def fogbugzOnFail(self,logfp):
print "Creating FogBuz Ticket"
cfp=Config(self.__fpath)
attempts=0
run=True
while run is True and attempts < 3:
try:
site=FogBugz(cfp.getVar("fogbugz","site","string"))
try:
site.logon(cfp.getVar("fogbugz","user","string"), cfp.getVar("fogbugz","pass","string"))
cfp=Config(self.__fpath)
with open(logfp,'rb') as fp:
print site.new(sTitle="The Parser "+os.path.join(self.__logbase,self.__execute)+" Failed",ixPersonAssignedTo="Andy",Files={"faillog.txt":fp})
attempts+=1
run=False
except Exception,e:
print str(e)
for frame in traceback.extract_tb(sys.exc_info()[2]):
print '\n'.join([str(x) for x in frame])
finally:
site.logoff()
except Exception,e:
print str(e)
for frame in traceback.extract_tb(sys.exc_info()[2]):
print '\n'.join([str(x) for x in frame])
class ProxyManager(object):
def __init__(self,fpath):
self.__fpath = fpath
if self.__fpath is None or os.path.exists(self.__fpath) is False:
raise ConfigurationNotFoundException("Valid Configuration must be Generated")
self._cfp = Config(self.__fpath)
def getProxies(self,number,domain):
return json.loads(urllib2.urlopen(urlparse.urljoin(self._cfp.getVar("proxymanager", "server", "String").strip(),"getProxy?auth=&number={}&domain={}".format(number,domain.strip())), None, timeout = 120).read())
def dropProxy(self,domain,ip):
return json.loads(urllib2.urlopen(urlparse.urljoin(self._cfp.getVar("proxymanager", "server", "String").strip(),"dropProxyForJson?auth=&domain={}&ip={}".format(domain.strip(),ip.strip())), None, timeout = 120).read())
def dropDomain(self,domain):
return json.loads(urllib2.urlopen(urlparse.urljoin(self._cfp.getVar("proxymanager", "server", "String").strip(),"dropDomainForJsony?auth=&domain={}".format(domain.strip())), None, timeout = 120).read())
def run(self):
"""
Executes the crawler as a separate process and monitors for completion. The worker itself is a Thread so run is the
necessary name of the method.
"""
print "Executing "+self.__execute
p=Psyco(self.__fpath)
cfp=Config(self.__fpath)
if self.__execute is not None and self.__logbase is not None:
try:
logfp=self.__logbase+"logs/"+self.__execute.replace(".xml","").replace(".xml","")+str(int(round(time.time())))+".txt"
self.__sema.increment(self.__lock)
try:
p.execute("INSERT INTO crawlers.ca_crim_parsers_running(name) VALUES('"+self.__execute+"')")
p.execute("DELETE FROM crawlers.ca_crim_parsers_not_run WHERE name LIKE '"+self.__execute+"'")
stdfile=open(logfp,'w+')
#the products config file will be in the base directory
cmd="/usr/bin/java -Xms"+cfp.getVar("Pentaho","xms","string").strip()+" -Xmx"+cfp.getVar("Pentaho","xmx","string").strip()+" -XX:+UseConcMarkSweepGC -Xcompactexplicitgc -Dbeansfile="+self.__logbase+self.__execute+" -jar "+self.__jarfile
print cmd
pipe=subprocess.Popen(shlex.split(cmd), stdout=stdfile,stderr=subprocess.STDOUT,shell=False)
ret=pipe.wait()
print "Completed "+self.__execute
p.execute("DELETE FROM crawlers.ca_crim_parsers_running WHERE name LIKE '"+self.__execute+"'")
if ret is 0:
p.execute("INSERT INTO crawlers.ca_crim_parsers_complete (name,crawlstamp) VALUES('"+self.__execute+"','"+str(self.__datestamp)+"')")
else:
print "PARSERS- Premature Detonation!!! Failure "+self.__execute
print str(ret)
if cfp.getVar("notification", "err_type","string") == 'fog':
self.fogbugzOnFail(logfp)
else:
self.emailOnFail(logfp)
finally:
print "Decrementing"
self.__sema.decrement(self.__lock)
except Exception, e:
print "Process Failed"
print str(e)
for frame in traceback.extract_tb(sys.exc_info()[2]):
fname,lineno,fn,text = frame
print "Error in %s on line %d" % (fname, lineno)
print "Worker Complete "+str(time.time())
def emailOnFail(self,logfp):
"""
Email on a failed crawler. Sender should not be internal for outlook purposes.
