run_spell_checks

#!/usr/bin/python3

import argparse
import os
import subprocess
import sys
import tempfile


IGNORED_FILES = [r"'\.(pdf|xz)$'", r"^po/", r"LICENSE"]


def run_command(command, cwd=None):
    res = subprocess.Popen(command, shell=True, stdout=subprocess.PIPE,
                           stderr=subprocess.PIPE, cwd=cwd)

    out, err = res.communicate()
    return (res.returncode, out.decode().strip(), err.decode().strip())



def write_file(filename, content):
    with open(filename, 'w+', encoding='utf-8') as f:
        f.write(content)


def _run_spellintian(ignored, cwd=None):
    ignore = " ".join(["-e " + i for i in IGNORED_FILES])
    ret, out, err = run_command("git ls-tree -rz --name-only HEAD | grep -Evz %s | xargs -r0 spellintian | grep -v \"(duplicate word)\"" % ignore,
                                cwd=cwd)
    if ret != 0:
        raise RuntimeError("Failed to run codespell: %s %s" % (out, err))
    errors = out.split("\n")
    return [e for e in errors if ignored and any(i for i in ignored if i not in e)]


def _run_codespell(ignored, cwd=None):
    ignored_files = " ".join(["-e " + i for i in IGNORED_FILES])
    ignored_words = "-L " + ",".join(ignored) if ignored else ""
    _ret, out, err = run_command("git ls-tree -rz --name-only HEAD | grep -Evz %s | xargs -r0 codespell %s" % (ignored_files, ignored_words),
                                cwd=cwd)
    if err:
        raise RuntimeError("Failed to run codespell: %s" % err)
    return out.strip().split("\n")


def _check_stable(project, ignored):
    errors = ["", ""]
    with tempfile.TemporaryDirectory() as dname:
        ret, _out, err = run_command("git clone https://github.com/storaged-project/%s" % project,
                                     cwd=dname)
        if ret != 0:
            raise RuntimeError("Failed to clone %s: %s" % (project, err))

        errors[0] = _run_spellintian(ignored, os.path.join(dname, project))
        errors[1] = _run_codespell(ignored, os.path.join(dname, project))

    return errors


def _check_latest(ignored):
    errors = ["", ""]

    errors[0] = _run_spellintian(ignored)
    errors[1] = _run_codespell(ignored)

    return errors


def run_checks(project, ignored):
    suc = True

    stable_errors = _check_stable(project, ignored)
    latest_errors = _check_latest(ignored)

    lintian_new = [e for e in latest_errors[0] if e and not e in stable_errors[0]]
    if lintian_new:
        print("New spelling errors found by lintian:\n%s\n" % "\n".join(lintian_new))
        suc = False

    codespell_new = [e for e in latest_errors[1] if e and not e in stable_errors[1]]
    if codespell_new:
        print("New spelling errors found by codespell:\n%s\n" % "\n".join(codespell_new))
        suc = False

    return suc


if __name__ == '__main__':

    argparser = argparse.ArgumentParser(description='spell checker')
    argparser.add_argument('-p', '--project', dest='project', action='store',
                           help='project to run tests for', required=True)
    argparser.add_argument('-i', '--ignore', dest='ignore', nargs='*', action='store',
                           help='ignored words')
    args = argparser.parse_args()

    suc = run_checks(args.project, args.ignore)

    sys.exit(0 if suc else 1)