pixi-glyphs

Glyphs is a multi-style text component for pixi.js that supports multiple TextStyles using HTML-like tags. Includes many additional features not found in PIXI.Text such as embedded images (sprites), underlines, justified layout, and more.

Inspired by the original pixi-multistyle-text and is the spiritual successor to pixi-tagged-text. The difference in pixi-glyphs vs pixi-tagged-text is that pixi-glyphs is designed from the ground up for Pixi 8, whereas pixi-tagged-text was built for Pixi 6.

Usage

// constructor
new Glyphs(text, styles, options);

• text - A string containing the text to display. The text can be decorated with tags just like html.
• styles - An object with tag names as keys and PIXI.TextStyle+ objects as the values. The key "default" is used to set base values for all text.
• options - An object with additional confirguration options.

All parameters are optional.

Text & tags

Text

The text can be any UTF-8 string.

• Multiline strings are supported (using "\n" or multiline strings wrapped in backticks).
• Emoji's are supported.
• RTL languages are NOT officially supported but seem to work in a limited capacity.

You can get the text with tags stripped out with the .untaggedText implicit getter.

Tags

• Tags are html-like, e.g. Normal text. <bold>Bold text</bold>
• Self-closing tags are ok e.g. <img />
• Nesting tags is supported. Closing tags must match in a FILO order as opening tags. <b>bold <i>bold italic</i></b>
• Attibutes added to tags will override any existing styles on the tag. <b>bold</b> <b fontStyle="italic">bold italic</b>
• Tags that do not have a matching style are NOT treated as tags. <i><bold>just italic!</bold></i> would render as <bold>just italic!</bold> if there were a style for i but not bold.
• No tag definitions are included by default. e.g. <b> won't be bold unless you include a style for b tags.
• Use backslash to escape tag characters and display them literally: \< renders as < and \> renders as >. For example, \<b\> displays the literal text <b> instead of being parsed as a tag. Use \\ to display a literal backslash before a < or >.

Styles

styles is a plain object containing one or more keys. Each key is the name of a tag to style. 'Global' styles that apply everywhere can be set under the default key. Note, some values such as wordWrapWidth are ignored if they are included in any styles other than default.

The style objects are modified versions (supersets) of PIXI.TextStyle (referred to as TextStyleExtended in the source code). In addition to the properties allowed in TextStyle, the following extended properties are added.

(By the way, there is an excellent TextStyle visual editor that you can use to preview how your style will appear. The style objects generated can be used in this component.)

