Strip romaji title and groups

[AerialAtom]

So right now I'm downloading a huge amount of doujins and its taking a while because I have to edit the metadata to remove the romanji title and group names. Most of the titles I edit have a format of this - romaji title | english title. Some groups with add their names without enclosures and it shows up on the title.

Is there a way to strip the romaji titles and leave only the english titles?

Scylla na Kanojo no Konkatsu Jijou ｜ A Scylla's Journey to Marriage =Dark Mac + CW=

[Squidy]

It's possible-- You can do pretty much anything you want with custom tokens under "Settings > Save to > General > Custom Tokens". The UI for creating them isn't so great and I'd like to rework it, but it does the job. If you let me know what website it is, I could help you get started.

It might also be possible to add a setting for "cleaning up" titles, but it's difficult when there are weird edge cases like this where the group adds their name to the title in a nonstandard way. If this is at least consistent across the site in question, it would be significantly easier; I could extract the English title and set the value of the %ALT_TITLE% token to it.

[AerialAtom]

Oh forgot the website, It's Exhentai/E-hentai mostly. The ones that have an =NAME= aren't many, like 121 out of 2400 but there's a lot of split titles, 1014 out of 2400, (just used the filter to check, artist are included too).

[AerialAtom]

Ok so found a way to get rid of the group names and romaji titles but I had to use Notepad++'s find and replace on regex mode.

So I got the answer from here. and some help on /spg/
The expression that I used is:
For group names
find:
=.*?=
replace:
blank

For dual titles
find:
:.*?\|
replace:
:

If I load all of the info.txt, I can replace all the characters and just leave the English title.
I added the one for group names to the Find and Replace in HDoujin and it worked to get rid of it but the dual titles didn't work.

I tried to do this with the program but I couldn't figure it out. For the two titles separated by |, I tried to use Custom Tokens. I ran into some trouble with setting the right bound.
https://exhentai.org/g/1941710/bc95d9d739/
https://e-hentai.org/g/1941458/73c1abc15a/