Chunking with Overlap

Split a list of tokens into fixed-size chunks with overlapping windows — a standard preprocessing step for RAG ingestion.

Signature: def chunk_text(tokens: list, chunk_size: int, overlap: int) -> list

• Each chunk has at most chunk_size tokens.
• Consecutive chunks share overlap tokens (so the step between chunks is chunk_size - overlap).
• The final chunk may be shorter if the tokens don't divide evenly.
• Return a list of token-list chunks.

Constraint: overlap < chunk_size.