Tomcat log file parser

tomcat_log_analyzer.py

#!/usr/bin/python
# -*- coding: UTF-8 -*-
# Version: 0.1
# Author: rad <radiofun@ya.ru>
# License: GPL 3
#

import sys
import codecs
import re
import argparse

class ErrorInfo:
    def __init__(self):
        self.dates = []
    def add_date(self, date):
        self.dates.append(date)
    @property
    def err_count(self):
        return str(len(self.dates))

def read_error(file):
    text = "";
    for line in file:
        if time_re.search(line):
            return text;
        text += line
    return text # срабатаывает для ошибки в конце файла, и если не сработал time_re то содержит текст всего лога, надо что-то делать
    
def read_errors(file_name, errors):
    counter = 0;
    file = codecs.open(file_name, mode="r", encoding="utf-8")
    old_line = None
    for line in file:
        if line.startswith("SEVERE"):
            mo = time_re.search(old_line)
            if not mo:
                continue
            date = mo.group()
            counter += 1
            error = line;
            other = read_error(file)
            if other:
                error += other
            if error:
                info = errors.get(error)
                if not info:
                    info = ErrorInfo()
                    errors[error] = info;
                info.add_date(date)
        else:
          old_line = line
          
if __name__ == "__main__":
    ap = argparse.ArgumentParser(description="Java log analyzer")
    ap.add_argument('log_files', 
                    metavar='log', 
                    type=str, 
                    nargs='+',
                    help='log files for analyzing')
    ap.add_argument('-t', 
                    dest='time_pattern', 
                    default="([\d.]+ [\d:]+)|([\w]+ \d+, \d{4} [\d:]+ [A|P]M)",
                    help='time regular expression patten (for example like "([\d.]+ [\d:]+)|([\w]+ \d+, \d{4} [\d:]+ [A|P]M)" )')
    ap.add_argument('-o', 
                    dest='out_file', 
                    help='analyze report file (XML)')
    args = ap.parse_args(sys.argv[1:])
    time_re = re.compile("^(?:" + args.time_pattern + ")", re.UNICODE)
    file_names = args.log_files
    errors = {}
    for file_name in file_names:
        read_errors(file_name, errors)
    with codecs.open(args.out_file, 'w', 'utf-8') as out:
        out.write('''<?xml-stylesheet href="#style" type="text/css"?>
<errors>
    <style id="style">
        style {
            display:none;
        }
        error, dates, date, title, text {
            display:block;
        }
        error {
            border-top: solid 1px gray;
        }
        count {
            background-color:red;
        }
        date {
            margin-left:40px;
            display: list-item;
            list-style-type:round;
        }
        text {
            font-family: monospace;
            white-space: pre;
        }
    </style>''')
        for error_text, info in errors.items():
            out.write("\t<error>\n" +
                  u"\t\t<title>Ошибка повторилась <count>" + info.err_count + u"</count> раз</title>\n" + 
                  "\t\t<dates>\n")
            for date in info.dates:
                out.write("\t\t\t<date>" + date + "</date>\n")
            out.write("\t\t</dates>\n" +
                  "\t<text><![CDATA[" + error_text + "]]></text>\n"
                  "</error>")
        out.write("</errors>\n")