You can not select more than 25 topics
Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
3.1 KiB
3.1 KiB
fastparse
A very simple and stupid parser, based on a statemachine and regular expressions.
It's not intended for complex languages. It's intended to easily write a simple parser for a simple language.
Usage
Pass a description of statemachine to the constructor. The description must be in this form:
new Parser(description)
description is {
// The key is the name of the state
// The value is an object containing possible transitions
"state-name": {
// The key is a regular expression
// If the regular expression matches the transition is executed
// The value can be "true", a other state name or a function
"a": true,
// true will make the parser stay in the current state
"b": "other-state-name",
// a string will make the parser transit to a new state
"[cde]": function(match, index, matchLength) {
// "match" will be the matched string
// "index" will be the position in the complete string
// "matchLength" will be "match.length"
// "this" will be the "context" passed to the "parse" method"
// A new state name (string) can be returned
return "other-state-name";
},
"([0-9]+)(\\.[0-9]+)?": function(match, first, second, index, matchLength) {
// groups can be used in the regular expression
// they will match to arguments "first", "second"
},
// the parser stops when it cannot match the string anymore
// order of keys is the order in which regular expressions are matched
// if the javascript runtime preserves the order of keys in an object
// (this is not standardized, but it's a de-facto standard)
}
}
The statemachine is compiled down to a single regular expression per state. So basically the parsing work is delegated to the (native) regular expression logic of the javascript runtime.
Parser.prototype.parse(initialState: String, parsedString: String, context: Object)
initialState: state where the parser starts to parse.
parsedString: the string which should be parsed.
context: an object which can be used to save state and results. Available as this in transition functions.
returns context
Example
var Parser = require("fastparse");
// A simple parser that extracts @licence ... from comments in a JS file
var parser = new Parser({
// The "source" state
"source": {
// matches comment start
"/\\*": "comment",
"//": "linecomment",
// this would be necessary for a complex language like JS
// but omitted here for simplicity
// "\"": "string1",
// "\'": "string2",
// "\/": "regexp"
},
// The "comment" state
"comment": {
"\\*/": "source",
"@licen[cs]e\\s((?:[^*\n]|\\*+[^*/\n])*)": function(match, licenseText) {
this.licences.push(licenseText.trim());
}
},
// The "linecomment" state
"linecomment": {
"\n": "source",
"@licen[cs]e\\s(.*)": function(match, licenseText) {
this.licences.push(licenseText.trim());
}
}
});
var licences = parser.parse("source", sourceCode, { licences: [] }).licences;
console.log(licences);