Print number of unique occurences of clang tidy diagnostics in a given log file

clang_tidy_stats.py

#!/usr/bin/env python3

import sys
import re
from typing import Dict, Set

clang_tidy_log = sys.argv[1]

# A clang-tidy diagnostic is of the form:
# <path>:<line>:<column>: <warning|error>: <message> [<diagnostic-name(s)>]
diagnostic_re = re.compile(
    r"^([^:]+):(\d+):(\d+): (warning|error): ([^[]+) \[([^]]+)\]")

# We keep a set of per file, per line, per column (which is probably overkill) diagnostics to remove duplicates
per_file_diagnostics: Dict[str, Dict[int, Dict[int, Set[str]]]] = {}
diagnostic_counts: Dict[str, int] = {}

diagnostics_parsed: int = 0
duplicates: int = 0

with open(clang_tidy_log) as fd:
    line = fd.readline()
    while line:
        match = diagnostic_re.search(line)
        if match:
            diagnostics_parsed += 1
            (file, line, column, type, msg, diagnostic) = match.groups()
            per_file_diagnostics.setdefault(file, {}).setdefault(
                line, {}).setdefault(column, set())
            if diagnostic not in per_file_diagnostics[file][line][column]:
                diagnostic_counts.setdefault(diagnostic, 0)
                diagnostic_counts[diagnostic] += 1
                per_file_diagnostics[file][line][column].add(diagnostic)
            else:
                duplicates += 1

        line = fd.readline()

print(
    f"Found {diagnostics_parsed - duplicates} diagnostics (out of "
    f"{diagnostics_parsed} parsed; {duplicates} were duplicates).")

sorted_diagnostics = {k: v for k, v in reversed(sorted(
    diagnostic_counts.items(), key=lambda item: item[1]))}

for (diagnostic, count) in sorted_diagnostics.items():
    print(f"{count:>5}: {diagnostic}")