good first issue
Repository metrics
- Stars
- (1 star)
- PR merge metrics
- (PR metrics pending)
説明
Goal
Component to split Markdown/semantic text into offset chunks for chunked retrieval (RAG). Might use headings/paragraphs as boundaries or basic token counts with overlaps.
Scope/Features
- Input: string
- Output: structured array of chunks (text, offset/indices)
- Configurable: chunk size, overlap, min paragraph, etc.
Rough WIT shape
split(input: string, params: chunk_params) -> list<chunk>- Chunk struct: {text, start, end, ...}
Labels
- good first issue