• Everything in PIXI.TextStyle including:
  • fontFamily - The font family. e.g. "Arial" or "Arial, Helvetica, sans-serif"
  • fontStyle - Font style. Options are "normal", "italic", "oblique". Default is "normal".
  • fontVariant - Font variant. Options are "normal", "small-caps". Default is "normal". Small-caps renders lowercase letters as smaller versions of capital letters.
  • fontWeight - Font weight. Options include "normal", "bold", "bolder", "lighter", "100" through "900". Default is "normal".
  • stroke - Stroke color for the text. Default is 0x000000 (black).
  • strokeThickness - Thickness of the stroke. Default is 0 (no stroke).
  • dropShadow - Enable drop shadow. Default is false.
  • dropShadowColor - Drop shadow color. Default is 0x000000.
  • dropShadowBlur - Drop shadow blur radius. Default is 0.
  • dropShadowAngle - Drop shadow angle in radians. Default is Math.PI / 6.
  • dropShadowDistance - Drop shadow distance. Default is 5.
  • dropShadowAlpha - Drop shadow alpha/opacity. Default is 1.
  • lineHeight - The height of each line of text. Can be a number or a percentage string. Default is 0 (auto).
  • leading - The space between lines of text. Default is 0.
  • lineJoin - Type of corner when drawing strokes. Options are "miter", "round", "bevel". Default is "miter".
  • miterLimit - Miter limit ratio when using "miter" lineJoin. Default is 10.
  • padding - Padding around the text. Default is 0.
  • trim - Trim transparent borders. Default is false.
  • textBaseline - The baseline of the text. Options are "alphabetic", "top", "hanging", "middle", "ideographic", "bottom". Default is "alphabetic".
  • whiteSpace - How whitespace and newlines are handled. Options are "normal", "pre", "pre-line". Default is "pre" (preserve whitespace and newlines).
  • wordWrap - Enable word wrapping. Default is true (changed from PIXI default of false).
  • wordWrapWidth - The width at which text wraps. Default is 500 (changed from PIXI default of 100).
  • breakWords - Break words if they're too long for the wordWrapWidth. Default is false.
  • fill - The fill color. Can be a string, number, or gradient. Default is 0x000000 (black).
• align - Has all the options from PIXI.TextStyle plus additional options for justified alignment:
  • "justify-left", "justify-right", and "justify-center" are all types of jutified alignment and they only differ in how they treat the last line of text.
  • "justify-all" justifies all lines of text even the last line.
  • "justify" is an alias for "justify-left".
• valign - Options are "top", "middle", "bottom", "baseline"
• textTransform - Options are "normal", "capitalize", "uppercase", "lowercase"
• fontSize - Has the added ability to use percentage-based sizes which are based on the fontSize in the parent tag. In the example: <small fontSize="10">small<big fontSize="300%">big</big></small>, "small" will be 10px and "big" will be 30px. The default fontSize is 26px.
• fontScaleWidth - Percentage to scale the font e.g. 0.5 = 50%
• fontScaleHeight - Percentage to scale the font e.g. 1.25 = 125%
• letterSpacing - Additional horizontal spacing between each character in pixels. Can be positive to spread letters apart or negative to bring them closer together. Default is 0.
• wordSpacing - Additional horizontal spacing for word separators (spaces) in pixels. Adjusts the width of spaces between words. Can be positive to increase spacing or negative to decrease spacing. Default is 0.
• lineSpacing - Additional vertical spacing between lines of text in pixels. This spacing is added between all lines within a paragraph. Default is 0.
• paragraphSpacing - Additional spacing between paragraphs that is added when you use an explicit carriage return rather than letting the text wrap at the end of a line. Default is 0. Can also be negative.
• breakLines - When breakLines is false, the text in the tag will ignore the wordWrapWidth property and never wrap to the next line unless you explicitly include a newline character. It essentially treats the whole tag as a single word. If a nested tag overrides this, only the text inside the nested tag will wrap. Default is true.
• imgSrc - ID of image to include in this tag (see imgMap under Options section)
• imgDisplay - How should the image be displayed. "block" is no scaling, "icon" scales the image to match the text-size and appear inline, "inline" displays at original size inline. Default is "inline".
• iconScale - If you use imgDisplay="icon", this value will scale the size of the icon relative to the text size. Default is 1.0 (or 100%)
• imgScale - Scales both width and height of an image by a percentage. e.g. "50%" makes the image half size.
• imgScaleX - Scales only the width of an image by a percentage. e.g. "200%" makes the image twice as wide.
• imgScaleY - Scales only the height of an image by a percentage. e.g. "75%" makes the image 3/4 of its original height.
• imgWidth - Sets the width of an image in pixels or as a percentage. e.g. 100 or "50%".
• imgHeight - Sets the height of an image in pixels or as a percentage. e.g. 50 or "150%".
• textDecoration - (i.e. underlines) Adds lines under, over, or through your text. Possible values are either "normal" or one or more of "underline", "overline", "line-through" (as a space separated string). Can also be set using the more fine-grained properties below. By default, decorations have color that matches the fill color of the text, thickness of 1 and offset 0. Note: You may need to enabled the drawWhitespace option in options to avoid seeing gaps in your text decorations between words.
• decorationColor - overrides the default color (fill) for all decorations.
• decorationThickness - overrides the default thickness (1) for all decorations.
• underlineColor - Sets the color of the underline. Default is same as fill
• underlineThickness - Sets the thickness of the underline. Default is 1.
• underlineOffset - Positions the underline above or below the default location. Default is 0.
• overlineColor - Sets the color of the overline. Default is same as fill
• overlineThickness - Sets the thickness of the overline. Default is 1.
• overlineOffset - Positions the overline above or below the default location. Default is 0.
• lineThroughColor - Sets the color of the line-through. Default is same as fill
• lineThroughThickness - Sets the thickness of the line-through. Default is 1.
• lineThroughOffset - Positions the line-through above or below the default location. Default is 0.
• adjustBaseline - Adjusts the position of the text above or below the baseline. Default is 0. Also see the adjustFontBaseline property in the options.
• topTrim - Adjusts the line height of text by trimming (or expanding) from the top of the text. Positive values reduce the line height by trimming from the top (useful for reducing gaps above large text), while negative values increase the line height. This property only affects the segments that have it set, and the overall line height is determined by the tallest effective segment after all individual topTrim adjustments. Default is 0. Example: <big topTrim="50">Big Text</big> reduces the line height of "Big Text" by 50 pixels from the top.
• highlightColor - Adds a background highlight color behind the text. Can be a hex number like 0xFFEB3B or a hex string like "#FFEB3B". The highlight appears as a continuous solid color box behind the text with no borders, spanning across spaces between words for a seamless highlight effect. When the same highlight color is applied to consecutive text segments, they will be rendered as a single continuous highlight box. Perfect for emphasizing important text, creating visual hierarchy, inline code blocks, or implementing syntax highlighting. Example: <highlight>This entire phrase is highlighted</highlight> will create one continuous highlight box.
• color - An alias for fill. It's recommended you just use either fill or color, but if both are set, fill will be used. If tags are nested, color on an inner tag can override fill in an outer tag.

