yavorskiy
comment-parser
comment-parser is a library helping to handle Generic JSDoc-style comments. It is
- language-agnostic – no semantics enforced. You decide what tags are and what they mean. And it can be used with any language supporting
/** */source comments. - no dependencies – it is compact and environment-agnostic, can be run on both the server and browser sides
- highly customizable – with a little code you can deeply customize how comments are parsed
- bidirectional - you can write comment blocks back to the source after updating or formatting
- strictly typed - comes with generated
d.tsdata definitions since written in TypeScript
npm install comment-parser💡 Check out the Playground
💡 Previous version lives in 0.x branch
Lib mainly provides two pieces Parser and Stringifier.
Parser
Let's go over string parsing:
const { parse } = require('comment-parser/lib')
const source = `
/**
* Description may go
* over few lines followed by @tags
* @param {string} name the name parameter
* @param {any} value the value of any type
*/`
const parsed = parse(source)Lib source code is written in TypeScript and all data shapes are conveniently available for your IDE of choice. All types described below can be found in primitives.ts
The input source is first parsed into lines, then lines split into tokens, and finally, tokens are processed into blocks of tags
Block
/**
* Description may go
* over multiple lines followed by @tags
* @param {string} name the name parameter
* @param {any} value the value parameter
*/Description
/**
* Description may go
* over multiple lines followed by @tagsTags
* @param {string} name the name parameter * @param {any} value the value parameter
*/Tokens
|line|start|delimiter|postDelimiter|tag |postTag|name |postName|type |postType|description |end|
|----|-----|---------|-------------|------|-------|-----|--------|--------|--------|--------------------------------|---|
| 0|{2} |/** | | | | | | | | | |
| 1|{3} |* |{1} | | | | | | |Description may go | |
| 2|{3} |* |{1} | | | | | | |over few lines followed by @tags| |
| 3|{3} |* |{1} |@param|{1} |name |{1} |{string}|{1} |the name parameter | |
| 4|{3} |* |{1} |@param|{1} |value|{1} |{any} |{1} |the value of any type | |
| 5|{3} | | | | | | | | | |*/ |Result
The result is an array of Block objects, see the full output on the playground
[{
// uppper text of the comment, overall block description
description: 'Description may go over multiple lines followed by @tags',
// list of block tags: @param, @param
tags: [{
// tokens.tag without "@"
tag: 'param',
// unwrapped tokens.name
name: 'name',
// unwrapped tokens.type
type: 'string',
// true, if tokens.name is [optional]
optional: false,
// default value if optional [name=default] has one
default: undefined,
// tokens.description assembled from a siongle or multiple lines
description: 'the name parameter',
// problems occured while parsing this tag section, subset of ../problems array
problems: [],
// source lines processed for extracting this tag, "slice" of the ../source item reference
source: [ ... ],
}, ... ],
// source is an array of `Line` items having the source
// line number and `Tokens` that can be assembled back into
// the line string preserving original formatting
source: [{
// source line number
number: 1,
// source line string
source: "/**",
// source line tokens
tokens: {
// indentation
start: "",
// delimiter, either '/**', '*/', '*', or ''. Mid lines may have no delimiters
delimiter: "/**",
// space between delimiter and tag
postDelimiter: "",
// tag starting with "@"
tag: "",
// space between tag and type
postTag: "",
// name with no whitespaces or "multiple words" wrapped into quotes. May occure in [name] and [name=default] forms
name: "",
// space between name and type
postName: "",
// type is has to be {wrapped} into curlies otherwise will be omitted
type: "",
// space between type and description
postType: "",
// description is basicaly rest of the line
description: "",
// closing */ marker if present
end: ""
}
}, ... ],
// problems occured while parsing the block
problems: [],
}];While .source[].tokens are not providing readable annotation information, they are essential for tracing data origins and assembling string blocks with stringify
options
interface Options {
// start count for source line numbers
startLine: number;
// escaping chars sequence marking wrapped content literal for the parser
fence: string;
// block and comment description compaction strategy
spacing: 'compact' | 'preserve';
// tokenizer functions extracting name, type, and description out of tag, see Tokenizer
tokenizers: Tokenizer[];
}examples
Stringifier
The stringifier is an important piece used by other tools updating the source code. It goes over Block.source[].tokens items and assembles them back to the string. It might be used with various transforms applied before stringifying.
const { parse, stringify, transforms: {flow, align, indent} } = require('comment-parser');
const source = `
/**
* Description may go
* over multiple lines followed by @tags
*
* @my-tag {my.type} my-name description line 1
description line 2
* description line 3
*/`;
const parsed = parse(source);
const transform = flow(align(), indent(0))
console.log(stringify(transform(parsed[0])));Result
/**
* Description may go
* over multiple lines followed by @tags
*
* @my-tag {my.type} my-name description line 1
description line 2
* description line 3
*/examples
Migrating from 0.x version
Code of pre-1.0 version is forked into 0.x and will phase out eventually. Please file the issue if you find some previously existing functionality can't be achieved with 1.x API. Check out migration notes.