ImagePreloader | Simple , fast and lightweight html-images preloader | Reactive Programming library

The extension property fox:external-document is originally meant to insert images as pages in the pdf output.

In order to use it with pdf files too you need the PDF Images Plug-In.

Just download and put (or link) the jar in the lib folder and everything should work as you expected.

It was easier than I thought. Before I call the transformation I register a costum EventListener to the User Agent of the Fop I'm using. This Listener just stores the Information what kind of Event was triggered, so I can throw an Exception if it's an ImageError.

My Listener:

You're preload function is returning back a batch of images in shape 64 x 128 x 128. It's not bothering with the channel dimension since it's 1 (grayscale). But tensorflow expects it.

I haven't used the tf.learn stuff, and it certainly seems annoying that they wouldn't return it in shape 64x128x128x1.

A quick look at the documentation didn't tell me what the return type of image_preloader is. I'm assuming X is a numpy array, in which case you just reshape it as so:

X_reshaped = np.reshape(X, (64,128,128,1))

There might be a slightly more elegant way to add the last dimension via tflearn that I don't know, but your problem is just that the last channels dimension needs to be trivially, but formally 1.

You could also reshape it using tensorflow with tf.reshape(X, [64,128,128,1]), just make sure your placeholder is accepting a shape (batch, height, width).