def main():
parser = argparse.ArgumentParser(
description=
'Extracting features as the input of LIBSVM from log. Output = logfile.instance'
)
parser.add_argument('harfolder', type=str, help='')
parser.add_argument('predictfile', type=str, help='')
args = parser.parse_args()
harfolder = args.harfolder
predicted_file = args.predictfile
# Ground truth
(all_real_pages, all_objects) = har.parse_pages_har(harfolder)
valid_urls = [i.root.url for i in all_real_pages]
# Reset nodes. One whole day wasted...Shit!@@
for node in all_objects:
node.bpointer = None
node.fpointer = []
# detected pages with SVM
(recut_pos_pages, recut_tp_pages) = get_svm_pages(all_objects, valid_urls,
predicted_file)
timetuple = pagetime(all_real_pages, recut_tp_pages)
# page timings
pagetimings = [
'{0} {1}'.format(i[0], i[1]) for i in timetuple
if i[0] > 0 and i[1] > 0
]
ofile = open('m_pagetime.txt', 'wb')
ofile.write('\n'.join(pagetimings))
ofile.close()
def main():
parser = argparse.ArgumentParser(description='Extracting features as the input of LIBSVM from log. Output = logfile.instance')
parser.add_argument('harfolder', type=str, help= '')
parser.add_argument('predictfile', type=str, help= '')
args = parser.parse_args()
harfolder = args.harfolder
predicted_file = args.predictfile
# Ground truth
(all_real_pages, all_objects) = har.parse_pages_har(harfolder)
valid_urls = [i.root.url for i in all_real_pages]
# Reset nodes. One whole day wasted...Shit!@@
for node in all_objects:
node.bpointer = None
node.fpointer = []
# detected pages with SVM
(recut_pos_pages, recut_tp_pages) = get_svm_pages(all_objects, valid_urls, predicted_file)
timetuple = pagetime(all_real_pages, recut_tp_pages)
# page timings
pagetimings = ['{0} {1}'.format(i[0], i[1]) for i in timetuple if i[0]>0 and i[1]>0]
ofile = open('m_pagetime.txt', 'wb')
ofile.write('\n'.join(pagetimings))
ofile.close()
def main():
parser = argparse.ArgumentParser(description='This program extracts real page \
urls of HAR files as groundtruth.')
parser.add_argument('harfolder', type=str, help= 'File folder containing HAR \
file(s). All the HAR files under \
this folder will be processed.')
parser.add_argument('output', type=str, help= 'Output data.')
args = parser.parse_args()
harfolder = args.harfolder
dumpfile = args.output
(all_pages, all_objs) = HAR.parse_pages_har(harfolder)
# Write to file
if os.path.exists(dumpfile):
os.remove(dumpfile)
ofile = open(dumpfile, 'wb')
i = 0 # counter
for p in all_pages:
if p.root:
i += 1
ofile.write("{0}\t{1}\n".format(p.root.url, p.root.start_time))
print('write {0} real pages to: {1}'.format(i, dumpfile))
ofile.flush()
ofile.close()
def main():
parser = argparse.ArgumentParser(
description=
'Extracting features as the input of LIBSVM from log. Output = logfile.instance'
)
parser.add_argument('harfolder', type=str, help='')
args = parser.parse_args()
harfolder = args.harfolder
# Ground truth
(all_real_pages, all_objects) = HAR.parse_pages_har(harfolder)
valid_pages = [i.root.identifier for i in all_real_pages]
for node in all_objects:
node.bpointer = None
node.fpointer = []
# detected pages with SVM
(pos_pages, tp_pages) = get_ss_pages(all_objects, valid_pages)
# objects status
print 'real pages:', len(all_real_pages)
print 'detected pages:', len(pos_pages)
(classified_right, classified_wrong, missed) = \
check_objects(all_real_pages, pos_pages)
print 'right {0} wrong {1} missed {2}'.format(len(classified_right),
len(classified_wrong),
len(missed))
def which_type(obj):
subtype_re = {
r'.*(jpeg|jpg|gif|png|bmp|ppm|pgm|pbm|pnm|tiff|exif|cgm|svg).*':
