word count like using `wc` using async state machine parsing. Inspired by: https://github.com/robertdavidgraham/wc2

wc.py

WAS_SPACE = 0
NEW_LINE = 1
NEW_WORD = 2
WAS_WORD = 3

SPACES = [9,10,11,12,13,32]
NEWLINE = 10

def init_table():
    # 0 => was space
    # 1 => was line
    # 2 => new word
    # 3 => was word
    table = [0,] * 1024

    # Transitions when not a space
    for c in range(256):
        table[WAS_SPACE* 256 + c] = NEW_WORD
        table[NEW_LINE * 256 + c] = NEW_WORD
        table[NEW_WORD * 256 + c] = WAS_WORD
        table[WAS_WORD * 256 + c] = WAS_WORD

    # Transitions when space
    for c in SPACES:
        table[WAS_SPACE* 256 + c] = WAS_SPACE;
        table[NEW_LINE * 256 + c] = WAS_SPACE;
        table[NEW_WORD * 256 + c] = WAS_SPACE;
        table[WAS_WORD * 256 + c] = WAS_SPACE;

    # Transitions when newline
    table[WAS_SPACE* 256 + NEWLINE] = NEW_LINE;
    table[NEW_LINE * 256 + NEWLINE] = NEW_LINE;
    table[NEW_WORD * 256 + NEWLINE] = NEW_LINE;
    table[WAS_WORD * 256 + NEWLINE] = NEW_LINE;
    return table

if __name__ == '__main__':
    # State Machine
    table = init_table()
    # Counts: num spaces, num new lines, num new words, num chars of words
    counts = [0,0,0,0]
    # Initial state
    state = WAS_SPACE

    # Read & parse
    # TODO: read in chunks for better efficiency
    with open('./test.txt', 'rb') as fd:
        for c in fd.read():
            state = table[state * 256 + c]
            counts[state] += 1

    print(f'{counts[1]} {counts[2]} {sum(counts)}')