*Required Parameter*
:param logfp: log file path
"""
print "Sending Alert Message"
cfp=Config(self.__fpath)
#get log as attachment
att=""
with open(logfp,'r') as fp:
att=fp.read()
#prep
body=MIMEMultipart('alternative')
body.attach(MIMEText("PARSER - Premature Detonation!!! Failed or Terminated Parser: "+os.path.join(self.__logbase,self.__execute)))
msg=MIMEText("The Parser "+os.path.join(self.__logbase,self.__execute)+" Failed")
msg=MIMEMultipart()
msg.attach(body)
msg['From']=cfp.getVar("mail","sender","string")
msg['To']=cfp.getVar("mail","recipients","string")
msg['Subject']="Failed CACRIM Parser"
part=MIMEBase('application',"octet-stream")
part.set_payload(att)
Encoders.encode_base64(part)
part.add_header('Content-Disposition','attachment',filename="faillog.txt")
#attach
msg.attach(part)
#send
mailobj=smtplib.SMTP(cfp.getVar("mail","host","string"))
mailobj.sendmail(cfp.getVar("mail", "sender","string"),[cfp.getVar("mail","recipients","string").split(",")],msg.as_string())
mailobj.quit()
print "Mail Sent"
def setup(self,folder,table,confPath,createImages=False):
'''
Setup status table and folder table.
*Required Parameters*
:param folder: the folder name to use
:param table: the table to use in setting up the status
:param confPath: the configuration path
*Optional Parameters*
:param createImages: whether or not to create an images folder
'''
cfp=Config(confPath)
os.mkdir(cfp.getVar("Images","fpath","string")+"FL/SOR/"+folder)
p=Psyco(confPath)
if createImages is True:
#update crawl specific tables
p.execute("INSERT INTO folders.names (folder,schema) VALUES ('"+folder+"','us_fl_crawlsor')")
os.mkdir(cfp.getVar("Images","fpath","string")+"FL/SOR/"+folder)
os.mkdir("Images","fpath","string"+"FL/SOR/"+folder+"/Images/")
def changeOnComplete(self,alterSchema,confPath,folder=None):
'''
Handles the final cleanup on the last part of the crawl.
*Required Parameters*
:param alterSchema: the schema to rename and use
:param confPath: the configuration path to use
*Optional Parameters*
:param folder: folder name to use
NOTE: The folder name should be the saem as provided from getFolder and/or to setup if used.
'''
cfp=Config(confPath)
p=Psyco(confPath)
rename="us_fl_crawlsor"+datetime.datetime.fromtimestamp(time.time()).strftime('%Y%m%d%H%M%S')
print("Renaming Schema to "+rename)
p.execute("ALTER SCHEMA us_fl_crawlsor RENAME TO "+rename)
if folder is not None:
#update the folder name information
p.execute("UPDATE folders.names SET status=true,schema="+rename+" WHERE folder="+folder)
p.execute("INSERT INTO "+cfp.getVar("finished","alterschema","string")+"."+cfp.getVar("finished","altertable","string")+"(filename,schema) VALUES('"+re.search('.*?\/([^\/]+\.py.*)',inspect.getfile(inspect.currentframe())).group(1)+"','"+rename+"')")
class Errors:
"""
Errors is the class that handles all error processing.
"""
# empty constructor
def __init__(self, fpath=None, logger=None):
"""
*Optional Parameters*
:param fpath: The file path to use.
:param logger: The logger to use.