'image',
r'.*(flash|flv).*': 'flash',
r'.*(css).*': 'css',
r'.*(javascript|js).*': 'js',
r'.*(html|htm).*': 'html',
}
if obj.type != None:
for regex in subtype_re.keys():
if re.match(re.compile(regex, re.I), obj.type):
return subtype_re[regex]
else:
continue
return 'others'
def stat(objects):
html = 0
js = 0
css = 0
flash = 0
image = 0
others = 0
for obj in objects:
objtype = which_type(obj)
if objtype == 'html':
html += 1
elif objtype == 'js':
js += 1
elif objtype == 'css':
css += 1
elif objtype == 'flash':
flash += 1
elif objtype == 'image':
image += 1
elif objtype == 'others':
others += 1
print 'html {0} js {1} css {2} flash {3} image {4} others{5}'.\
format(html, js, css, flash,image,others)
stat(classified_right)
stat(classified_wrong)
stat(missed)
def main():
parser = argparse.ArgumentParser(description='Extracting features as the input of LIBSVM from log. Output = logfile.instance')
parser.add_argument('harfolder', type=str, help= '')
args = parser.parse_args()
harfolder = args.harfolder
# Ground truth
(all_real_pages, all_objects) = HAR.parse_pages_har(harfolder)
valid_pages = [i.root.identifier for i in all_real_pages]
for node in all_objects:
node.bpointer = None
node.fpointer = []
# detected pages with SVM
(pos_pages, tp_pages) = get_timetype_pages(all_objects, valid_pages)
# objects status
print 'real pages:',len(all_real_pages)
print 'detected pages:', len(pos_pages)
(classified_right, classified_wrong, missed) = \
check_objects(all_real_pages, pos_pages)
print 'right {0} wrong {1} missed {2}'.format(len(classified_right), len(classified_wrong), len(missed))
def which_type(obj):
subtype_re = {
r'.*(jpeg|jpg|gif|png|bmp|ppm|pgm|pbm|pnm|tiff|exif|cgm|svg).*': 'image',
r'.*(flash|flv).*': 'flash',
r'.*(css).*': 'css',
r'.*(javascript|js).*': 'js',
r'.*(html|htm).*': 'html',
}
if obj.type != None:
for regex in subtype_re.keys():
if re.match(re.compile(regex, re.I), obj.type):
return subtype_re[regex]
else:
continue
return 'others'
def stat(objects):
html = 0
js = 0
css = 0
flash = 0
image = 0
others = 0
for obj in objects:
objtype = which_type(obj)
if objtype == 'html':
html += 1
elif objtype == 'js':
js += 1
elif objtype =='css':
css += 1
elif objtype == 'flash':
flash += 1
elif objtype == 'image':
image += 1
elif objtype == 'others':
others += 1
print 'html {0} js {1} css {2} flash {3} image {4} others{5}'.\
format(html, js, css, flash,image,others)
stat(classified_right)
stat(classified_wrong)
stat(missed)
def main():
parser = argparse.ArgumentParser(description='Extracting features as the input of LIBSVM from log. Output = logfile.instance')
parser.add_argument('harfolder', type=str, help= '')
parser.add_argument('predictfile', type=str, help= '')
args = parser.parse_args()
harfolder = args.harfolder
predicted_file = args.predictfile
# Ground truth
(all_real_pages, all_objects) = har.parse_pages_har(harfolder)
valid_urls = [i.root.url for i in all_real_pages]
# Reset nodes. One whole day wasted...Shit!@@
for node in all_objects:
node.bpointer = None
node.fpointer = []
# Ground truth pagetime
dumpfile = 'pagetime_gt.txt'
ofile = open(dumpfile, 'wb')
for page in all_real_pages:
ofile.write(str(page.total_seconds())+'\n')
ofile.close()
# detected pages with SVM
(recut_pos_pages, recut_tp_pages) = get_svm_pages(all_objects, valid_urls, predicted_file)
# page timings
# pagetimings = [str(i.total_seconds()) for i in recut_pos_pages if i.total_seconds() > 0]
# ofile = open('pagetime_svm_pos.txt', 'wb')
# ofile.write('\n'.join(pagetimings))
# ofile.close()
# pagetimings = [str(i.total_seconds()) for i in recut_tp_pages if i.total_seconds() > 0]
# ofile = open('pagetime_svm_tp.txt', 'wb')
# ofile.write('\n'.join(pagetimings))
# ofile.close()
# objects status
(classified_right, classified_wrong, missed) = \
