# Best Practice
Foxit PDF SDK for Web runs in a browser sandbox in a network environment. Choosing a correct website operation scheme and Foxit PDF SDK for Web configuration can make Foxit PDF SDK for Web run faster. The following section give references on website operation optimization and Foxit PDF SDK for Web configuration.
# Website assets optimization
# Gzip and Brotli compression
Compression is a way to shrink the assets size and reduce the downloading time. The following table shows the compressed size using gzip
and brotli
on UIExtension.css
and UIExtension.full.js
.
File | Original size | Gzip | Brotli |
---|---|---|---|
UIExtension.css | 1.2M | 213kb | 156kb |
UIExtension.full.js | 2.6M | 534kb | 443kb |
NOTE: Although the brotli compression algorithm provided by Google is superior to gzip in compression ratio. But brotli is not natively supported by all browsers, such as Microsoft's IE. Decompression of brotli in IE requires the use of a JavaScript engine. This time-consuming process offsets the advantages of Brotli and consumes website loading performance.
# Cache
Caching resource files can avoid downloading the same assets again and again. The /lib library in the SDK and the font files in /external are recommended for front-end caching. To learn more, check out Google (opens new window) and Mozilla (opens new window) for HTTP cache.
# Foxit PDF SDK for Web configuration
# Read only
If the following scenario is your current needs, it is recommended that you load the Foxit PDF SDK for Web read-only to improve rendering performance.
Applicable scenario:
- complex PDF documents generated by CAD
- page rendering speed is high preference
- no page editing requirements
Code Example:
<script src="path/to/UIExtension.full.js"></script>
<script src="path/to/allInOne.js"></script>
<script>
var pdfui = new UIExtension.PDFUI({
...
viewerOptions:{
customs: {
getDocPermissions: function () {
return 0;// 0 means ReadOnly
}
}
...
})
</script>
or
<script src="path/to/PDFViewCtrl.full.js"></script>
<script>
var pdfviewer = new PDFViewCtrl.PDFViewer({
...
customs: {
getDocPermissions: function () {
return 0;// 0 means ReadOnly
}
}
...
})
</script>
# Brotli compression
The core of Foxit PDF SDK for Web is the wasm / asm module compiled by emscripten. The module size is 8M / 13M, and the loading time varies depending on the browser performance. These two modules are compressed using Brotli by default. But Brotli is not natively supported by all browsers, such as Microsoft's IE, click here (opens new window) to see the browser support for Brotli. Decompressing brotli in IE needs to use the browser's JavaScript engine. This process takes time, and may offset the advantages of Brotli and then result a performance penalty.
It is recommended that you select the most suitable configuration by enabling and disabling Brotli in your test environment.
Code Example:
<script src="path/to/UIExtension.full.js"></script>
<script src="path/to/allInOne.js"></script>
<script>
var pdfui = new UIExtension.PDFUI({
...
viewerOptions:{
jr: {
brotli:{
core:false,// the default value is true which means to enable brotli,false means no brotli compression
}
}
...
})
</script>
or
<script src="path/to/PDFViewCtrl.full.js"></script>
<script>
var pdfviewer = new PDFViewCtrl.PDFViewer({
...
jr: {
brotli:{
core:false,// the default value is true which means to enable brotli,false means no brotli compression
}
}
...
})
</script>
# Preload webassembly artifacts
Starting from version 7.1.1, Foxit PDF SDK for Web provides a script file called "preload-jr-worker.js" to load webWorker scripts and wasm/asm in advance, this can greatly save document rendering time.
Code Example:
<body>
<div id="pdf-ui"></div>
<script>
var licenseSN = "Your license SN";
var licenseKey = "Your license Key";
</script>
<!-- Add the preload-jr-worker.js-->
<script src="./lib/preload-jr-worker.js"></script>
<script>
var readyWorker = preloadJrWorker({
workerPath: './lib/',
enginePath: './lib/jr-engine/gsdk',
fontPath: './external/brotli',
licenseSN: licenseSN,
licenseKey: licenseKey
})
</script>
<script src="./lib/UIExtension.full.js"></script>
<script>
var pdfui = new UIExtension.PDFUI({
viewerOptions: {
libPath: './lib', // the library path of web sdk.
jr: {
readyWorker: readyWorker,
}
},
renderTo: '#pdf-ui', // the div (id="pdf-ui").
appearance: UIExtension.appearances.adaptive,
addons: [
'```'
]
});
...
