1 #============================================================================ 

     2 #Name        : readHTML.py 

     3 #Part of     : Helium 

     4 

     5 #Copyright (c) 2009 Nokia Corporation and/or its subsidiary(-ies).

     6 #All rights reserved.

     7 #This component and the accompanying materials are made available

     8 #under the terms of the License "Eclipse Public License v1.0"

     9 #which accompanies this distribution, and is available

    10 #at the URL "http://www.eclipse.org/legal/epl-v10.html".

    11 #

    12 #Initial Contributors:

    13 #Nokia Corporation - initial contribution.

    14 #

    15 #Contributors:

    16 #

    17 #Description:

    18 #===============================================================================

    19 

    20 # Name: readHTML.py

    21 # Synopsis: This script creates a CSV file from the Log File Summary (.html)

    22 

    23 import htmllib

    24 import sys

    25 import formatter

    26 import re

    27 

    28 class HTMLComponent(object):

    29     """

    30     Represents a component in the log file summary

    31     """

    32     def __init__(self):

    33         self._name = ''

    34         self._errorCount = 0

    35         self._warningCount = 0

    36         

    37     def getCSV(self):

    38         return self._name + ',' + str(self._errorCount) + ',' + str(self._warningCount)

    39         

    40     def __setName(self, n):

    41         self._name = n

    42     def __setErrorCount(self, n):

    43         self._errorCount = n

    44     def __setWarningCount(self, n):

    45         self._warningCount = n

    46         

    47     name = property(None, __setName)    

    48     errorCount = property(None, __setErrorCount)

    49     warningCount = property(None, __setWarningCount)

    50 

    51 class LogHTMLParser(htmllib.HTMLParser):

    52     """

    53     Parse the scan2log HTML file into CSV

    54     """

    55     def __init__(self, verbose=0):

    56         self.anchors = {}

    57         f = formatter.NullFormatter()

    58         htmllib.HTMLParser.__init__(self, f, verbose)

    59         

    60         self.state = -1 #represents column, 0 is first

    61         self.printFlag = False

    62         self.errorCount = 0

    63         self.warningCount = 0

    64         

    65         self._components = []

    66         self.component = None

    67         

    68     def __getComponents(self):

    69         return self._components

    70         

    71     components = property(__getComponents)

    72         

    73     def handle_data(self, text):

    74         text = text.strip()

    75         

    76         #ignore plain text links that appear eg. [9]

    77         p = re.compile('\[[0-9]*\]')

    78         if not text or p.match(text):

    79             return

    80       

    81         #start of area to parse

    82         if (text == 'Component'):

    83             self.state = 0

    84   

    85         #end of area to parse

    86         if (text == 'By Command'):

    87             self.state = -1

    88           

    89         #reset column if we get lost

    90         #if (self.state > 0 and not text.isdigit()):

    91         #    self.state = 0

    92         

    93         if (self.state == 0):

    94             self.component = HTMLComponent()

    95             self.component.name = text

    96         if (text.isdigit() and self.state == 2):

    97             self.errorCount += int(text)

    98             self.component.errorCount = text

    99         if (text.isdigit() and self.state == 3):

   100             self.warningCount += int(text)

   101             self.component.warningCount = text

   102             

   103         #if there are more than/ equal 5 errors or 50 warnings we print this row

   104         if (text.isdigit() and ((self.state == 2 and (int(text) >= 5)) or 

   105           (self.state == 3 and (int(text) >= 50)))):

   106             self.printFlag = True

   107 

   108         if (self.state == 5):

   109             if (self.printFlag):                

   110                 self.components.append(self.component)

   111                 

   112             self.printFlag = False

   113           

   114         if (self.state >= 0):

   115             self.state += 1

   116             self.state %= 6

   117         

   118 def main():    

   119     if len(sys.argv) != 3:

   120         print "Usage: readHTML.pl LogFile.html errors.csv"

   121         sys.exit(1)

   122     

   123     parser = LogHTMLParser()

   124     

   125     inputFile = file( sys.argv[1], 'rb' )

   126     outFile = file( sys.argv[2], 'w' )

   127     

   128     outFile.write("Component,Errors (more than 5),Warnings (more than 50)\n")

   129 

   130     parser.feed(inputFile.read())

   131 

   132     for c in parser.components:

   133         outFile.write(c.getCSV() + "\n")

   134     

   135     outFile.write("Total," + str(parser.errorCount) + "," + str(parser.warningCount) + "\n")

   136     

   137     inputFile.close()

   138     outFile.close()

   139     parser.close()

   140         

   141 if __name__ == '__main__' :

   142     main()