"""
self._errors = {}
self._fpath = fpath
self._cfp = None
self.errCo = -1
if self._fpath is not None:
self._cfp = Config(self._fpath)
self.logger = logger
@property
def errCo(self):
return self.__errco
@errCo.setter
def errCo(self, errCo):
self.__errCo = errCo
def printError(self, msg):
"""
General Error Printing that checks for a log (Python2.7 cannot override print)
:param msg: The message
"""
if self.logger is not None:
self.logger.error(msg)
else:
print (msg)
# print traceback frames
def printFrames(self, frames):
"""
Prints traceback frames.
:param frames: The exception frames.
"""
for frame in frames:
fname, lineno, fn, text = frame
self.printError("Error in %s on line %d" % (fname, lineno))
def postToErrorTables(self, sourcename, e, frames):
"""
Takes in traceback frames and posts them to database. The
expectation is that the data collected here is useful in
anomoly detection.
An error code of -1 specifies that an attribute has no code. This is a property that can be changed
:param sourcename: The name of the source.
:param frames: The traceback frames from the error.
:param e: The error, Exception instance.
"""
code = self.__errCo
description = None
etype = None
etype = str(type(e))
if "HTTPError" in etype:
code = e.code
description = str(e)
elif hasattr(e, "errno") is True:
code = e.errno
if "URLError" in etype:
description = e.reason()
else:
description = str(e)
if sourcename is None:
self.printError("Please Specify a Source Name")
else:
if self._fpath is not None and self._cfp is not None and type is not None and description is not None:
p = Psyco(self._fpath)
p.execute(
"INSERT INTO "
+ self._cfp.getVar("Error", "table", "string")
+ " (source,type,code,description) VALUES('"
+ sourcename
+ "','"
+ etype
+ "','"
+ str(code)
+ "','"
+ description
+ "')"
)
elif type is None or description is None:
self.printError("Type or Description Not Found for the Error. Cannot insert To Database.")
else:
self.printError("Please Specify a Config Path")
# handle http error
def httpError(self, e, link, infailcodes, inkillswitch, incurrpos, frames):
"""
Handles httpErrors
*Required Parameters*
:param e: error instance
:param link: url address when the error occurred
:param infailcodes: fail codes for the url
:param inkillswitch: number of failures to the current point
:param incurrpos: current position in the link list on fail
:param frames: the traceback frames at failure
"""
currpos = incurrpos
killswitch = inkillswitch
failcodes = infailcodes
print "Failed to Get Page " + link
if "URLError" in str(type(e)):
self.printError("URLError")
print e.reason
failcodes += str(e.reason) + ","
elif "HTTPError" in str(type(e)):
self.printError("HTTPError")
print e.code
failcodes += str(e.code) + ","
else:
self.printError("Exception Error")
print str(e)
failcodes += str(e) + ","
if killswitch < 5:
killswitch += 1
else:
self.printError("Critical Failure for " + link)
ks = 0
currpos += 1
self.printFrames(frames)
return (failcodes, killswitch, currpos)
# crawlersetup failure
def crawlerSetupError(self, e, frames):
"""
Handles failures in crawler setup while taking in the error
instance and traceback frames.
"""
self.printError("Crawler Setup Failed")
self.error(e, frames)
# crawler process failure other than in the pull
def crawlerFail(self, e, frames, restart):
"""
Handles a failure in the crawler.
:param e: The exception.
:param frames: The frames to use.
:param restart: Whether or not the process will restart.
"""
print "Failed to Get Pages."
if restart is True:
self.printError("Will Restart This Iteration")
self.error(e, frames)
# Prints an error without a header
def error(self, e, frames):
"""
General error method for handling error instances and traceback
frames.
:param e: The exception.
:param frames: The exception frames.
"""
self.printError(str(e))
self.printFrames(frames)
def parse(self):
"""
Run the parser.
