84 lines
3.1 KiB
Markdown
84 lines
3.1 KiB
Markdown
# dfa
|
|
|
|
Compiles a regular expression like syntax to fast deterministic finite automata.
|
|
Useful for pattern matching against non-string sequences.
|
|
|
|
## Example
|
|
|
|
This example matches [Hangul](https://en.wikipedia.org/wiki/Hangul) syllables. The symbols defined in the machine are Unicode character categories which could be mapped from code points.
|
|
|
|
Machine definition:
|
|
|
|
```coffeescript
|
|
# define symbols
|
|
X = 0; # Other character
|
|
L = 1; # Leading consonant
|
|
V = 2; # Medial vowel
|
|
T = 3; # Trailing consonant
|
|
LV = 4; # Composed <LV> syllable
|
|
LVT = 5; # Composed <LVT> syllable
|
|
M = 6; # Tone mark
|
|
|
|
# define variables
|
|
decomposed = L V T?;
|
|
partial = LV T?;
|
|
composed = LVT;
|
|
|
|
# define main state machine pattern
|
|
main = (decomposed | partial | composed) M?;
|
|
```
|
|
|
|
Visualized, the machine looks like this (double circles are accepting states):
|
|
|
|
|
|
|
|
Compiling and using the machine:
|
|
|
|
```javascript
|
|
import compile from 'dfa/compile';
|
|
import fs from 'fs';
|
|
|
|
let stateMachine = compile(fs.readFileSync('hangul.machine', 'utf8'));
|
|
|
|
// find matches
|
|
for (let [startIndex, endIndex] of stateMachine.match([0, 1, 2, 3, 0, 4, 6]) {
|
|
console.log('match:', startIndex, endIndex);
|
|
}
|
|
```
|
|
|
|
Output:
|
|
```
|
|
match: 1 3
|
|
match: 5 6
|
|
```
|
|
|
|
## Syntax
|
|
|
|
A state machine file contains a list of assignment statements. Comments are also allowed
|
|
and are started with the `#` character. Each statement is an assignment of a variable name
|
|
to a value or expression. Assigning a variable to a number produces a symbol, which is
|
|
added to the state machine's alphabet. Assigning a variable to an expression allows
|
|
for substitutions into later expressions. The special `main` variable should always be
|
|
assigned to at the end of the file, and is the final expression that will be compiled.
|
|
|
|
A subset of common regular expression syntax is supported. A list of operators and their
|
|
precedence is below. Operators with the same precedence are evaluated left to right.
|
|
|
|
| Precedence | Syntax | Type | Meaning |
|
|
| ---------- | ---------- | --------------| ------------------------------------------ |
|
|
| 1 | `a \| b` | Alternation | Matches either `a` or `b` |
|
|
| 2 | `a b` | Concatenation | Matches `a` followed by `b` |
|
|
| 3 | `a*` | Repetition | Matches zero or more occurrences of `a` |
|
|
| 3 | `a+` | Repetition | Matches one ore more occurrences of `a` |
|
|
| 3 | `a?` | Optional | Matches zero or one occurrence of `a` |
|
|
| 3 | `a{n}` | Repetition | Matches exactly n occurrences of `a` |
|
|
| 3 | `a{n,}` | Repetition | Matches n or more occurrences of `a` |
|
|
| 3 | `a{,n}` | Repetition | Matches up to n occurrences of `a` |
|
|
| 3 | `a{n,m}` | Repetition | Matches n to m occurrences of `a` |
|
|
| 4 | `t:<expr>` | Tag | Tags the following expression with tag `t` |
|
|
| 5 | `(<expr>)` | Grouping | Groups an expression |
|
|
|
|
## License
|
|
|
|
MIT
|