Encoding API

The piko.sh/piko/wdk/encoding package provides:

• A generic Base-X encoder (*Encoding) usable with any non-ambiguous alphabet.
• Convenience functions for Base-36 and Base-58 (the most common short-string formats).
• UUID v7 generators (sortable IDs).
• Short-UUID helpers that pack a UUID into a compact Base-58 string.

There are no Base64() / Base32() / Hex() factory functions. Go's standard library already covers those formats (encoding/base64, encoding/base32, encoding/hex). Use them directly.

Base-X encoder

type Encoding struct { /* ... */ }

func NewEncoding(alphabet string) (*Encoding, error)

func (enc *Encoding) EncodeBytes(data []byte) string
func (enc *Encoding) DecodeBytes(input string) ([]byte, error)
func (enc *Encoding) EncodeUint64(value uint64) string
func (enc *Encoding) DecodeUint64(input string) (uint64, error)

NewEncoding rejects empty alphabets and alphabets with repeated characters (ErrAlphabetEmpty, ErrAlphabetAmbiguous). The alphabet length implies the base.

Alphabet constants

Callers that need a custom encoder constructed from a known alphabet can pass these directly (NewEncoding(StdBase64Alphabet)).

Base-36

func EncodeBytesBase36(data []byte) string
func DecodeBytesBase36(input string) ([]byte, error)
func EncodeUint64Base36(value uint64) string
func DecodeUint64Base36(input string) (uint64, error)

Base-58

func EncodeBytesBase58(data []byte) string
func DecodeBytesBase58(input string) ([]byte, error)
func EncodeUint64Base58(value uint64) string
func DecodeUint64Base58(input string) (uint64, error)

The Base-58 alphabet excludes the visually ambiguous 0, O, I, and l so encoded strings are safe to read aloud or transcribe.

UUID v7

UUID v7 values are sortable by creation time, which makes them cheap to index in a database.

func NewV7At(t time.Time) (uuid.UUID, error)
func NewV7AtFromReader(r io.Reader, t time.Time) (uuid.UUID, error)
func NewV7MinAt(t time.Time) uuid.UUID
func NewV7MaxAt(t time.Time) uuid.UUID
func ResetLastV7timeAt()

Short-UUID helpers

Pack a UUID into a compact human-friendly string. Three flavours:

// Base64URL short string (RFC 4648 URL-safe, 22 chars).
func UUIDToShortString(id uuid.UUID) string
func ShortStringToUUID(s string) (uuid.UUID, error)
func MustShortStringToUUID(s string) uuid.UUID

// Base-58 short string (21-22 chars). Uses the unambiguous Base-58 alphabet.
func UUIDToBase58String(id uuid.UUID) string
func Base58StringToUUID(s string) (uuid.UUID, error)
func MustBase58StringToUUID(s string) uuid.UUID

// Even shorter (drops the variant/version bits and reconstructs them on decode).
// Requires you to specify which version when decoding. Output is 21 chars.
func UUIDToShorterString(id uuid.UUID) string
func ShorterStringToUUID(s string, version int) (uuid.UUID, error)
func ShorterStringToUUIDv1(s string) (uuid.UUID, error)
func ShorterStringToUUIDv4(s string) (uuid.UUID, error)
func ShorterStringToUUIDv7(s string) (uuid.UUID, error)
func MustShorterStringToUUID(s string, version int) uuid.UUID

UUIDToShortString and UUIDToBase58String are different encodings, not aliases. Pick UUIDToShortString when you want the canonical URL-safe representation. Pick UUIDToBase58String when humans read the string aloud or transcribe it. The "shorter" form trades flexibility for length. Because the encoded string omits the version, you must know which version you are decoding.

Errors

Both match via errors.Is.

See also

• Source: wdk/encoding/.
• encoding/base64, encoding/base32, encoding/hex in the Go standard library for those alphabets.