0
Sign In Sign in to leave a comment.