Rendering PDF document using img tag with Base64 content #829

[sonnb]

The original issue from the original repository so someone can pickup and implement later: danfickle#829

I want to display PDFs and images inline with the PDF I'm generating, but I only have the raw data for each file (and their associated MIME type). For images, I can do this with:

<img 
   src="data:image/png;base64,{base64 encoded data here}"
   width="500" 
   height="500"
/>

However, when I do the same thing with a PDF (following the guidance in this issue), the PDF can't be resolved.

This is the HTML that I used:

<img 
   src="data:@file/pdf;base64,{base64 encoded data here}"
   page="1"
   width="500" 
   height="500"
/>

And the error I get is:

com.openhtmltopdf.load WARNING:: Unrecognized image format for: data:@file/pdf;base64,{base64 encoded data here} => java.io.IOException:: Unrecognized Image format

Is this expected behavior, or is this a bug with how PDFs are handled via the img tag?

Thanks!

[siegelzc]

I don't know how Daniel got the image working given a PDF file. For the data url, openhtmltopdf uses the javax.imageio package to decode your image data. We only support data types which are supported by default by this package. A data type is only supported if it returns true from javax.imageio.spi.ImageReaderSpi # canDecodeInput.

Here's a glimpse at what is supported: (inheritors of javax.imageio.spi.ImageReaderSpi)

If you really need this functionality, ImageIO has a plugin system designed to add codec capabilities for other data formats, and it looks like they can be injected into its internal registry via a scan of the classpath. See javax.imageio.spi.IIORegistry # registerApplicationClasspathSpis.

[syjer]

hi @sonnb , currently you can only render pdf file as external resource of an img element (and the file name must end with ".pdf"), see

openhtmltopdf/openhtmltopdf-pdfbox/src/main/java/com/openhtmltopdf/pdfboxout/PdfBoxReplacedElementFactory.java

Line 82 in 494a923

} else if (srcAttr.endsWith(".pdf")) {

for the special case handling of pdf as img.

In theory it should be quite easy to add base64 support for pdf too (tbh, this section of the code should be cleaned up and generalized).

[syjer]

@sonnb , I've opened a PR (#81) as the changes were not intrusive.

[sonnb]

Thanks @siegelzc and @syjer for the explanation and implementation.