"""
cfp=Config(self.__fpath)
runlist=[]
files=self.getXmlFiles(cfp.getVar("base", "basedir","string"))
print "Files Found: ",
print sorted(files)
#get completed crawlers and parsers
p=Psyco(self.__fpath)
data=p.getData("SELECT q1.name,q1.stamp FROM (SELECT * FROM (SELECT name,max(date) as stamp FROM crawlers.ca_crim_weekly_complete GROUP BY name) as q1 UNION (SELECT name,max(date) FROM crawlers.ca_crim_monthly_complete GROUP BY name UNION SELECT name,max(date) FROM crawlers.ca_crim_daily_complete GROUP BY name)) as q1 LEFT OUTER JOIN (SELECT name,crawlstamp FROM crawlers.ca_crim_parsers_complete) as q2 ON q1.stamp = q2.crawlstamp AND regexp_replace(q1.name,'.py','') = regexp_replace(q2.name,'.xml','') WHERE q2.name IS NULL")
nilList=[]
for xfile in data:
if re.sub('\.py.*','',xfile[0])+".xml" not in files:
nilList.append(re.sub('\.py.*','',xfile[0])+".xml")
if len(nilList) >0:
print "Files Do Not Exist for the Following (the name but not the extension must match)",
print sorted(nilList)
else:
print "All Crawlers Have Corresponding Parsers"
#get datestamp dict -- assumes that the query gets the max(date)
dates={}
for xfile in data:
fp=re.sub('\.py.*','',xfile[0])+".xml"
if fp not in dates:
dates[fp]=xfile[1]
for xfile in data:
if xfile is not None:
fp=xfile[0]
if fp.replace('.py','.xml') in files:
runlist.append(fp.replace('.py','.xml'))
print "Execution List: ",
print sorted(runlist)
nilList=[]
for xfile in data:
if re.sub('\.py.*','',xfile[0])+".xml" not in files:
nilList.append(re.sub('\.py.*','',xfile[0])+".xml")
if len(nilList) >0:
print "Parsers that may not have been Found",
print sorted(nilList)
else:
print "All Completed and Found crawlers Accounted For"
#run the crawlers from the run list
if len(runlist) > 0:
pnum=0
pool=[]
#get the semaphore and lock
sema=IncSemaphore()
lock=Lock()
#max processes
maxprocs=cfp.getVar("opts","maxdel","int")
print "Max Processes: "+str(maxprocs)
#run
cfp=Config(self.__fpath)
while len(runlist) >0:
if pnum<maxprocs and len(runlist)>0 and runlist[0] in files:
ex=runlist.pop()
print "Starting "+ex
w=Worker(self.__fpath,ex.strip(),cfp.getVar("base","logbase","string"),cfp.getVar("db", "dbname","string"),cfp.getVar("db", "user","string"),cfp.getVar("db", "passw","string"),cfp.getVar("db", "host","string"), cfp.getVar("db", "port","string"),sema,lock,dates[ex],cfp.getVar("jars","parser","string"))
w.start()
pool.append(w)
pnum+=1
print "Continuing"
while sema.getValue() >= maxprocs and len(pool) > 0:
print "Waiting for Parsers to Complete"
time.sleep(random.randint(1,120))
for proc in pool:
if not proc.isAlive():
pool.remove(proc)
del proc
pnum-=1
if len(pool) is 0 and sema.getValue() >0:
sema.setValue(0, lock)
gc.collect()
del gc.garbage[:]
for proc in pool:
if not proc.isAlive():
pool.remove(proc)
del proc
pnum-=1
if len(pool) is 0 and sema.getValue() >0:
sema.setValue(0, lock)
gc.collect()
del gc.garbage[:]
if sema.getValue()==0 and len(runlist)==0:
break
time.sleep(30)
print "Completed Loop. Awaiting any final Completions."
gc.collect()
del gc.garbage[:]
while sema.getValue()>0 and len(pool)>0:
print "Waiting for Completion"
time.sleep(random.randint(1,120))
if len(pool) is 0 and sema.getValue() >0:
sema.setValue(0, lock)
print "Current Execution List Complete. Will Restart Soon!"
def start(self):
"""
Start the controller. The controller is not multi-processed.
Paths are configured, executables obtained, workers run; logs compressed from
here.
