Skip to content

High-pass data compression filter for inline styling (embedded CSS).

Notifications You must be signed in to change notification settings

zm-cttae/dom-inline-style-filter

Repository files navigation

dom-inline-style-filter

dom-inline-style-filter library filters inline style declarations for a standalone DOM element tree by computed effect.

  • As web developers, we would like elements that ship only with inline styling to be light so that they can be included in NPM packages.
  • A main use case of this is SVG screenshots of HTML elements.
  • Even after a filter algorithm to filter out user agent styling when inlining the style, there is some way to go with data size.

Usage

dominlinestylefilter(node)

Parameter: node - a HTMLElement with all style rules embedded as inline style attributes or <style> tags.

Returns: a Promise that resolves to node. Within node, all inline styling has been filtered to the minimum declarations that produce the same computed style.

dominlinestylefilter.sync(node)

Synchronous version. Returns node when the styling compression is completed.

Optimizations

  1. When traversing DOM tree of node, group nodes by descending node depth.

    CSS inheritance is computed on the DOM tree via preorder traversal and is additive-cumulative (increases styling data).

    For the filter op which is subtractive, we want to traverse the tree in the opposite direction to avoid more recomputation.

    The algorithm sorts elements in the node tree by descending node depth. (This is known as reverse level order traversal.)

    This gives us a 30% to 40% speed boost. This also ensures declarations are only removed when they really can be inherited.

  2. When filtering each inline style declaration by computed effect, go for the most hyphenated properties first.

    In CSS, shorthands consistently have less hyphens than their longhand.

    We want to filter out scenarios where a CSS property matches their shorthand, e.g. block-size -> height or border-color -> border.

    The algorithm does a radix sort with bitmasks for standard, custom and vendored proprties, then subsorts by descending hyphen count.

    In tests this filtered another 50% of inline styling. We also get a 20-40% speed boost because we're not setting as many properties back.

Performance

filterAuthorInlineStyles

The filterAuthorInlineStyles algorithm was proven to be a single-pass deterministic compression.

Data was collected from manual testing on the output of the domtoimage.toSvg function in the dom-to-image NPM package.

Large file inputs

$O(N)$ growth for inputs at large filesizes $|F| &gt;&gt; 1e6 \text{ bytes}$.

Wikipedia article demo Value
Number of nodes 5558 nodes
Initial declaration count 2107482 (377.54 declarations / node)
Pre-compression bytes 33.79mb
Reductions 27901890
Processing time 7921.5ms (1.85 ms/node)
Total reduction 27.90mb
Output declaration count 263099 (47.13 / node)
Post-compression bytes 5.89mb
Compression ratio 5.74

filterWinningInlineStyles

The filterWinningInlineStyles algorithm was proven to be a high-pass multi-pass - $N \approx 4$ - deterministic compression in two modes.

Data was collected from manual testing on the output of the domtoimage.toSvg function in the dom-to-image-more NPM package.

Large file inputs

$O(log(N))$ growth for inputs at large filesizes $|F| &gt;&gt; 1e6 \text{ bytes}$.

Wikipedia article demo Value
Number of nodes 5558 nodes
Initial declaration count 186985 (33.65 declarations / node)
Pre-compression bytes 3.88mb
Reductions [2970178, 95584, 16774, 0]
Processing time 7921.5ms (1.43 ms/node)
Total reduction 3.09mb
Output declaration count 83512 (15.03 / node)
Post-compression bytes 793.7kb
Compression quotients [0.5379, 0.8891, 0.9800, 1]
Compression ratio 4.88
Decay formula $1-exp(-33 / 17 \cdot N)$

Graph

Small file results

$O(c \cdot N), \space c \space \approx \space 4$ growth for inputs at small filesizes $|F| &lt;&lt; 1e6\space\text{ bytes}$.

Code screenshot demo Value
Number of nodes 420 nodes
Initial declaration count 17551 (41.79 declarations / node)
Pre-compression bytes 429160b
Reductions [298806, 40420, 0]
Processing time 5445.2ms (12.96 ms / node)
Total reduction 339226b
Post-compression bytes 89934b
Output declaration count 7637 (18.18 / node)
Compression quotients [0.696, 0.931, 1]
Total quotient (compound) 4.77
Decay formula $1-exp(-23 / 11 \cdot N)$