Zanaptak.BinaryToTextEncoding

A binary-to-text encoder/decoder library for .NET and Fable. Provides base 16, base 32, base 46, base 64, and base 91 codecs. Supports custom character sets.

Output example

Example of a random 16-byte array (same size as a GUID) encoded in each base:

• Base 16: 3A319D0D6BA340E8CFFA6E8F65236B71
• Base 32: HIYZ2DLLUNAORT72N2HWKI3LOE
• Base 46: G7YXHjqTF4THH7KYYxCBr4sM
• Base 64: OjGdDWujQOjP+m6PZSNrcQ
• Base 91: 7M515sme(-[9YfN?/LIf

Encoded bits per character

The base values in this library have been chosen because they can encode an integral number of bits as either 1 or 2 characters, making the conversion relatively efficient since groups of bits can be directly converted using lookup arrays.

• Base 16: 4 bits per character
• Base 32: 5 bits per character
• Base 46: 5.5 bits per character (11 bits per 2 characters)
• Base 64: 6 bits per character
• Base 91: 6.5 bits per character (13 bits per 2 characters)

Usage

Add the NuGet package to your project:

dotnet add package Zanaptak.BinaryToTextEncoding

C#

using Zanaptak.BinaryToTextEncoding;

// Default codec
var originalBytes = new byte[] { 1, 2, 3 };
var encodedString = Base32.Default.Encode(originalBytes);
var decodedBytes = Base32.Default.Decode(encodedString);

// Custom character set
var customBase32 = new Base32("BCDFHJKMNPQRSTXZbcdfhjkmnpqrstxz");
var customOriginalBytes = new byte[] { 4, 5, 6 };
var customEncodedString = customBase32.Encode(customOriginalBytes);
var customDecodedBytes = customBase32.Decode(customEncodedString);

// Wrap output
var randomBytes = new byte[100];
new System.Random(12345).NextBytes(randomBytes);
Console.WriteLine(Base91.Default.Encode(randomBytes, 48));
//  Output:
//  r]g^oP{ZKd1>}lC{C*P){O96SL8z%0TW,4BfEof}%!b@a#:6
//  nN<c#=}80|srYHUy6$XP}4x945a~,ItFPS;U%a^<DMA]@m|#
//  12tC]*5+BoT-4Th,oVR9wvIv;Iym

F#

open Zanaptak.BinaryToTextEncoding

// Default codec
let originalBytes = [| 1uy; 2uy; 3uy |]
let encodedString = Base32.Default.Encode originalBytes
let decodedBytes = Base32.Default.Decode encodedString

// Custom character set
let customBase32 = Base32("BCDFHJKMNPQRSTXZbcdfhjkmnpqrstxz")
let customOriginalBytes = [| 4uy; 5uy; 6uy |]
let customEncodedString = customBase32.Encode customOriginalBytes
let customDecodedBytes = customBase32.Decode customEncodedString

// Wrap output
let randomBytes = Array.create 100 0uy
System.Random(12345).NextBytes(randomBytes)
printfn "%s" (Base91.Default.Encode(randomBytes, 48))
//  Output:
//  r]g^oP{ZKd1>}lC{C*P){O96SL8z%0TW,4BfEof}%!b@a#:6
//  nN<c#=}80|srYHUy6$XP}4x945a~,ItFPS;U%a^<DMA]@m|#
//  12tC]*5+BoT-4Th,oVR9wvIv;Iym

Notes

• Sortable
  • If the character set is in ASCII order, then an ASCII string sort of the encoded outputs is the same as a numeric sort of the inputs, for inputs of the same length in bytes.
  • Note however that some of the default character sets used by different base values in this library are not in ASCII order because they are aligned with traditional implementations. A custom character set should be used if sortability is required.
• Case-sensitive
  • Encoding/decoding is case-sensitive using the exact characters in the character set. If case-insensitivity is required, it must be handled externally by converting to the case used in the character set.
• Padding not supported
  • To reduce complexity, this library does not support padding. Padding does not affect encode/decode accuracy, only string length normalization. It is not needed when exact string lengths are known or otherwise delimited (such as quoted JSON strings). If required, it must be handled externally by trimming or appending as necessary.

Built-in character sets

Base16	Description	Characters
StandardCharacterSet	(Default) Standard hexadecimal notation, ASCII-sortable	0123456789ABCDEF
ConsonantsCharacterSet	Excludes numbers, vowels, and some confusable letters, ASCII-sortable	BCDFHJKMNPQRSTXZ

Base32	Description	Characters
StandardCharacterSet	(Default) RFC 4648 section 6	ABCDEFGHIJKLMNOPQRSTUVWXYZ234567
HexExtendedCharacterSet	RFC 4648 section 7, ASCII-sortable	0123456789ABCDEFGHIJKLMNOPQRSTUV
ConsonantsCharacterSet	Excludes numbers, vowels, and some confusable letters, ASCII-sortable	BCDFHJKMNPQRSTXZbcdfhjkmnpqrstxz

Base46	Description	Characters
SortableCharacterSet	(Default) Excludes vowels and some confusable characters, ASCII-sortable	234567BCDFGHJKMNPQRSTVW XYZbcdfghjkmnpqrstvwxyz
LettersCharacterSet	Excludes numbers and some confusable letters, ASCII-sortable	ABCDEFGHJKMNPQRSTUVWXYZ abcdefghjkmnpqrstuvwxyz

Base64	Description	Characters
StandardCharacterSet	(Default) RFC 4648 section 4	ABCDEFGHIJKLMNOPQRSTUVWXYZabcdef ghijklmnopqrstuvwxyz0123456789+/
UrlSafeCharacterSet	RFC 4648 section 5	ABCDEFGHIJKLMNOPQRSTUVWXYZabcdef ghijklmnopqrstuvwxyz0123456789-_
UnixCryptCharacterSet	Unix crypt password hashes, ASCII-sortable	./0123456789ABCDEFGHIJKLMNOPQRST UVWXYZabcdefghijklmnopqrstuvwxyz

Base91	Description	Characters
SortableQuotableCharacterSet	(Default) Excludes " ' \ characters, ASCII-sortable	!#$%&()*+,-./0123456789:;<=>?@A BCDEFGHIJKLMNOPQRSTUVWXYZ[]^_`a bcdefghijklmnopqrstuvwxyz{|}~

Base91Legacy	Description	Characters
LegacyCharacterSet	(Default) Original 'basE91' character set	ABCDEFGHIJKLMNOPQRSTUVWXYZabcdef ghijklmnopqrstuvwxyz0123456789!# $%&()*+,./:;<=>?@[]^_`{|}~"

Legacy 'basE91' compatibility

This library provides two base 91 implementations: Base91 and Base91Legacy. They are not compatible; the encoded output of one cannot be decoded by the other.

The main Base91 algorithm works like the other BaseXX algorithms in the library. It encodes with constant-width (each 2-character pair encodes exactly 13 bits) in big-endian order (most-significant character fist, representing the most-significant bits of the most-significant byte). The default character set is in ASCII order to preserve sortability of input, and excludes the characters ", ', and \ to make it more easily quotable in programming languages.

Base91Legacy is based on the previously existing basE91 algorithm. It encodes with a variable-width mechanism (some 2-character pairs can encode 14 bits instead of 13) which can result in slightly smaller encoded strings. Each two-character pair in the output is swapped compared to the main algorithm (least-significant char of the pair first), so sorting by string is not meaningful regardless of character set. Its default character set includes the " character, making it inconvenient to use in some programming languages and data formats such as JSON.

Benchmarks

See the benchmark project.