Note that pixi-glyphs changes some default values from PIXI.TextStyle:

• wordWrap: false -> true
• wordWrapWidth: 100 -> 500
• fill, stroke, & dropShadowColor - "black" -> 0x000000

'Default' default styles

Some styles (fontSize, color, etc.) are set by default when you call new Glyphs() but they can all be overridden. The complete default styles are:

{
  align: "left",
  valign: "baseline",
  wordWrap: true,
  wordWrapWidth: 500,
  fill: 0x000000,
  fontSize: 26,
  stroke: 0x000000,
  dropShadowColor: 0x000000,
  imgDisplay: "inline",
  iconScale: 1.0,
  breakLines: true
}

To get a complete list, you can view the static property Glyphs.defaultStyles or look in the source code for DEFAULT_STYLES

Options

The third parameter in the Glyphs constructor is a set of options.

• debug - If true, generates debug information which is overlaid on the text during draw(). default is false.

• debugConsole - If true, logs debug information to the console during draw(). default is false.

• splitStyle - Allows you to specify how the text should be split into PIXI.Text objects when rendered. This would affect things like animations that operate on each individual piece of text within the component. Possible values are are "words" (default) and "characters".

• imgMap - An object that maps string ids like "myImage" to Sprite objects like PIXI.Sprite.from("./myImage.png"). This id will be used in the imgSrc style property to make a tag with that id render the sprite. As of v2.1.4, in addition to Sprites, you can also use any value supported by Sprite.from() or Texture.from() (including ImageHTMLElement or url strings). When a style contains imgSrc="myImage", the matching sprite is used. By default, each of the keys you provide here will automatically be added as a style in the tagStyles (equivalent to { myImage: { imgSrc: "myImage"}}) so you can add a tag <myImage />. default is {}.

  Note on icon scaling: pixi-glyphs automatically enables linear filtering and mipmaps on textures to ensure icons scale smoothly. However, if a texture has already been rendered to the GPU before being passed to pixi-glyphs, mipmaps may not generate correctly. If you experience pixelated icons, load your textures with mipmaps enabled:

  const texture = await PIXI.Assets.load({
    src: 'icon.png',
    data: { scaleMode: 'linear', autoGenerateMipmaps: true }
  });
  const icon = new PIXI.Sprite(texture);

• adjustFontBaseline - For fonts that do not align correctly with the baseline, this adjusts the position of the text relative to the baseline. This should be an object with font names for keys and a value which is a string percentage (of the font's ascent, or height above the baseline) or a numerical value in pixels to adjust the offset. E.g. {"arial": "80%"}. Since this value changes for each fontFamily and for each fontSize it's recomended that you use the percentage values. Default is no adjustments.

