Reformats piped mypy output to group by the error code.

mypy_groupby.py

"""
Reformats piped mypy output to group by the error code.

Author:  Michael Osthege
License: MIT

Usage
-----
mypy -p mypackage --show-error-codes | python mypy_groupby.py
"""
import pandas
import sys
from typing import Iterator


def mypy_to_pandas(input_lines: Iterator[str]) -> pandas.DataFrame:
    current_section = None
    data = {
        "file": [],
        "line": [],
        "type": [],
        "section": [],
        "message": [],
    }
    for line in input_lines:
        line = line.strip()
        elems = line.split(":")
        if len(elems) < 3:
            continue
        try:
            file, lineno, message_type, *_ = elems[0:3]
            message_type = message_type.strip()
            if message_type == "error":
                current_section = line.split("  [")[-1][:-1]
            message = line.replace(f"{file}:{lineno}: {message_type}: ", "").replace(
                f"  [{current_section}]", ""
            )
            data["file"].append(file)
            data["line"].append(lineno)
            data["type"].append(message_type)
            data["section"].append(current_section)
            data["message"].append(message)
        except Exception as ex:
            print(elems)
            print(ex)
    return pandas.DataFrame(data=data).set_index(["file", "line"])


if __name__ == "__main__":
    df = mypy_to_pandas(sys.stdin)
    for section, sdf in df.reset_index().groupby("section"):
        print(f"\n\n[{section}]")
        for row in sdf.itertuples():
            print(f"{row.file}:{row.line}: {row.type}: {row.message}")
    print()
    print(f"{len(df)} rows in total.")