Added read/write support for image blobs

[yannickulrich]

This adds support for the native image insertions in notebooks on 3.27 or later. It is similar to PR #51 but parses the whole object rather than just the filename. The construction is based on remarkable-shapes which already has this ability in ruby.

On the top-level, images are defined in a SceneImageInfoBlock (SceneAssetBlock in #51, block type 0x0E) via a UUID. They can have a flag which currently always evaluates to [17, 0]. They are then references in the tree as a SceneImageItemBlock (ScenePathItemBlock in #51, block type 0x0F, item type 0x07). This uses the UUID defined earlier and includes image placement as a list of image vertices (x, y, u, v tuples)

[
  left, top, 0.0, 0.0,
  right, top, 1.0, 0.0,
  right, bottom, 1.0, 1.0,
  left, bottom, 0.0, 1.0
]

I have so far made no attempt at parsing this further since it should probably be done by rmc instead. The image block also contains a magic sequence of integers.

I have verified that this can read and write files generated by the remarkable and the remarkable-shapes library but haven't added any pytests yet.

Side note: I've also added a read_lww_bytes function for the image flag and the UUID when referenced in the SceneImageItemBlock. There's probably a better way of doing this but I couldn't find one quickly.