• scaleIcons - When true, images in the imgMap that use imgDisplay: icon will scale with the text when fontScaleWidth or fontScaleHeight are set. Default is true.

• drawWhitespace - When true, whitespace characters are rendered as their own Text objects. Default is false.

• wrapEmoji - When true, emoji characters are automatically wrapped in a special tag <__EMOJI__> which allows the user to specify styles for emoji only. A base style is automatically defined for __EMOJI__ that sets the font-family to sans-serif. This helps to solve some issues with rendering on certain fonts that don't handle emojis well. Default is true. (This feature may be removed in future updates)

• skipUpdates - When true, update() will not be called when text or styles are changed; it must be called explicitly or overridden using the skipUpdate parameter in functions such as setText(). default is false.

• skipDraw - When true, draw() will not be called by update() it must be called explicitly or overridden using the skipDraw parameter, e.g. myGlyphs.update(false). default is false.

• supressConsole - When true, prevents warnings and other messages from being logged to the console, however, this does not affect debugConsole.

• errorHandler - A handler function that will receive non-fatal warnings and errors generated by the component. This will not catch any exceptions, only internal messaging such as when you try use an unknown tag. The handler should be in the format (e:ErrorMessage) => void where ErrorMessage is an object with the shape { code: string, message: string, type: "warning" | "error" }.

• overdrawDecorations - When using text decorations (e.g. underline) an additional length in pixels equal to overdrawDecorations will be added to both sides of the line. Default is 0.

• lineSpacing - Sets the default line spacing for all text. Can also be set per-tag in styles. Default is 0.

• paragraphSpacing - Sets the default paragraph spacing for all text. Can also be set per-tag in styles. Default is 0.

To see a list of the default options, you can view the static property Glyphs.defaultOptions or look in the source code for DEFAULT_OPTIONS

A note about the component architecture

Glyphs allows you to control text with multiple styles and embedded images as though it were a single Pixi component. It was inspired by pixi-multistyle-text but is structurally very different.

pixi-multistyle-text composes bitmap snapshots of text objects into a single canvas. Conversely, pixi-glyphs creates a separate PIXI.Text component for each word, or word segment. Using multiple Text components allows developers to have control over individual words or even characters for the purposes of animation or other effects. It also makes embedding sprites into the layout easier. Cosmetically, they're very similar, however, the overhead of creating multiple Text objects is much larger potentially making Glyphs a heavier slower component.

Another similar component is @pixi/text-html which renders a bitmap from an HTML element using the browser's native rendering. While pixi-glyphs should theoretically render more consistently cross-browsers, this could be a good option if pixel perfection and cross-browser support is not a concern.

Child DisplayObjects

Glyphs generates multiple display objects when it renders the text with draw(). Developers can access these children if desired for example to add additional interactivity or to animated individual elements.

Please note that by default, these are recreated every time text or style properties change (technically, whenever draw() is called). Manipulating the children directly may cause your view to become out of sync with the text and styles properties.

These properties are available:

• textContainer - The Container layer which holds all the text fields rendered by draw.
• spriteContainer - The Container layer which holds all the sprites rendered by draw if you're using an image map (imgMap).
• debugContainer - The SpContainerite layer which holds all debug overlay information (if you're using the debug: true setting).
• decorationContainer - The Container layer which holds all text decorations (underlines).
• textFields - An array containing all the text fields generated by draw.
• sprites - If you're using an image map (imgMap), this array stores references to all the Sprites generated by draw.
• spriteTemplates - The sprites in sprites and spriteContainer are actually clones of the originals passed in via the imgMap option. To get the originals, access them this way.
• decorations - Array of Graphic objects which render the text decorations.

Life-cycle & optional updates

In order to maximize performance of Glyphs, it helps to understand how it renders the input.

1. Constructor - creating the component is the first step. You can set text, styles, and options in the constructor.
2. update() - Update generates a list of plain JS object tokens that hold information on what type of text to create and where to position it. The tokens contain all the information you need to draw the text and are saved as the instance member tokens but also returned by the update() method. Aside from decoupling from the render code, this allows us to write tests to verify every step of the lexing, styling and layout independently without drawing anything at all. By default, this is called every time the text or style definitions are changed (e.g. setTagStyles(), setText()). This is a fairly expensive process but usually faster than draw().
3. draw() - Draw creates the child objects based on the data generated by update(). It clears any existing children (if needed) then recreates and positions them. This is probably the costliest method in the life-cycle. By default, this is called automatically by update().
4. Of course, you won't see anything on your screen until your component is added to a visible PIXI container that's part of the stage in a pixi app.

