I wanted to share a little a script I wrote to make a diff between two log files. The basic idea is to compare the log messages but ignore certain fields that are different every time, like the time stamp. What I do is remove every digit (0-9) and hex digit (a-f A-F) from every line and then call difflib to make the comparison. Then the original lines are presented in and html format.
So here is the script:
#!/usr/bin/python
import difflib
import sys
import re
import itertools
def charJunk (ch):
ch = re.sub (r"[A-Fa-f0-9]", "", ch)
if len(ch) == 0:
return True
else:
return False
def filter (line):
return re.sub (r"[A-Fa-f0-9]", "", line)
def printResultsHtml (result, file1Name, file2Name):
text1 = '<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" \n\
"http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd"> \n\
\n\
<html> \n\
\n\
<head> \n\
<meta http-equiv="Content-Type" \n\
content="text/html; charset=ISO-8859-1" /> \n\
<title></title> \n\
<style type="text/css"> \n\
table.diff {font-family:Courier; border:medium;} \n\
.diff_header {background-color:#e0e0e0} \n\
td.diff_header {text-align:right} \n\
.diff_next {background-color:#c0c0c0} \n\
.diff_add {background-color:#aaffaa} \n\
.diff_chg {background-color:#ffff77} \n\
.diff_sub {background-color:#ffaaaa} \n\
</style> \n\
</head>\n\
<body>\n'
table_header = '<table class="diff" id="difflib_chg_to0__top"\n\
cellspacing="0" cellpadding="0" rules="groups" >\n\
<colgroup></colgroup> <colgroup></colgroup> <colgroup></colgroup>\n\
<colgroup></colgroup> <colgroup></colgroup> <colgroup></colgroup>\n\
<thead><tr><th class="diff_next"><br /></th><th colspan="2"\n\
class="diff_header">\n'
print text1
print table_header,
print file1Name,
print '</th><th class="diff_next"><br /></th><th colspan="2" class="diff_header">',
print file2Name,
print '</th></tr></thead>',
index1 = 0
index2 = 0
for line in result:
if line.startswith ('-'):
print '<tr><td class="diff_next"></td><td class="diff_header">' + str (index1 +1) + '</td>',
print '<td nowrap="nowrap"><span class="diff_sub">' + lines1[index1] + '</span></td>',
print '<td class="diff_next"></td><td class="diff_header">' + '</td>',
print '<td nowrap="nowrap">' + '</td></tr>',
index1 += 1
elif line.startswith ('+'):
print '<tr><td class="diff_next"></td><td class="diff_header">' + '</td>',
print '<td nowrap="nowrap">' + '</td>',
print '<td class="diff_next"></td><td class="diff_header">' + str(index2 + 1) + '</td>',
print '<td nowrap="nowrap"><span class="diff_add">' + lines2[index2] + '</span></td></tr>',
index2 += 1
elif line.startswith ('?'):
pass
else:
if index1 >= len (lines1):
print "Reeeeeeeeediox"
break
else:
print '<tr><td class="diff_next"></td><td class="diff_header">' + str (index1 +1) + '</td>',
print '<td nowrap="nowrap">' + lines1[index1] + '</td>',
print '<td class="diff_next"></td><td class="diff_header">' + str (index2 +1) + '</td>',
print '<td nowrap="nowrap">' + lines2[index2] + '</td></tr>',
index1 += 1
index2 += 1
print text2
def printResultsHtml2 (result, file1Name, file2Name):
text1 = '<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" \n\
"http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd"> \n\
\n\
<html> \n\
\n\
<head> \n\
<meta http-equiv="Content-Type" \n\
content="text/html; charset=ISO-8859-1" /> \n\
<title></title> \n\
<style type="text/css"> \n\
table.diff {font-family:Courier; border:medium;} \n\
.diff_header {background-color:#e0e0e0} \n\
td.diff_header {text-align:right} \n\
.diff_next {background-color:#c0c0c0} \n\
.diff_add {background-color:#aaffaa} \n\
.diff_chg {background-color:#ffff77} \n\
.diff_sub {background-color:#ffaaaa} \n\
#one \n\
{border:1px solid red;overflow:hidden; \n\
float: left; \n\
width: 48%;overflow-x: scroll; \n\
} \n\
#two \n\
{border:1px solid red;overflow:hidden; \n\
float: left; \n\
width: 48%;overflow-x: scroll; \n\
} \n\
.c1 {width: 200px;} \n\
#wrapper \n\
{ \n\
float: left; \n\
float/**/: none; \n\
} \n\
/* easy clearing */ \n\
#wrapper:after \n\
{ \n\
content: ' + "'.'; \n" +\
