Code Craft: Hierarchical Path Key-Value Store

OOD Code Craft round. Implement a hierarchical key-value store using directory-like paths (e.g. `/a/b/c`). The store starts with a root path `/` whose value is `"root"`. Operations: `create`, `set`, `get`, `delete` (leaf-only; root cannot be deleted). Requirements Initial state: root path / exists with value "root". create(path, value): create a new path with the given value. Only succeeds if the parent path already exists. set(path, value): update the value at an existing path. get(path): return the value stored at the path. delete(path): delete a path only if it is a leaf node. Root cannot be deleted. Method signatures (from the prompt): class PathKeyValueStore: def create(self, path: str, value: str) -> str def set(self, path: str, value: str) -> str def get(self, path: str) -> str def remove(self, path: str) -> str Return values are left to the candidate to define; "OK" / the affected value / a boolean string are all accepted. Notes Two natural representations: Tree of nodes — each node holds value + children map. Path operations split on / and walk children. delete checks len(children) == 0. Most natural for create parent-exists check. Flat dict — {path: value} plus parent-existence checks on create. Simpler code but delete needs a children-count helper (any(p.startswith(path + '/') for p in store)). The tree version is cleaner for the leaf-check on delete; the flat version is faster to type. Either is acceptable. Edge cases: trailing slash normalization (/a/b vs /a/b/), root special-case in all four operations, empty path string handling, set on a non-existent path (clarify: error vs upsert). Concurrency follow-up is common: "how would you make this thread-safe?" — per-node RW lock or a single store-level lock; mention copy-on-write for read-heavy workloads. Persistence follow-up: write-ahead log of mutations; rebuild tree on startup; snapshot periodically. Preparation Write the tree-based implementation in under 25 minutes cold. Memorize the path-splitting boilerplate in your language of choice (path.strip('/').split('/')). Have a 60-second answer for thread safety and one for persistence. Pre-write 5 test cases: create with missing parent, set on non-existent, delete of non-leaf, delete of root, get on missing path.