Exercise 2 solution

    import re
    from collections import Counter

    log_re = re.compile(r"""
        (?P<date>\d{4}-\d{2}-\d{2})\s
        (?P<time>\d{2}:\d{2}:\d{2})\s
        (?P<level>[A-Z]+)\s
        \[(?P<source>\w+)\]\s
        (?P<msg>.*)
    """, re.X)
    kv_re = re.compile(r"(\w+)=(\S+)")

    def parse_line(line):
        m = log_re.fullmatch(line)
        if not m:
            return None
        rec = m.groupdict()
        rec["fields"] = dict(kv_re.findall(rec.pop("msg")))
        return rec

    def parse_log(text):
        records, bad = [], []
        for n, line in enumerate(text.splitlines(), 1):
            rec = parse_line(line)
            if rec:
                records.append(rec)
            else:
                bad.append((n, line))
        return records, bad

    text = """\
    2026-11-02 10:00:01 INFO [web] start
    2026-11-02 10:00:02 WARN [db] slow
    2026-11-02 10:00:03 INFO [web] ok
    2026-11-02 10:00:04 ERROR [web] boom
    2026-11-02 10:00:05 WARN [web] slow
    2026-11-02 10:00:06 INFO [db] ok
    """
    recs, _ = parse_log(text)
    c = Counter(r["level"] for r in recs)
    for lvl, n in sorted(c.items(), key=lambda kv: (-kv[1], kv[0])):
        print(lvl, n)
    # INFO 3
    # WARN 2
    # ERROR 1

Note: when you paste the log text, the lines must start at the left margin
(the indentation above is only for this file). A line with leading spaces
would not match, and parse_log would report it as skipped.

(All lines were run, with the log text at the left margin, and printed
exactly what is shown.)

Explanation: the sort key (-count, name) sorts by count descending and then
by name ascending, because negating the count reverses only that part.
Counter.most_common() would also sort by count, but it does not promise any
particular order for ties, so an explicit key is safer.