Module ulid 

Lowercase ULID minting and validation (SPEC § The id field, format v0.4).

A ULID is a 128-bit identifier — a 48-bit millisecond timestamp followed by 80 random bits — rendered as 26 characters of Crockford base32. db.md’s canonical form is lowercase (YAML-clean, shell-friendly, reads like the rest of the frontmatter). dbmd write mints one for every new content file that carries no id; the leading timestamp makes freshly minted ids time-sortable, and the 80 random bits make offline minting coordination-free.

Std-only by design. The toolkit’s dependency discipline (zero AI deps, minimal tree) rules out pulling a ulid/rand stack for one mint call. Randomness comes from RandomState — std’s hasher keys are seeded from OS entropy — mixed with the wall clock, the PID, and a process-global counter. That is not cryptographic randomness and does not need to be: the id’s contract is store-scoped uniqueness (DUP_ID is the backstop), not unguessability.

Functions

is_ulid

True when s is a well-formed lowercase ULID: exactly 26 chars, all lowercase Crockford base32, first char 0–7 (so the value fits 128 bits). Uppercase or mixed-case forms are not the db.md canonical form and return false — as does any other opaque id, which stays legal in a store (SPEC: the ULID form is recommended, never a validation gate).

mint

Mint a fresh lowercase ULID: 48-bit Unix-millisecond timestamp + 80 random bits, encoded as 26 chars of lowercase Crockford base32.