Parser for gnu global output

parse-gnu-global.py

import re
import urllib
import subprocess
import collections


def drop_dupes(iterator):
    return collections.OrderedDict.fromkeys(iterator).keys()

def read_lines(path):
    with open(path) as f:
        return f.read().splitlines()


def execute_global(args):
    return subprocess.check_output(["global", "--result", "cscope", "--encode-path", " "] + args)


def parse_global_output(output):
    def parse_line(line):
        sourcefile, symbol, lineno, context = line.split(' ', 3)
        sourcefile = urllib.unquote(sourcefile)
        lineno = int(lineno) - 1
        return sourcefile, symbol, lineno

    return [parse_line(line) for line in drop_dupes(output.splitlines())]


def find_symbol_matches(line, symbol):
    SYMBOL_RE = "(^|\W)%s($|\W)"

    def _match_span(match):
        span_start, span_end = match.span()
        prefix, suffix = match.groups()
        return (span_start + len(prefix), span_end - len(suffix))

    return map(_match_span, re.finditer(SYMBOL_RE % re.escape(symbol), line))


def find_references(sourcefile):
    source_lines = read_lines(sourcefile)
    output = execute_global(["-f", sourcefile, "-r"])

    references = [[] for _ in range(len(source_lines))]
    for _, symbol, lineno in parse_global_output(output):
        matches = find_symbol_matches(source_lines[lineno], symbol)
        references[lineno].append((symbol, matches))
    return references


def find_definitions(symbol):
    output = execute_global(["-d", symbol])
    return [(sourcefile, lineno) for sourcefile, _, lineno in parse_global_output(output)]


if __name__ == '__main__':
    import sys
    import pprint
    if sys.argv[1] == '-r':
        print(find_references(sys.argv[2]))
    else:
        print(find_definitions(sys.argv[1]))