check_objects(all_real_pages, recut_pos_pages)
print 'right {0} wrong {1} missed {2}'.format(len(classified_right), len(classified_wrong), len(missed))
def which_type(obj):
subtype_re = {
r'.*(jpeg|jpg|gif|png|bmp|ppm|pgm|pbm|pnm|tiff|exif|cgm|svg).*': 'image',
r'.*(flash|flv).*': 'flash',
r'.*(css).*': 'css',
r'.*(javascript|js).*': 'js',
r'.*(html|htm).*': 'html',
}
if obj.type != None:
for regex in subtype_re.keys():
if re.match(re.compile(regex, re.I), obj.type):
return subtype_re[regex]
else:
continue
return 'others'
def stat(objects):
html = 0
js = 0
css = 0
flash = 0
image = 0
others = 0
for obj in objects:
objtype = which_type(obj)
if objtype == 'html':
html += 1
elif objtype == 'js':
js += 1
elif objtype =='css':
css += 1
elif objtype == 'flash':
flash += 1
elif objtype == 'image':
image += 1
elif objtype == 'others':
others += 1
print 'html {0} js {1} css {2} flash {3} image {4} others{5}'.\
format(html, js, css, flash,image,others)
stat(classified_right)
stat(classified_wrong)
stat(missed)
def main():
parser = argparse.ArgumentParser(
description=
'Extracting features as the input of LIBSVM from log. Output = logfile.instance'
)
parser.add_argument('harfolder', type=str, help='')
parser.add_argument('predictfile', type=str, help='')
args = parser.parse_args()
harfolder = args.harfolder
predicted_file = args.predictfile
# Ground truth
(all_real_pages, all_objects) = har.parse_pages_har(harfolder)
valid_urls = [i.root.url for i in all_real_pages]
# Reset nodes. One whole day wasted...Shit!@@
for node in all_objects:
node.bpointer = None
node.fpointer = []
# Ground truth pagetime
dumpfile = 'pagetime_gt.txt'
ofile = open(dumpfile, 'wb')
for page in all_real_pages:
ofile.write(str(page.total_seconds()) + '\n')
ofile.close()
# detected pages with SVM
(recut_pos_pages, recut_tp_pages) = get_svm_pages(all_objects, valid_urls,
predicted_file)
# page timings
# pagetimings = [str(i.total_seconds()) for i in recut_pos_pages if i.total_seconds() > 0]
# ofile = open('pagetime_svm_pos.txt', 'wb')
# ofile.write('\n'.join(pagetimings))
# ofile.close()
# pagetimings = [str(i.total_seconds()) for i in recut_tp_pages if i.total_seconds() > 0]
# ofile = open('pagetime_svm_tp.txt', 'wb')
# ofile.write('\n'.join(pagetimings))
# ofile.close()
# objects status
(classified_right, classified_wrong, missed) = \
check_objects(all_real_pages, recut_pos_pages)
print 'right {0} wrong {1} missed {2}'.format(len(classified_right),
len(classified_wrong),
len(missed))
def which_type(obj):
subtype_re = {
r'.*(jpeg|jpg|gif|png|bmp|ppm|pgm|pbm|pnm|tiff|exif|cgm|svg).*':
'image',
r'.*(flash|flv).*': 'flash',
r'.*(css).*': 'css',
r'.*(javascript|js).*': 'js',
r'.*(html|htm).*': 'html',
}
if obj.type != None:
for regex in subtype_re.keys():
if re.match(re.compile(regex, re.I), obj.type):
return subtype_re[regex]
else:
continue
return 'others'
def stat(objects):
html = 0
js = 0
css = 0
flash = 0
image = 0
others = 0
for obj in objects:
objtype = which_type(obj)
if objtype == 'html':
html += 1
elif objtype == 'js':
js += 1
elif objtype == 'css':
css += 1
elif objtype == 'flash':
flash += 1
elif objtype == 'image':
image += 1
elif objtype == 'others':
others += 1
print 'html {0} js {1} css {2} flash {3} image {4} others{5}'.\
format(html, js, css, flash,image,others)
stat(classified_right)
stat(classified_wrong)
stat(missed)
