Add PaddleOCR-Standalone and Batch mode support

[timminator]

This PR adds support for the PaddleOCR-Standalone version I have also made.

This enables the usage of PaddleOCR without requiring the user to install python, paddlepaddle and paddleocr. Instead it prompts the user to download the standalone version, similar to when using tesseract 5.5.0 for the first time.
The PR still detects if the user installed PaddleOCR beforehand and does not prompt the user in this case to download PaddleOCR again.
Furthermore it checks if the system supports the GPU version or not. If not, the CPU version is used instead.

This PR also adds one of two desired features mentioned in #9250.

Edit: The second commit also adds Batch mode support when using PaddleOCR.

Fixes: #9250

[timminator]

Hi! @niksedk:
I think you pushed this commit in the wrong repo...
I am actually kinda puzzled that you could push a commit into my fork, but thats no problem.
I want to update my commit/add a commit and I am not sure if your added commit is now also in the Subtitle Edit repo. I saw that one change was also done in another commit there, but this change:

subtitleedit/src/libse/SubtitleFormats/Pac.cs

Lines 2772 to 2773 in a679c60

	var result = list?.Where(c => c.Value.Character == letter).Cast<KeyValuePair<int, SpecialCharacter>?>().FirstOrDefault();
	return result;

is not in the main repo. Are you aware of that?
Before I update/add another commit to my branch I wanted to ask you this...

[niksedk]

Yes, I did commit to the wrong repro... I did not know I could do that, LOL
The commit is also in the main SE branch.

I will try to look at your PR tomorrow.

Also, do you know what this is; https://github.com/PaddlePaddle/PaddleOCR/blob/main/deploy/cpp_infer/readme.md ?

[timminator]

I will try to look at your PR tomorrow.

That would be really cool. :-)

I've also added now batch mode support for PaddleOCR.
It is backwards compatible, so with the current version of paddleocr via pip (2.9.1) it is not active, but with the next version it will automatically activate it. With my standalone version it is used already.
I've also added a check that the paddleocr version has to be version 2.7 or greater due to us using the model PP-OCRv4 which is not supported on older versions. It prompts the user to upgrade paddleocr in this case.

I've had to change quite a few thing for this, but I tried to test everything to make sure that it works correctly.
I was not sure how to test the function DoBatch() - I think it is used when using SubtitleEdit from the command line? I've updated it to the best of my knowledge - but maybe you could test this one?

I'm quite happy how this turned out - I hope you like it aswell! :-)

Regarding your link: I've not seen this before and have not tried this out. Also the instructions for Windows are only in chinese...
But it seems interesting...

Edit:
I've done some speed benchmarks on 250 subtitles using the original implementation and the new batch mode. I've done them using Python and the Standalone version:

Mode	GPU (Batch)	GPU (Original)	CPU (Batch)	CPU (Original)
Python	34s	35m	1m 54s	30m
Standalone	22s	-	1m 40s	-

The Batch mode is soo much better.
Interestingly using just the CPU on the Original mode is faster than using the GPU.
The compiled Standalone version is once again a little bit faster.

[timminator]

Thank you for your quick merge! I really appreciate it! :-)

[niksedk]

Thank you for making PaddleOCR a lot better in SE :)