Draft: Move translation memory entry metadata to third line

The same data is displayed per translation memory (TM) entry, but the data is ordered differently and formatted better. The target string is always bold and at the top of the entry, followed by the source string. Metadata follows on the bottom line, separated clearly by bullet points rather than spread between different lines with odd formatting, as it was before. If the sources match by >=95% then the N% match is in bold to indicate a good match, on 100% match it says "Perfect match" instead of a %. If there is no keyboard shortcut associated with the TM entry then no shortcut info is shown rather than an empty placeholder.

• Seems like it's possible to store image data in an xliff file (https://docs.oasis-open.org/xliff/v1.2/os/xliff-core.html)? Is this what the reformatting and calling cursor.insertImage() is dealing with in insertContent()? I guess I'd need xliff data with images to test this out. I've never seen it before. If this is supposed to handle images in .xliff files, then the previous TM entry target strings were not handling images in .xliffs at all and should also be changed to handle them properly (display images in the TM?? I guess?). Otherwise perhaps we can ignore whatever insertContent() was doing.