nginx日志访问量，多维度统计，__fields表示每条日志以空格分隔后索引对应的字段名称。


如0表示第一个字段客户端IP（client_ip），要统计其它字段，只需在这里添加即可。


## 直接上代码

# -*- coding:utf-8 -*-

import sys

try:
    import json
except ImportError:
    import simplejson as json


class AnalysisNginxLog(object):
    ''' analysis nginx proxy access log

    proxy log format:
        log_format  proxy  '$remote_addr - $host [$time_local] "$request" $status'
                        ' $body_bytes_sent "$http_referer" '
                         '"$http_user_agent" "$http_x_forwarded_for/$upstream_addr" "$request_time/$upstream_response_time/$upstream_status/$userId" "$xm_logid"';
    '''

    __fields = [
        (0, 'client_ip'),
        (2, 'domain'),
        (3, 'access_time'),
        # (4, 'time_zone'),
        (5, 'http_method'),
        (6, 'uri'),
        (7, 'http_version'),
        (8, 'request_status'),
        (9, 'request_body_size')
    ]

    __filter_args = [
        'uri',
        'access_time'
    ]

    def __init__(self, logfile):
        self.logfile = logfile
        self.data = []

        self._load_data()

    def _open(self):
        with open(self.logfile) as fp:
            while True:
                line = fp.readline().strip()
                if not line:
                    break
                yield line

    def get_data(self):
        self._load_data()
        return self.data

    def _filter1(self, data):
        return data.strip().strip('#').strip('"')

    def _filter2(self, data):
        if not isinstance(data, dict):
            raise Exception('Invalid Parameters, must be type of dict!')

        for key, value in data.items():
            if key not in self.__filter_args:
                continue

            if '?' in value:
                v = value.split('?')[0]
                data[key] = v
            elif '[' in value:
                data[key] = value.strip('[')

        return data

    def _load_data(self):
        for line in self._open():
            li = [x.strip() for x in line.split()]

            dict_tmp = {}
            for i, field in self.__fields:
                try:
                    dict_tmp[field] = self._filter1(li[i])
                except IndexError as e:
                    print('Error: {}, at {}'.format(
                        e, line
                    ))
                    dict_tmp[field] = None

            self.data.append(self._filter2(dict_tmp))

    def counter_by_field(self, field, top):
        fields = [x[1] for x in self.__fields]

        if field not in fields:
            raise Exception('Invalid Parameters!')

        result = {}

        for log in self.data:
            if log[field] in result:
                result[log[field]] += 1
            else:
                result[log[field]] = 1

        result = sorted(result.items(), key=lambda d: d[1], reverse=True)[:top]

        print(json.dumps(result, indent=2))
        return result


def _help():
    print('Usage: %s [log_file] [colums] [topN] ' % sys.argv[0])
    sys.exit()


def counter_log(logfile, colums, top=10):
    nginx = AnalysisNginxLog(logfile)

    cols = colums.split(',')
    for col in cols:
        print('{c} [{col}] Top {top} {c}'.format(c='-' * 30, col=col, top=top))
        nginx.counter_by_field(col, int(top))


if __name__ == '__main__':

    if len(sys.argv) <= 2:
        _help()

    counter_log(*sys.argv[1:])



## 运行结果

# > python analysis_nginx_proxy_log.py
Usage: analysis_nginx_proxy_log.py [log_file] [colums] [topN]


> python analysis_nginx_proxy_log.py pss_00-06.log uri,client_ip 3
------------------------------ [uri] Top 3 ------------------------------
[
  [
    "/api/pss/GetBatchSnDetails", 
    2298
  ], 
  [
    "/api/xmss/CheckSubHome", 
    1456
  ], 
  [
    "/api/xmss/GetPhyStock", 
    1299
  ]
]
------------------------------ [client_ip] Top 3 ------------------------------
[
  [
    "10.108.38.18", 
    2150
  ], 
  [
    "10.114.135.253", 
    1403
  ], 
  [
    "10.104.3.11", 
    1253
  ]
]