The methods that normally trigger an update are:

• setText() & myText.text = (implicit setter)
• setTagStyles() & tagStyles = (implicit setter)
• setStyleForTag()
• setDefaultStyle() & defaultStyle = (implicit setter)
• removeStylesForTag()

The methods that normally trigger a draw:

• update()

Please note that direct changes to styles or other objects will not trigger an automatic update unless you use one of the above methods. For example:

const t = new Glyphs("<big>Big text</big>", { big: { fontSize: 25 } }); // renders "Big text" at 25px

t.getStyleForTag("big").fontSize = 100; // The change to the style wasn't detected. It still renders "Big text" at 25px

t.update(); // now it renders correctly.

t.textFields[0].visible = false; // Makes the word "Big" disappear.

t.draw(); // recreates the text fields restoring "Big"

skipUpdates & skipDraw

If performance is becoming an issue, you can use the skipUpdates and skipDraw flags in the options object with new Glyphs() to disable automatic updates and automatic drawing (more on that below). This gives you control over when the update() or draw() function will be called. However, the component can become out of sync with what you see on your screen so use this with caution.

Several other individual functions, such as setText() also give you the option to skipUpdate on an as needed basis.

// Create a new Glyphs but disable automatic updates and draws.
const t = new Glyphs("", {}, {skipUpdate: true, skipDraw: true});
const words = ["lorem", "ipsum", ... ];
// add words until the length of text is > 500 characters.
while (t.untaggedText.length <= 500) {
  t.text += words[i];
}

// Normally, update() will draw() also, but we've disabled that.
// t.tokens will be up dated to match the new text but it will not appear on the screen.
t.update();
t.textContainer.length; // 0 - text fields never got created.

// Manually call draw to generate the PIXI.Text fields
t.draw();
t.textContainer.length; // This will now contain all the PIXI.Text objects created by draw.

Destructor

Like with other Pixi components, when you're ready to remove your Glyphs object from the stage and make it ready for garbage collection, use the .destroy() method. Unlike some other Pixi components, Glyphs will automatically destroy its child objects, such as fragments of text and debug drawings (with the exception of imgMap, see below). Child objects can still be retained by passing options to the destructor, e.g. .destroy(false).

Destroying imgMap sources

destroy() will not by default destroy the textures used for the imgMap since these are often shared between multiple Glyphs objects.

To destroy the source textures in imgMap use the method .destroyImgMap(). Please note that this must be called before destroying the Glyphs instance and will throw an error if you try to call it on an already-destroyed instance.

Contributing

If you'd like to contribute, a great place to start is to log an issue. If you are interested in bug-fixing, please take a look at the issues list and notify one of the owners that you're interested in working on it. Please make sure to test the code you write! If you're not sure how, just ask for help!

Build instructions

npm install
npm run build

The node-canvas package, used for testing, has some additional dependencies. For Mac users, there's a homebrew Brewfile that you can install using brew bundle. For other users, see the instructions for the package. On Apple Silicon, this issue was helpful with troubleshooting.

npm should automatically install peer dependencies (including the very important pixi.js) but in my experience, you may have to run npm install again after adding any additional packages.

VSCode Users

If you're using the vscode-jest extension, you may need some additional packages to get the tests to run in your IDE. If you're on a Mac you can use brew bundle install to install these packages.

Demo

You can view examples using the command:

npm run docs

This will start a simple HTTP server running locally on port 8080. Navigate to http://localhost:8080/demo

Dependencies

The build process is slightly complex and some of the build dependencies are at the max major version that doesn't break. These depdencies seem to cause issues at certain versions...

| package | max version | | -------------------------- | ----------- | | jest, @types/jest, ts-jest | 26 | | emoji-regex | 9 |

License

MIT, see LICENSE.md for details.