External Images

[Irmitya]

I was finding out how to export images, and after getting an image "extracted", I wanted to know how duplicate filenames would be handled and nested folders.
So, I opened a .pur file that's composed of images somewhere on my computer, and found that NONE of the images were loaded in the purformat.images list,

import PIL, io
PIL.Image.open(io.BytesIO(img.pngBinary)).save(fR"C:\purtoimage\{img.transforms[0].name}.png")

I use pureref as a notes/bookmarking tool, and wanted to use this to finally be able to automate it but without any usage of external images, I can only use it as an extractor.

[FyorDev]

I've just realized I never took local images into account in the case of the 'embed local images in save file' setting in PureRef being turned off. Meaning the image is not packed and saved but only linked in the file, an option I've always had on and didn't consider.
I'll add functionality for this in a while, next week after exams. If this is indeed the issue, you can circumvent this for now by saving the file (I'd suggest under a different name) with the 'embed local images in save file' setting in PureRef turned on. Be sure to turn it off again to avoid unlinking your other files upon saving.

For files with packed images, this piece of code extracts an image from a PureRef file.

import purformat
from PIL import Image
import io

purobj = purformat.PurFile()
purobj.read("NewScene.pur")
img = purobj.images[0]

Image.open(io.BytesIO(img.pngBinary)).save(fR"{img.transforms[0].name}.png")

If this doesn't work for you I'd like to have a file that has this issue so I can fix it.

About nested folders and duplicate filenames, this depends entirely on how you handle file saving. If you save different images with the same name in the same place they would be overwritten. If this is an issue you could check if the file already exists or use the PurGraphicsImageItem source.

[Irmitya]

I already have the code setup for extraction.
I just got stopped at the file things because the name of one of my files had at least one illegal character ?, and while I could just swap/remove all illegal characters \:*?"<>|, I could probably break something if the character was relevant to the file's location rather than name.

I'm not in a hurry to do anything; I need to have code/scripts ready with a new process validated to migrate from what I currently do.

My first goal is to detach images from .pur files, so that I can access/change the files elsewhere, as well as stop it from taking so long to save (dozens-of-seconds to several-minutes).
My second goal is to be able to run a script to autoupdate text on my bookmarks....which while I was testing, I found a problem #3 that makes the first task more cumbersome and the second non-viable.

[FyorDev]

I'm aware this was not the original issue, but unpacked external image links are now supported since 3b680e6. External image links have 0xFFFFFFFF as the binary and the source is used as the link. You can now read and write a file while retaining links, or add new images with links yourself.