I have the following situation:

In Polish the name of the color will get a different ending, depending on the gender. So I have for example for "red" "czerwony, czerwona, czerwone".

This makes it impossible to translate red to one of those in a highly repetitive file, where I have partially >2000 such position. To make this translation fast, I translated all colors like this "red = kolor: czerwony" (which simply means adding the word "color" before each color name, what makes this repetitive again, as then the name of the color does not depend on gender). This was a good idea, until the customer has changed his file. Now in some places I have this structure:

Part name
color name

which translates perfectly to

Part name
color: name

But then he added

Part name
Color
color name

and I cannot now translate like above

Part name
Color
color: name

Using the advanced display filter I could display all segments, where I have in source either color or color name (for example (Color|red|blue<yellow<brown) in source and "^color: ..." in Polish. But the problem is, I need only to change this segments, where "Color" is followed by "color: ...".

But in my view sometimes the segments do not really follow each other.

Looks like this:

Does anyone have an idea, how to automate this?