def classify(htmlSeries, sm):
global BASEFILESTORAGEDIR
# Make the useless folder page.
uselessPagesFolder = chkmkFolderStructure(BASEFILESTORAGEDIR + '/useless/')
# List all the files, in the folder.
listOfFiles = os.listdir(BASEFILESTORAGEDIR)
listOfFiles = [BASEFILESTORAGEDIR + p for p in listOfFiles]
# Now start the loop and process every file
for l in range(0, len(listOfFiles)):
# Choose a file randomly.
page = random.choice(listOfFiles)
# Extract the content of the file
c = gzip.open(page, 'rb')
contents = c.read()
c.close()
# Write to a tmp file.
tmpFilename = '/tmp/' + page.split('/')[-1]
f = file(tmpFilename, 'w')
f.write(contents)
f.close()
# Generate html series of this file, tphs --> testPageHtmlSeries
tphsUrl = 'file://' + tmpFilename
tphs = KastParsersLib.html2TagSignal(tphsUrl)
# dftDistance scoreboard
dftDistanceScoreboard = []
for d in htmlSeries:
# Now calculate the score and append them to an array.
dftDistanceScoreboard.append(KastParsersLib.dftDistance(tphs, d))
# Now calculate average.
s = KastGenericFunctionsLib.calcAvg(dftDistanceScoreboard)
# Score is less than mean similarity measure, move it to the useless folder.
if s < sm:
os.system(page, uselessPagesFolder)
def main(targetWebsite, configFile):
global unseenUrlList
global BASELOGDIR
global BASELOCKFILEDIR
global BASEFILESTORAGEDIR
global BASEERRORLOGDIR
global BASECONTENTDIR
global contentLogFile
global mode
# Extract website name
sitename = KastGenericFunctionsLib.extractWebSiteName(targetWebsite)
# First generate the folder structure if its does not exist.
BASELOGDIR = KastGenericFunctionsLib.chkmkFolderStructure(BASELOGDIR)
BASELOCKFILEDIR = KastGenericFunctionsLib.chkmkFolderStructure(BASELOCKFILEDIR)
BASEFILESTORAGEDIR = KastGenericFunctionsLib.chkmkFolderStructure(BASEFILESTORAGEDIR + sitename + '/')
BASEERRORLOGDIR = KastGenericFunctionsLib.chkmkFolderStructure(BASEERRORLOGDIR)
BASECONTENTDIR = KastGenericFunctionsLib.chkmkFolderStructure(BASECONTENTDIR)
# Now generate the task/target specific filenames.
lockFile = BASELOCKFILEDIR + sitename + '.lock'
errorLog = BASEERRORLOGDIR + sitename + '.error'
contentLogFile = BASECONTENTDIR + sitename + '-' + str(round(time.time(), 2))
# Now check if the lock file exists and proceed with crawling.
if os.path.exists(lockFile):
KastGenericFunctionsLib.logException(sitename + ' crawl in progress - Exiting - ' + str(time.time()), BASELOGDIR + sitename + '.exit.log')
sys.exit(-1)
# Make a lock file.
if mode == 'p':
lf = file(lockFile, 'w')
lf.close()
# Read the config file into a Dictionary/Hash structure.
targetWebsiteConfigs = KastParsersLib.kastConfigFileParser(configFile)
if targetWebsiteConfigs == {}:
KastGenericFunctionsLib.logException('Target website configs could not extracted - ' + str(time.time()), errorLog)
sys.exit(-1)
# Obtain the list of URLs from the above data structure and generate time domain
# perfect series representation of html content.
htmlSeries = [KastParsersLib.html2TagSignal(url) for url in targetWebsiteConfigs['SampleURLS']]
# Calculate the average similarity measure.
similarityMeasure = KastParsersLib.calculateThresholdDftDistanceScore(htmlSeries)
# Populate the unseenUrlList
unseenUrlList = KastParsersLib.populateUnseenUrlList(targetWebsite, unseenUrlList)
if unseenUrlList == []:
logException('Seed URL List is malformed. Crawl engine is exiting - ' + str(time.time()), errorLog)
sys.exit(-1)
# Start crawling
crawl(targetWebsite)
# Now apply the Page classification algorithm to preserve only the pages of interest.
classify(htmlSeries, similarityMeasure)
# Apply the CSS rules for scrapping content, this will serve as a simple rule engine template.
contentExtractionRules = targetWebsiteConfigs['ContentExtractionRules']
extractContent(contentExtractionRules)
# Convert the log file into RDF N Triples file
predicateList = targetWebsiteConfigs['PredicateList']
nTriplesFile = table2RDFNTriplesConverter(contentLogFile, predicateList)
# Now log all the information to AllegroGraphDB
store2db(nTriplesFile)
def crawl(targetWebsite):
global sitename
global errorLog
global unseenUrlList
global visitedUrlList
global BASEFILESTORAGEDIR
# Now start the crawling rountine.
while (1):
if unseenUrlList != []:
# Choose a page randomly
page = random.choice(unseenUrlList)
# Fetch the content.
r = KastParsersLib.fetchURL(page)
# Clean the content.
r = KastParsersLib.cleanHtml(r)
# Write the content to a file, in the designated folder.
filename = KastGenericFunctionsLib.extractWebSiteName(page) + '-' + str(round(time.time(), 2))
# Replace all '/' with [kastSlash]
filename = string.replace(filename, '/', '[kastSlash]')
f = gzip.open(BASEFILESTORAGEDIR + filename + '.gz', 'wb')
f.write(r)
f.close()
# Convert to DOM and apply the CSS rule engine
d = pq(r)
ele_a = d('a')
# Extract the hyperlinks
links_a = KastParsersLib.extractHyperlinks(ele_a)
# Convert to absolute links.
unseenUrlListTmp = KastParsersLib.convert2AbsoluteHyperlinks(links_a, targetWebsite)
# Now check how many of these links exist in Visited URL list.
for link in unseenUrlListTmp:
if not visitedUrlList.__contains__(link):
unseenUrlList.append(link)
# Now append this page processed to visited URLs list.
visitedUrlList.append(page)
# Now remove the same link from unseenUrlList.
unseenUrlList.remove(page)
# Condition to end the crawl.
# Debug ON, turn off in production.
pdb.set_trace()
if unseenUrlList == []:
return
