Examples

Hands-on examples showing atdata in action

These examples are self-contained demonstrations of atdata’s key features. Each one builds a working pipeline from scratch so you can see real code paths end-to-end.

Example	What It Shows
Typed Dataset Pipeline	Define a sample type, write shards, iterate with batching
Lens Transformations	View datasets through different schemas without copying data
Manifest-Powered Queries	Build per-shard manifests and query samples by metadata
Index-Managed Datasets	Use Index with LocalDiskStore for managed dataset storage
Multi-Split Datasets	Work with train/test splits via DatasetDict and load_dataset
Dataset Profiler	Profile datasets visually using manifest aggregates and matplotlib
Lens Graph Explorer	Visualize lens transformation networks with Mermaid diagrams
Query Cookbook	Practical query recipes with the typed proxy DSL and scatter plots

New to atdata?

Start with the Quick Start tutorial for foundational concepts, then come back here for deeper worked examples.