Melody is a language that compiles to ECMAScript regular expressions, while aiming to be more readable and maintainable.
Note: these are for the currently supported syntax and may change
Batman Theme try in playground
16 of "na";
2 of match {
<space>;
"batman";
}
// 🦇🦸♂️
Turns into
(?:na){16}(?: batman){2}
Twitter Hashtag try in playground
"#";
some of <word>;
// #melody
Turns into
#\w+
Introductory Courses try in playground
some of <alphabetic>;
<space>;
"1";
2 of <digit>;
// classname 1xx
Turns into
[a-zA-Z]+ 1\d{2}
Indented Code (2 spaces) try in playground
some of match {
2 of <space>;
}
some of <char>;
";";
// let value = 5;
Turns into
(?: {2})+.+;
Semantic Versions try in playground
<start>;
option of "v";
capture major {
some of <digit>;
}
".";
capture minor {
some of <digit>;
}
".";
capture patch {
some of <digit>;
}
<end>;
// v1.0.0
Turns into
^v?(?<major>\d+)\.(?<minor>\d+)\.(?<patch>\d+)$
You can try Melody in your browser using the playground
Read the book here
cargo install melody_cli
git clone https://github.com/yoav-lavi/melody.git
cd melody
cargo install --path crates/melody_cli
- macOS binaries (
aarch64
andx86_64
) can be downloaded from the release page
-
Brew (macOS and Linux)
Installation instructions
brew install melody
-
Arch Linux (maintained by @ilai-deutel)
Installation instructions
-
Installation with an AUR helper, for instance using
paru
:paru -Syu melody
-
Install manually with
makepkg
:git clone https://aur.archlinux.org/melody.git cd melody makepkg -si
-
-
Installation instructions
-
Declarative installation using
/etc/nixos/configuration.nix
:{ pkgs, ... }: { environment.systemPackages = with pkgs; [ melody ]; }
-
Imperative installation using
nix-env
:nix-env -iA nixos.melody
-
USAGE:
melody [OPTIONS] [INPUT_FILE_PATH]
ARGS:
<INPUT_FILE_PATH> Read from a file
Use '-' and or pipe input to read from stdin
OPTIONS:
-f, --test-file <TEST_FILE>
Test the compiled regex against the contents of a file
--generate-completions <COMPLETIONS>
Outputs completions for the selected shell
To use, write the output to the appropriate location for your shell
-h, --help
Print help information
-n, --no-color
Print output with no color
-o, --output <OUTPUT_FILE_PATH>
Write to a file
-r, --repl
Start the Melody REPL
-t, --test <TEST>
Test the compiled regex against a string
-V, --version
Print version information
See the changelog here or in the release page
... of
- used to express a specific amount of a pattern. equivalent to regex{5}
(assuming5 of ...
)... to ... of
- used to express an amount within a range of a pattern. equivalent to regex{5,9}
(assuming5 to 9 of ...
)over ... of
- used to express more than an amount of a pattern. equivalent to regex{6,}
(assumingover 5 of ...
)some of
- used to express 1 or more of a pattern. equivalent to regex+
any of
- used to express 0 or more of a pattern. equivalent to regex*
option of
- used to express 0 or 1 of a pattern. equivalent to regex?
All quantifiers can be preceded by lazy
to match the least amount of characters rather than the most characters (greedy). Equivalent to regex +?
, *?
, etc.
<char>
- matches any single character. equivalent to regex.