# Tiling size
Foxit PDF SDK for Web performs raster scan when rendering the page. If the currently rendered page layout is too large, the rendering speed of the page will be extremely slow. It is recommended that you enable tileSize rendering mode when opening this large page layout. Currently, the supported tileSize range is 500-3000px. In our internal comprehensive test, the rendering speed is optimal with the tileSize
being set as 1200px. But it may vary with your document complex. You can set different tileSize such as 200, 3600, and etc. in your test environment according to the needs of the actual scenario to obtain the most suitable configuration scheme.
Applicable scenario:
- Complex documents with large page layout
Code Example:
<script src="path/to/UIExtension.full.js"></script>
<script src="path/to/allInOne.js"></script>
<script>
var pdfui = new UIExtension.PDFUI({
...
viewerOptions:{
tileSize:1200,
...
}
...
})
</script>
or
<script src="path/to/PDFViewCtrl.full.js"></script>
<script>
var pdfviewer = new PDFViewCtrl.PDFViewer({
...
tileSize:1200,
...
})
</script>
# Tiling size and zoom
Foxit PDF SDK for Web opens PDF with fitWidth
by default. For large but simple documents, fitWidth zoom ratio improves rendering speed. However,for CAD type of documents with large layout and objects on a page, fitWidth may slow down the page rendering on the contrary. That is because there are more page objects are required to render in the same height of viewport -- see example below. For this type of document, the solution we recommend is to reduce the rendering object by adjusting the page zoom ratio and tileSize, thus increasing the rendering speed. You can do test in your environment to get the most appropriate configuration for your actual scenario.
Example
Take the visual range of 800 * 600, page object 3000 * 4000, and tileSize 200 as an example, let's see how many page objects are required to render in the different zoom ratios.
Zoom | Origin of Coordinate Space | Page Objects |
---|---|---|
fitWidth | the top-left corner | 3000*2400 |
0.5 | the top-left corner | 1600*1200 |
1 | the top-left corner | 800*600 |
Code Example:
<script src="path/to/PDFViewCtrl.full.js"></script>
<script>
var pdfviewer = new PDFViewCtrl.PDFViewer({
...
defaultScale = '0.5',// or '1'
tileSize:200,
...
})
</script>
# Rendering mode
Starting in version 8.5, the option annotRenderingMode has been deprecated. By default, Foxit PDF SDK for Web uses the native mode (using WebAssembly as the rendering engine) to render annotations and form controls. The native render has been significantly optimized to ensure the rendering quality and speed, so the canvas rendering is no longer required.
# Document loading
# Synchronous loading
Synchronous loading is to first obtain the complete binary stream of the file for loading, which is a compromised way of memory and performance. For documents between 50M and 500M, this method is recommended.
Code Example:
<script src="path/to/UIExtension.full.js"></script>
<script src="path/to/allInOne.js"></script>
<script>
var pdfui = new UIExtension.PDFUI({...})
var blob = getBlob();
pdfui.openPDFByFile(blob)
</script>
# Asynchronous loading
Asynchronous loading does not require a complete file stream, only the required part is obtained during loading. When the file is too large (greater than 500MB) and cannot be put in memory at all, or when you only need to request part of the document at a time, it is recommended to load the document in this way to get a good performance experience.
Code Example:
<script src="path/to/UIExtension.full.js"></script>
<script>
var pdfui = new UIExtension.PDFUI({...})
pdfui.openPDFByHttpRangeRequest({
range:{
url:'../../../docs/FoxitPDFSDKforWeb_DemoGuide.pdf',
}
})
</script>
# Loading document from memory arrayBuffer
Loading from arrayBuffer is to store the entire file stream to and load from in wasm/asm memory. For small local documents (less than 500MB) , or when the entire document stream can be obtained in a short time, it is recommended to load in this way. This method has the advantages of high reading efficiency and fast loading speed. To enable this method, pass in the callback function getLoadingMode() at the time of constructing the PDFUI. When it returns 1, it means that it is loaded from memory arrayBuffer.
Code Example:
<script src="path/to/UIExtension.full.js"></script>
<script>
var pdfui = new UIExtension.PDFUI({
...
customs:{
getLoadingMode:function(file){return 1}
}
...
})
</script>
If you have implemented your own file open control, you can use the following method to load:
var pdfui = new UIExtension.PDFUI({...})
...//event bind context
{
var arrayBuffer=getArrayBuffer();
pdfui.openPDFByFile(arrayBuffer);
}
...