[DNSB]

That content.opf file is so far from compliant, you'd need a space telescope to set any hint of compliance. 2045 occurences of 'id="page_1"' for instance. Multiple pages of references to id for pages that do not exist. Multiple items in the main.css file with non-existent selectors such as sanad, mateen and sheer. Perhaps .sanad, .mateen and .sheer for those colour settings?

It's a bloody miracle that anything could open this abomination. The first time I've had an epub break epubcheck.

As for opening this in KOReader? The fastest answer is to use ePub ebooks that are valid ePub ebooks. Perhaps try an epub to epub conversion in calibre to see if calibre is able to fix enough to make this *&$*!@)# work with KOReader.

Edit: I've attached a calibre epub-epub conversion output to this message. The file size shrank drastically however the filename is now an ASCII transliteration of the original Arabic name.