"""
#get configuration file
cfp=Config(self.__fpath)
if not os.path.isdir(cfp.getVar("base","basedir","string")+"logs"):
os.mkdir(cfp.getVar("base","basedir","string")+"logs")
#setup variables
maxprocs=cfp.getVar("opts", "maxdel","int")
#the threadpool
pool=[]
#set up db
p=Psyco(cfp.getVar("db", "dbname","string"),cfp.getVar("db", "user","string"),cfp.getVar("db", "passw","string"),cfp.getVar("db", "host","string"), cfp.getVar("db", "port","string"))
p.prep()
p.execute("CREATE TABLE IF NOT EXISTS data.failedurls(source text,url text)")
prioritylist=[]
if cfp.getVar("priorities","usepriority","int") is 1:
prioritylist=cfp.getVar("priorities","priority_list","list")
else:
prioritylist=[x[0] for x in p.getData("SELECT name FROM crawlers.sor_complete WHERE name NOT ILIKE '%ParserController%' ORDER BY date ASC")]
#get the python files
#get the semaphore and lock
sema=IncSemaphore()
lock=Lock()
#the loop
run=True
print "Starting Crawlers"
print "MAX: "+str(maxprocs)
procnum=0
print "Base Directory is "+os.path.split(inspect.stack()[0][1])[0]
while run is True:
execs=self.cleanUp(self.getPyFiles(os.path.split(inspect.stack()[0][1])[0]),cfp.getVar("restart","waitdays","string"),p)
execs=list(set(execs))
execs=[x for x in execs if x not in prioritylist]
prioritylist.extend(execs)
execs=prioritylist
print "Executables: "+str(execs)
startnew=maxprocs-sema.getValue()
while len(pool) >0 or procnum is 0 and len(execs)>0:
print "Processes Left "+str(startnew)
print "Number of Running Processes "+str(sema.getValue())
for proc in pool:
if not proc.isAlive():
pool.remove(proc)
del proc
procnum-=1
execs=self.repopulate(execs,cfp.getVar("restart","waitdays","int"),p,maxprocs-sema.getValue(),cfp.getVar("base","basedir","string"))
execs=list(set(self.cleanUp(self.getPyFiles(cfp.getVar("base","basedir","string")),cfp.getVar("restart","waitdays","string"),p)))
if len(pool) is 0 and sema.getValue() >0:
sema.setValue(0, lock)
print "Pnum: "+str(procnum)
if procnum < len(execs) and sema.getValue() < maxprocs:
if float(psutil.swap_memory()[3])>2.0 and float(psutil.cpu_times_percent(interval=1,percpu=False)[4])>10.0:
print "Resources Low! Waiting for resources to free."
print "SMEM: "+str(psutil.swap_memory()[3])
print "IOWait"+str(float(psutil.cpu_times_percent(interval=1,percpu=False)[4]))
run=True
while run is True:
avgsmem=0
avgiowait=0
avgmem=0
for i in range (0,32):
time.sleep(.5)
avgsmem+=psutil.swap_memory()[3]
avgmem+=psutil.virtual_memory()[2]
avgiowait+=float(psutil.cpu_times_percent(interval=1,percpu=False)[4])
avgsmem/=32
avgmem/=32
avgiowait/=32
print "Avg. Swap Space: "+str(avgsmem)
print "Avg. Mem: "+str(avgmem)
print "Avg. IO Wait: "+str(avgiowait)
if avgiowait<10 and avgsmem<2:
print "Resource Level Acceptable. Continuing!"
run=False
else:
print "Resources Low! Waiting for resources to free."
try:
ex=execs[procnum]
print "Starting "+ex
w=Worker(self.__fpath,ex.strip(),cfp.getVar("base","logbase","string"),cfp.getVar("db", "dbname","string"),cfp.getVar("db", "user","string"),cfp.getVar("db", "passw","string"),cfp.getVar("db", "host","string"), cfp.getVar("db", "port","string"),sema,lock)
w.start()
pool.append(w)
time.sleep(5)
procnum+=1
execs.remove(ex)
except Exception, e:
print "Failed to Start a Crawler"
print "Crawler Was: "+ex
print str(e)
for frame in traceback.extract_tb(sys.exc_info()[2]):
fname,lineno,fn,text = frame
print "Error in %s on line %d" % (fname, lineno)
elif sema.getValue()>0:
print "Waiting for Crawlers to Complete "+str(sema.getValue())
print "Waiting To Start ",
print execs
time.sleep(30)
startnew=maxprocs-sema.getValue()