Linter for any plain text files

textlint.py

#!/usr/bin/env python
import argparse
import fnmatch
import os
import re
import sys

MIXED_TABS_SPACES = re.compile(r"^\t* +(\t+)")
TRAILING_WHITESPACE = re.compile(r"(.*)([ \t\r\f\v]+)$")

TEXTCHARS = bytearray({7,8,9,10,12,13,27} | set(range(0x20, 0x100)) - {0x7f})

parser = argparse.ArgumentParser(description='Lint general text files without parsing code.')
parser.add_argument('directory', help='directory to walk')
parser.add_argument('-e', '--exclude', action='append', help='exclude this glob pattern',
                    default=[])
args = parser.parse_args()


def is_binary_string(bytes):
    """Does the string look like binary data, i.e. non-text?"""
    return bool(bytes.translate(None, TEXTCHARS))


def mixed_indent_line(line):
    """Checks for spaces and then tabs in indentation"""
    m = MIXED_TABS_SPACES.match(line)
    if m:
        return "{}: E101 indentation contains mixed spaces and tabs".format(m.start(1))


def trailing_whitespace(line):
    """Checks for whitespace just before end of a line"""
    m = TRAILING_WHITESPACE.match(line)
    if m:
        if m.group(1):
            return "{}: W291 trailing whitespace".format(m.start(2))
        else:
            return "1: W293 blank line contains whitespace"


def final_newline(line):
    """Checks for a final newline in a file"""
    if not line.endswith('\n'):
        return "1: W292 no newline at end of file"


def mixed_indent_file(line, file_indent_type):
    """Checks for lines that don't match the file's indent style"""
    line_indent_type, result = None, None
    if line.startswith(' '):
        line_indent_type = 'space'
    elif line.startswith('\t'):
        line_indent_type = 'tab'
    if (
        file_indent_type is not None
        and line_indent_type is not None
        and line_indent_type != file_indent_type
    ):
        result = "1: E101 indentation contains mixed spaces and tabs"
    return line_indent_type, result


def output(filename, line_no, result):
    if result is not None:
        print("{0}:{1}:{2}".format(filename, line_no, result))


for dirpath, dirnames, filenames in os.walk(args.directory):
    for name in filenames:
        fullname = os.path.join(dirpath, name)
        if any(fnmatch.fnmatch(name, exclude) for exclude in args.exclude):
            continue
        with open(fullname, 'rb') as infile:
            if is_binary_string(infile.read(1024)):
                continue
            infile.seek(0)
            line_no = 1
            last_line = '\n'
            file_indent_type = None
            for line in infile:
                line_indent_type, result = mixed_indent_file(line, file_indent_type)
                if file_indent_type is None:
                    file_indent_type = line_indent_type
                results = (
                    result,
                    mixed_indent_line(line),
                    trailing_whitespace(line),
                )
                for result in results:
                    output(infile.name, line_no, result)
                line_no += 1
                last_line = line
            output(infile.name, line_no, final_newline(last_line))