linewrap

A fork of wordwrap that's faster and more powerful, supporting HTML, ANSI Color Codes, multiple paragraphing styles, and more.

On a 3.4GHz Sandy Bridge core, Linewrap achieves roughly 20MB/s when wrapping at 80 columns, or 15MB/s if wrapping at 20 columns.

Linewrap is almost backwards compatible with wordwrap. The behavior only differs in some edge cases where I believe wordwrap didn't make the best choice. You probably won't notice any difference in normal usage.

Install

npm install linewrap

Usage

    var linewrap = require('linewrap');

    // Wrap the string at 20 columns, using Windows-style line breaks.
    var wrap = linewrap(20, {lineBreak: '\r\n' /*, other options */});
    console.log(wrap('You and your whole family are made out of meat.'));

    // Wrap the string at 20 columns, prepend 10 spaces to each line, and
    // skip HTML tags when counting columns for wrapping.
    var wrap = linewrap(10, 30, {skipScheme: 'html' /*, other options */});
    console.log(wrap('You and your <b>whole family</b> are made out of <i>meat</i>.'));

Options

Skip strings

Relevant options: skip, skipScheme.

Sometimes certain characters in the text are used to control styling, annotate additional information, etc, and are not intended to be displayed. Examples include HTML tags and ANSI color codes. These characters shouldn't be counted when doing a wrap.

Supported values of skip:

1. RegExp.
2. string.

The specified regular expression or string is matched against the input, and all matching sequences in the input are simply copied to the output and are ignored by the wrapping algorithm.

skipScheme can take one of the following values: "ansi-color", "html", and "bbcode". They are pre-configured regular expressions for common tasks.

When both options are specified, skip takes precedence.

Line break strings

Relevant options: lineBreak, lineBreakScheme.

To support custom line breaks, there are actually two parameters that need to be specified: a regular expression that is used to match line breaks in the input (P1), and a string that is used as line breaks in the output (P2).

Supported values of lineBreak:

1. string. It is used as P2, and a RegExp object is created from the string to be used as P1.
2. [RegExp, string]. The RegExp object is used as P1, and the string is used as P2.
3. [string, string]. A RegExp object is created from the first string and used as P1, the second string is used as P2.
4. RegExp. It is used as P1. We will match the regular expression against the input and use the first match as P2. If no match is found, an exception is thrown. Not Recommended

You can, for example, use /\n/ as P1 and "<br>" as P2 to convert the string from one format to another.

lineBreakScheme can take one of the following values: "unix", "dos", "mac", "html", and "xhtml". Each scheme specifies both P1 and P2 for the specific scenario.

When both options are specified, lineBreak takes precedence.

Existing line breaks

Relevant option: respectLineBreaks.

This option controls how to treat existing line breaks in the input. It's important for supporting various paragraphing styles.

Supported values:

1. "all" Default. All existing line breaks are preserved.
2. "none". All existing line breaks are discarded.
3. "multi". Only 2 or more consecutive line breaks (there can be whitespaces between them) are preserved, single line breaks are discarded. This can be used to support the paragraphing style that inserts a blank line between paragraphs, so that each paragraph is re-formatted, but the paragraph structure is preserved.
4. "m<num>". A number is specified to indicate how many consecutive line breaks are preserved. For example, "multi" is equivalent to "m2".
5. "s<num>". A number is specified to indicate line breaks that are immediately followed by at least how many whitespaces are preserved. This can be used to support the paragraphing style that indents the first line of each paragraph.

Whitespaces

Relevant option: whitespace.

This option controls whether preceding and trailing whitespaces are stripped from the output. The original wordwrap isn't consistent in this area: it strips preceding whitespaces of all lines except the first one, and it strips trailing whitespaces of some lines but not others.

Supported values:

1. "default" Default. Both preceding and trailing whitespaces are stripped. This is the most similar to wordwrap's behavior.
2. "collapse". In addition to "default", also collapse consecutive whitespaces within each line.
3. "line". Similar to "default", but doesn't strip preceding whitespaces of lines preserved from the input (not wrapped by us). This option can be used with the "s<num>" options of respectLineBreaks to support the indenting paragraphing style, so that the indentations to mark new paragraphs are preserved. Preceding whitespaces are also significant in markup languages like Markdown.
4. "all". All whitespaces are preserved. In this mode, whitespaces are treated like other non-alphabetical characters that are displayed but can be wrapped at any position.

Additional indentation in preserved lines

Relevant option: preservedLineIndent.

This option applies additional indentation to lines preserved from the input. This can be used, for example, to convert from the blank-line paragraphing style to the indenting paragraphing style.

preservedLineIndent must be a non-negative integer specifying the amount of the indentation. If specified, all preserved lines will be indented by this amount.

Additional indentation in wrapped lines

Relevant option: wrapLineIndent, wrapLineIndentBase.

This option applies additional indentation to wrapped lines. This allows fine control over the alignment of wrapped text, illustrated in the following example:

Red: the color of blood, rubies
     and strawberries.
Green: the color of growing grass
       and leaves, of emeralds,
       and of jade.
Blue: the color of the clear sky
      and the deep sea.

Supported values of wrapLineIndentBase:

1. RegExp.
2. string.

wrapLineIndent must be an integer specifying the amount of the indentation. If only wrapLineIndent is specified, it must be a non-negative integer, and all wrapped lines are indented by this amount. If wrapLineIndentBase is also specified, it's searched in each preserved line: if found, all wrapped lines belonging to this preserved line are indented by the sum of wrapLineIndent and the index of the match (if the sum is a positive integer); if not found, the wrapped lines belonging to this preserved line are not indented.

For example, wrapLineIndentBase: ':', wrapLineIndent: 2 can be used to generate the result in the above example.

Hard wrapping

Relevant option: mode.

Supported values:

1. "soft" Default. Split chunks by /(\S+\s+/ and don't break up chunks which are longer than the wrap length. So if a single word is longer than the wrap length it will overflow.
2. "hard". Split chunks with /\b/ and break up chunks longer than the wrap length.

Tab width

Relevant option: tabWidth.

All \t characters are replaced with a certain number of spaces before doing the wrap. This option controls how many spaces to replace a \t. Default is 4.

Presets

Relevant option: preset.

Are you overwhelmed by the sheer amount of options? Worry not, presets are to the rescue!

Each preset contains values for one or more options. You can specify either a single preset or an array of presets. If multiple presets in the array set the same option, the last one wins.

Supported values:

1. "html". Sets skipScheme and lineBreakScheme to "html", and whitespace to "collapse".

Options that are set explicitly take predence to those set by a preset.

You are welcome to suggest new schemes and presets by creating an issue.

Acknowledgements

Thanks to James Halliday for wordwrap.

License

MIT License