<space>
- matches a space character. equivalent to regex<whitespace>
- matches any kind of whitespace character. equivalent to regex\s
or[ \t\n\v\f\r]
<newline>
- matches a newline character. equivalent to regex\n
<tab>
- matches a tab character. equivalent to regex\t
<return>
- matches a carriage return character. equivalent to regex\r
<feed>
- matches a form feed character. equivalent to regex\f
<null>
- matches a null characther. equivalent to regex\0
<digit>
- matches any single digit. equivalent to regex\d
or[0-9]
<vertical>
- matches a vertical tab character. equivalent to regex\v
<word>
- matches a word character (any latin letter, any digit or an underscore). equivalent to regex\w
or[a-zA-Z0-9_]
<alphabetic>
- matches any single latin letter. equivalent to regex[a-zA-Z]
<alphanumeric>
- matches any single latin letter or any single digit. equivalent to regex[a-zA-Z0-9]
<boundary>
- Matches a character between a character matched by<word>
and a character not matched by<word>
without consuming the character. equivalent to regex\b
<backspace>
- matches a backspace control character. equivalent to regex[\b]
All symbols can be preceeded with not
to match any character other than the symbol
<start>
- matches the start of the string. equivalent to regex^
<end>
- matches the end of the string. equivalent to regex$
Note: these are not supported when testing in the CLI (-t
or -f
) as the regex engine used does not support unicode categories. These require using the u
flag.
<category::letter>
- any kind of letter from any language<category::lowercase_letter>
- a lowercase letter that has an uppercase variant<category::uppercase_letter>
- an uppercase letter that has a lowercase variant.<category::titlecase_letter>
- a letter that appears at the start of a word when only the first letter of the word is capitalized<category::cased_letter>
- a letter that exists in lowercase and uppercase variants<category::modifier_letter>
- a special character that is used like a letter<category::other_letter>
- a letter or ideograph that does not have lowercase and uppercase variants
<category::mark>
- a character intended to be combined with another character (e.g. accents, umlauts, enclosing boxes, etc.)<category::non_spacing_mark>
- a character intended to be combined with another character without taking up extra space (e.g. accents, umlauts, etc.)<category::spacing_combining_mark>
- a character intended to be combined with another character that takes up extra space (vowel signs in many Eastern languages)<category::enclosing_mark>
- a character that encloses the character it is combined with (circle, square, keycap, etc.)
<category::separator>
- any kind of whitespace or invisible separator<category::space_separator>
- a whitespace character that is invisible, but does take up space<category::line_separator>
- line separator character U+2028<category::paragraph_separator>
- paragraph separator character U+2029
<category::symbol>
- math symbols, currency signs, dingbats, box-drawing characters, etc<category::math_symbol>
- any mathematical symbol<category::currency_symbol>
- any currency sign<category::modifier_symbol>
- a combining character (mark) as a full character on its own<category::other_symbol>
- various symbols that are not math symbols, currency signs, or combining characters
<category::number>
- any kind of numeric character in any script<category::decimal_digit_number>
- a digit zero through nine in any script except ideographic scripts<category::letter_number>
- a number that looks like a letter, such as a Roman numeral<category::other_number>
- a superscript or subscript digit, or a number that is not a digit 0–9 (excluding numbers from ideographic scripts)
<category::punctuation>
- any kind of punctuation character<category::dash_punctuation>
- any kind of hyphen or dash<category::open_punctuation>
- any kind of opening bracket<category::close_punctuation>
- any kind of closing bracket<category::initial_punctuation>
- any kind of opening quote<category::final_punctuation>
- any kind of closing quote<category::connector_punctuation>
- a punctuation character such as an underscore that connects words<category::other_punctuation>
- any kind of punctuation character that is not a dash, bracket, quote or connectors
<category::other>
- invisible control characters and unused code points<category::control>
- an ASCII or Latin-1 control character: 0x00–0x1F and 0x7F–0x9F<category::format>
- invisible formatting indicator<category::private_use>
- any code point reserved for private use<category::surrogate>
- one half of a surrogate pair in UTF-16 encoding<category::unassigned>
- any code point to which no character has been assigned
These descriptions are from regular-expressions.info
... to ...
- used with digits or alphabetic characters to express a character range. equivalent to regex[5-9]
(assuming5 to 9
) or[a-z]
(assuminga to z
)
"..."
or'...'
