Targeting manual translations to specific text paths?

[chrominance]

Working on a manual translation for a game, and I'm running into issues with some of the dialogue.

There are occasions where, in the original dialogue, a character will mention another character's game as their only line. For the translation, we might want to add punctuation or extra text to this dialogue for context, but since the original dialogue text is indistinguishable from the character's name, any extra text we add to the translation now shows up everywhere that character's name appears alone, ex. in character profile screens, party screens, etc.

For example: in dialogue, a character named Valeria tells another character that they should drop the formalities when talking; instead of Miss Valeria they should just call her Valeria, leading to the character line "Just Valeria." In the original Japanese text, however, the character just says "ヴァレリア". If I try to add a translation for ヴァレリア=Just Valeria., then any time Valeria appears in other screens, ex. the party members screen or her character profile, her name shows up as "Just Valeria." instead of "Valeria". It also means that any dialogue where Valeria is the speaker also shows her name as "Just Valeria.", ex. "Just Valeria.: Did you want me to go to the store today?"

There are a few options for translation scoping in the docs, but none of them work here. There's only one game EXE, so scoping by EXE doesn't work. This isn't an issue with screen or window resolution, so scoping by resolution doesn't work either. Scoping by layer ID seemed promising, but it turns out all story dialogue in the game uses the same layer ID, including the text boxes that display the speaker's name. So while we can fix the issue with "Just Valeria." showing up outside of dialogue, we can't fix the issue where "Just Valeria." shows up in the speaker's name textbox when reading dialogue.

Checking the text paths via EnableTextPathLogging, I can see that the speaker label and the dialogue text use different text paths, and targeting them would probably solve most of our problem; however, there's no way to scope translations that finely. Is this something we could potentially do via some other method, or by modifying the plugin code to expose this somehow?