#!/usr/bin/env python3
"""nginx 访问日志慢请求/大响应分析（只读）
用法: python3 nginx-slow.py <logfile> [host子串] [最近N行] [起始时间标记]
"""
import re
import sys

log = sys.argv[1]
host = sys.argv[2] if len(sys.argv) > 2 else ''
tail_lines = int(sys.argv[3]) if len(sys.argv) > 3 else 20000
since = sys.argv[4] if len(sys.argv) > 4 else ''

LINE = re.compile(
    r'^(?P<ip>\S+) \S+ \S+ \[(?P<time>[^\]]+)\] "(?P<req>[^"]*)" (?P<status>\d{3}) (?P<bytes>\d+)'
    r'(?: "[^"]*" "[^"]*")?(?P<rest>.*)$'
)

rows = []
with open(log, 'r', errors='replace') as fh:
    lines = fh.readlines()[-tail_lines:]

for line in lines:
    if host and host not in line:
        continue
    if since and since not in line:
        continue
    m = LINE.match(line)
    if not m:
        continue
    times = re.findall(r'(\d+\.\d+)', m.group('rest'))
    rt = float(times[-1]) if times else None
    rows.append(
        {
            'ip': m.group('ip'),
            'time': m.group('time'),
            'req': m.group('req'),
            'status': int(m.group('status')),
            'bytes': int(m.group('bytes')),
            'rt': rt,
        }
    )

print(f'匹配请求数: {len(rows)}')
if not rows:
    sys.exit(0)

withrt = [r for r in rows if r['rt'] is not None]
if withrt:
    ts = [r['rt'] for r in withrt]
    slow = [r for r in withrt if r['rt'] > 1]
    print(f'耗时: 平均 {sum(ts)/len(ts):.3f}s  中位 {sorted(ts)[len(ts)//2]:.3f}s  最大 {max(ts):.2f}s  >1s 的 {len(slow)} 个')
    print('\n== 最慢 12 个 ==')
    for r in sorted(withrt, key=lambda x: -x['rt'])[:12]:
        print(f"  {r['rt']:8.2f}s  {r['status']}  {r['bytes']:>9} B  {r['time'][12:]}  {r['req']}")

print('\n== 响应体最大 8 个 ==')
for r in sorted(rows, key=lambda x: -x['bytes'])[:8]:
    rt = f"{r['rt']:.2f}s" if r['rt'] is not None else '  ?  '
    print(f"  {r['bytes']:>9} B  {rt}  {r['status']}  {r['req']}")

from collections import Counter
c = Counter(r['req'].split(' ')[1].split('?')[0] if ' ' in r['req'] else r['req'] for r in rows)
print('\n== 路径 TOP 10 ==')
for path, n in c.most_common(10):
    print(f'  {n:5}  {path}')