- used to mark a literal part of the match. Melody will automatically escape characters as needed. Quotes (of the same kind surrounding the literal) should be escaped
`...`
- added directly to the output without any escaping
capture
- used to open acapture
or namedcapture
block. capture patterns are later available in the list of matches (either positional or named). equivalent to regex(...)
match
- used to open amatch
block, matches the contents without capturing. equivalent to regex(?:...)
either
- used to open aneither
block, matches one of the statements within the block. equivalent to regex(?:...|...)
ahead
- used to open anahead
block. equivalent to regex(?=...)
. use after an expressionbehind
- used to open anbehind
block. equivalent to regex(?<=...)
. use before an expression
Assertions can be preceeded by not
to create a negative assertion (equivalent to regex (?!...)
, (?<!...)
)
-
let .variable_name = { ... }
- defines a variable from a block of statements. can later be used with.variable_name
. Variables must be declared before being used. Variable invocations cannot be quantified directly, use a group if you want to quantify a variable invocationexample:
let .a_and_b = { "a"; "b"; } .a_and_b; "c"; // abc
/* ... */
,// ...
- used to mark comments (note:// ...
comments must be on separate line)
The Melody file extensions are .mdy
and .melody
melody_compiler
- The Melody compiler 📦 📖melody_cli
- A CLI wrapping the Melody compiler 📦 📖melody_wasm
- WASM bindings for the Melody compiler
Last measured on v0.19.0
Measured on an 8 core 2021 MacBook Pro 14-inch, Apple M1 Pro using criterion:
-
8 lines:
compiler/normal (8 lines) time: [4.0579 µs 4.0665 µs 4.0788 µs] slope [4.0579 µs 4.0788 µs] R^2 [0.9996538 0.9995633] mean [4.0555 µs 4.0806 µs] std. dev. [11.018 ns 26.342 ns] median [4.0500 µs 4.0852 µs] med. abs. dev. [5.6889 ns 35.806 ns]
-
1M lines:
compiler/long input (1M lines) time: [400.97 ms 402.31 ms 403.53 ms] mean [400.97 ms 403.53 ms] std. dev. [773.42 µs 2.9886 ms] median [401.22 ms 403.39 ms] med. abs. dev. [59.042 µs 3.5129 ms]
-
Deeply nested:
compiler/deeply nested time: [5.8085 µs 5.8291 µs 5.8514 µs] slope [5.8085 µs 5.8514 µs] R^2 [0.9992861 0.9992461] mean [5.8064 µs 5.8519 µs] std. dev. [21.027 ns 49.152 ns] median [5.7949 µs 5.8583 µs] med. abs. dev. [3.3348 ns 64.628 ns]
To reproduce, run cargo bench
or cargo xtask benchmark
🐣 - Partially implemented
❌ - Not implemented
❔ - Unclear what the syntax will be
❓ - Unclear whether this will be implemented
Melody | Regex | Status |
---|---|---|
not "A"; |
[^A] |
🐣 |
variables / macros | 🐣 | |
<...::...> |
\p{...} |
🐣 |
not <...::...> |
\P{...} |
🐣 |
file watcher | ❌ | |
multiline groups in REPL | ❌ | |
flags: global, multiline, ... |
/.../gm... |
❔ |
(?) | \# |
❔ |
(?) | \k<name> |
❔ |
(?) | \uYYYY |
❔ |
(?) | \xYY |
❔ |
(?) | \ddd |
❔ |
(?) | \cY |
❔ |
(?) | $1 |
❔ |
(?) | $` |
❔ |
(?) | $& |
❔ |
(?) | x20 |
❔ |
(?) | x{06fa} |
❔ |
any of "a", "b", "c" * |
[abc] |
❓ |
multiple ranges * | [a-zA-Z0-9] |
❓ |
regex optimization | ❓ | |
standard library / patterns | ❓ | |
reverse compiler | ❓ |
* these are expressable in the current syntax using other methods