finite automata – Modeling the replacement of regular expressions via a DFA

I wish to model the following common construct in code via a finite state automaton for the purpose of static analysis:

s = 
t = instead (, )
# What is the set of possible strings of t?

More formally, from the DFA $ R_1 $, $ R_2 $, and $ R_3 $, I want to find an automaton $ R_o $ it is the result of the replacement of the substrings of $ R_1 $ which are accepted by $ R_2 $ by a string that is accepted by $ R_3 $.

Unfortunately, while the literature seems to be discussing it, I can not precisely define the implementation of this operation. In addition, I have trouble defining the string set $ R_o $ should accept. Is there an algorithm to solve this problem? Is it even soluble in general?