Package barky provides utilities for handling hierarchical key/value data structures that commonly appear in configuration formats such as JSON, YAML, or TOML. It is designed to transform nested data into a flat representation, while preserving enough metadata to reconstruct paths, detect conflicts, and manage data from multiple sources. Key features include: Flattening: Nested maps, slices, and arrays can be converted into a flat map[string]string using dot notation for maps and index notation for arrays/slices. For example, {"db": {"hosts": ["a", "b"]}} becomes {"db.hosts[0]": "a", "db.hosts[1]": "b"}. Path handling: The package defines a Path abstraction that represents hierarchical keys as a sequence of typed segments (map keys or array indices). Paths can be split from strings like "foo.bar[0]" or joined back into their string form. Storage: A Storage type manages a collection of flattened key/value pairs. It builds and maintains a hierarchical tree internally to prevent property conflicts (e.g., treating the same key as both a map and a value). Storage also associates values with the files they originated from, which allows multi-file merging and provenance tracking. Querying: The Storage type provides helper methods for retrieving values, checking for the existence of keys, enumerating subkeys, and iterating in a deterministic order. Typical use cases: Overall, barky acts as a bridge between deeply nested structured data and flat, queryable representations that are easier to work with in configuration management, testing, or data transformation pipelines.