'display: block; \n\
height: 0; \n\
clear: both; \n\
visibility: hidden; \n\
} \n\
#wrapper \n\
{ \n\
display: inline-block; \n\
} \n\
/*\*/ \n\
#wrapper \n\
{ \n\
display: block; \n\
} \n\
/* end easy clearing */ \n\
</style> \n\
</head>\n\
<body>\n'
text2= '<table class="diff" summary="Legends">\n\
<tr> <th colspan="2"> Legends </th> </tr>\n\
<tr> <td> <table border="" summary="Colors">\n\
<tr><th> Colors </th> </tr>\n\
<tr><td class="diff_add"> Added </td></tr>\n\
<tr><td class="diff_chg">Changed</td> </tr>\n\
<tr><td class="diff_sub">Deleted</td> </tr>\n\
</table></td>\n\
<td> <table border="" summary="Links">\n\
<tr><th colspan="2"> Links </th> </tr>\n\
<tr><td>(f)irst change</td> </tr>\n\
<tr><td>(n)ext change</td> </tr>\n\
<tr><td>(t)op</td> </tr>\n\
</table></td> </tr>\n\
</table>\n\
</body>\n\
</html>\n'
table_header = '<table class="diff" id="difflib_chg_to0__top"\n\
cellspacing="0" cellpadding="0" rules="groups" >\n\
<colgroup></colgroup> <colgroup></colgroup> <colgroup></colgroup>\n\
<colgroup></colgroup> <colgroup></colgroup> <colgroup></colgroup>\n\
<thead><tr><th class="diff_next"><br /></th><th colspan="2"\n\
class="diff_header">\n'
table_header2 = '<table class="diff" id="difflib_chg_to0__top"\n\
cellspacing="0" cellpadding="0" rules="groups" >\n\
<colgroup></colgroup> <colgroup></colgroup> <colgroup></colgroup>\n\
<colgroup></colgroup> <colgroup></colgroup> <colgroup></colgroup>\n\
<thead><tr><th class="diff_next"><br /></th><th colspan="2"\n\
class="diff_header">\n'
print text1
print '<div id="wrapper"><div id="one">'
print table_header,
print file1Name,
print '</th></tr></thead>',
r1, r2 = itertools.tee (result)
index1 = 0
index2 = 0
for line in r1:
if line.startswith ('-'):
print '<tr><td class="diff_next"></td><td class="diff_header">' + str (index1 +1) + '</td>',
print '<td nowrap="nowrap"><span class="diff_sub">' + lines1[index1] + '</span></td></tr>',
index1 += 1
elif line.startswith ('+'):
print '<tr><td class="diff_next"></td><td class="diff_header"> ' + '</td>',
print '<td nowrap="nowrap">+ ' + '</td></tr>',
index2 += 1
elif line.startswith ('?'):
pass
else:
if index1 >= len (lines1):
print "Reeeeeeeeediox"
break
else:
print '<tr><td class="diff_next"></td><td class="diff_header">' + str (index1 +1) + '</td>',
print '<td nowrap="nowrap">' + lines1[index1] + '</td></tr>',
index1 += 1
index2 += 1
print '</table></div><div id="two">'
print table_header2,
print file2Name,
print '</th></tr></thead>',
index1 = 0
index2 = 0
for line in r2:
if line.startswith ('-'):
print '<tr><td class="diff_next"></td><td class="diff_header">' + '</td>',
print '<td nowrap="nowrap">- ' + '</td></tr>',
index1 += 1
elif line.startswith ('+'):
print '<tr><td class="diff_next"></td><td class="diff_header">' + str(index2 +1 ) + '</td>',
print '<td nowrap="nowrap"><span class="diff_add">' + lines2[index2] + '</span></td></tr>',
index2 += 1
elif line.startswith ('?'):
pass
else:
if index1 >= len (lines1):
print "Reeeeeeeeediox"
break
else:
print '<tr><td class="diff_next"></td><td class="diff_header">' + str (index2 +1) + '</td>',
print '<td nowrap="nowrap">' + lines2[index2] + '</td></tr>',
index1 += 1
index2 += 1
print '</table></div></div>'
def printResults (result):
index1 = 0
index2 = 0
for line in result:
print (index1, index2),
if line.startswith ('-'):
print '- ' + lines1[index1],
index1 += 1
elif line.startswith ('+'):
print '+ ' + lines2[index2],
index2 += 1
elif line.startswith ('?'):
pass
else:
if index1 >= len (lines1):
break
else:
print ' ' + lines1[index1],
index1 += 1
index2 += 1
def printResults2 (result):
for line in result:
print line,
file1 = open (sys.argv[1])
file2 = open (sys.argv[2])
lines1 = file1.read().splitlines (1)
lines2 = file2.read().splitlines (1)
filteredLines1 = []
filteredLines2 = []
for line in lines1:
filteredLines1.append (filter (line))
for line in lines2:
filteredLines2.append ( filter (line))
d = difflib.Differ ()
results = d.compare (filteredLines1, filteredLines2)
printResultsHtml2 (results, sys.argv[1], sys.argv[2])
