Parse a 1D language

Question

Given a string containing only 0's 1's, 2's and brackets, output the grammar tree of the string.

A 2 requires 2 arguments - one to the left and one to the right

A 1 requires a single argument - to either the left or right

A 0 doesn't require any arguments and is the base case

A pair of brackets counts as one argument and the contents of the brackets are evaluated separately from the rest of the string. Nested brackets are possible

A input string will always be a complete tree with no characters falling off. The string will also only have a single correct solution. Note that the functions are commutative and any arrangement of arguments for 2 will be acceptable. You will not have to handle input that doesn't conform to these requirements.

The output grammar format will be in the form function(arguments) recursively

Test cases

0 --> 0
01 --> 1(0)
020 --> 2(0,0)
101 --> 1(1(0))
0120 --> 2(1(0),0)
0120210 --> 2(1(0),2(0,1(0)))
01210 --> 2(1(0),1(0))
(020)210 --> 2(2(0,0),1(0))
((020)20)1 --> 1(2(0,2(0,0)))

Answer 1

Python 3.6 (pre-release), 199

Saved 6 bytes thanks to Morgan Thrapp

import re
def f(s):s=s.strip('()');i,j=[m.start()if m else-1for m in(re.search(c+'(?![^(]*\))',s)for c in'21')];return i>0and f'2({f(s[:i])},{f(s[i+1:])})'or j>=0and f'1({f(s[:j])or f(s[j+1:])})'or s

Explanation & ungolfed version:

import re

def f(s):
    s=s.strip('()')
    # Search for '2' and '1' outside of brackets
    i, j = [m.start() if m else -1
            for m in (re.search(c + '(?![^(]*\))', s) for c in '21')]

    if i > 0:
        # Call `f(s[:i])` and `f(s[i+1:])`, concatenate the results
        return f'2({f(s[:i])},{f(s[i+1:])})'
    elif j>=0:
        # Call `f(s[:j])` and `f(s[j+1:])`, choose the non-empty result
        return f'1({f(s[:j]) or f(s[j+1:])})'
    else:
        return s