ignore errors when translating large files, save the intermediate result

[TaunT]

Several times when translating a CBZ file, the program crashed with an error and stopped translation. This is very unpleasant, considering that each request, for example, to GPT4 costs money and the intermediate result of translating 10-20 pages was not saved anywhere. Can you make an option to intercept such errors? If some page is not successful, then the program ignores it, leaves its original version and moves on, and then at the end it writes which pages are untranslated so that you can complete them one by one. Otherwise, now I have to manually do everything one image at a time (((