Andersen's analysis

andersen.py

def parse_statement(s):
    a, b = s.split(' = ')
    A, x = a[:-1], a[-1]
    B, y = b[:-1], b[-1]
    return (A, x, B, y)


def do_andersen(statements, iterations):
    statements = [parse_statement(s) for s in statements]
    pt = {}

    for _ in range(iterations):
        for s in statements:
            A, x, B, y = s
            if A == '':
                xs = {x}

            elif A == '*':
                xs = {z for z in pt[x]}

            if B == '&':
                ys = {y}

            elif B == '':
                ys = {z for z in pt[y]}

            elif B == '*':
                ys = pt[y]
                ys = set.union(*[pt[z] for z in ys])

            elif B == '**':
                ys = {z for z in pt[y]}
                ys = set.union(*[pt[z] for z in ys])
                ys = set.union(*[pt[z] for z in ys])

            for x in xs:
                if x not in pt:
                    pt[x] = set()
                pt[x] |= ys

    return pt


if __name__ == '__main__':
    statements = ['c = &b',
                  '*c = &c',
                  'a = c',
                  'c = &d',
                  '*c = **a',
                  '*c = *b',
                  '*a = &a',
                  ]
    print(do_andersen(statements, 1))
    print(do_andersen(statements, 2))
    print(do_andersen(statements, 3))