textlint_fix_col.py

#!/usr/bin/env python
# See: https://github.com/textlint/textlint/issues/852
import json
import subprocess
import sys
from functools import cache
from os import getenv
from unicodedata import east_asian_width

r"""
USAGE:

Set makeprg/errorformat in your ~/.vimrc like:

> " text
> autocmd BufNewFile,BufRead *.txt setlocal filetype=text
> autocmd FileType text
>     \ setlocal
>         \ makeprg=nkf\ -w\ %\ \\\\|textlint_fix_col.py\ --stdin
>             \\ \\\\|awk\ 'BEGIN{FS=\":\";OFS=\":\"}/:/{$1=\"%\";print}'
>         \ errorformat=%f:%l:%v:\ %m\ [%t%*[a-zA-Z0-9/]],
>             \%-G,
>             \%-G%.%#\ problem,
>             \%-G%.%#\ problems
>
> " markdown
> autocmd BufNewFile,BufRead *.md,*.mkd setlocal filetype=markdown
> autocmd FileType markdown
>     \ setlocal
>         \ colorcolumn=0
>         \ spell
>         \ previewheight=3
>         \ makeprg=textlint_fix_col.py\ --plugin\ markdown
>         \ errorformat=%f:%l:%v:\ %m\ [%t%*[a-zA-Z0-9/]],
>             \%-G,
>             \%-G%.%#\ problem,
>             \%-G%.%#\ problems
"""


@cache
def read_file(path):
    with open(path, "r") as f:
        return f.readlines()


def main():
    args = sys.argv[1:]
    if "--" in args:
        index = args.index("--")
    else:
        index = len(args)
    opt_stdin = "--stdin" in args[:index]
    args = ["npx", "textlint", *args[:index], "--format", "json", *args[index:]]

    lines = None
    if opt_stdin:
        lines = sys.stdin.readlines()
        result = subprocess.run(args,
                                capture_output=True,
                                input=bytes("".join(lines), "utf-8"))
    else:
        result = subprocess.run(args,
                                capture_output=True)

    obj = json.loads(result.stdout)
    for row in obj:
        file_path = row["filePath"]

        if not opt_stdin:
            lines = read_file(file_path)

        for err in row.get("messages", []):
            line = lines[err["line"] - 1]
            col = err["column"] - 1
            binary = bytearray(line, encoding="utf-16le")[0:col * 2]

            display_col = 0
            for char in binary.decode("utf-16le"):
                if char == "\t":
                    display_col += int(getenv("TEXTLINT_TAB_WIDTH", 4))
                elif east_asian_width(char) in ("F", "W"):
                    display_col += 2
                else:
                    display_col += 1

            print("%s:%d:%d: %s [Error/%s]" % (file_path,
                                               err["line"],
                                               display_col + 1,
                                               err["message"],
                                               err["ruleId"],))


if __name__ == "__main__":
    main()