def __init__(self, source=None, logger=None):
# General Attributes
self.source = source
self.reader = None
self.drp = None
self.count = 0
self.Mcount = 99999
if source is not None:
self.ingestDir = PXPaths.RXQ + self.source.name
if self.source.type == 'filter' or self.source.type == 'filter-bulletin' :
self.ingestDir = PXPaths.FXQ + self.source.name
self.logger = source.logger
elif logger is not None:
self.logger = logger
self.pxManager = PXManager() # Create a manager
self.pxManager.setLogger(self.logger) # Give the logger to the the manager
self.pxManager.initNames() # Set rx and tx names
self.clientNames = self.pxManager.getTxNames() # Obtains the list of client's names (the ones to wich we can link files)
self.filterNames = self.pxManager.getFxNames() # Obtains the list of filter's names (the ones to wich we can link files)
if source is not None:
if self.source.name in self.filterNames : self.filterNames.remove(self.source.name)
self.sourlientNames = self.pxManager.getTRxNames() # Obtains the list of sourlient's names (the ones to wich we can link files)
self.allNames = self.clientNames + self.filterNames + self.sourlientNames # Clients + Sourlients names
self.clients = {} # All the Client/Filter/Sourlient objects
self.fileCache = None # product processed.
self.dbDirsCache = CacheManager(maxEntries=200000, timeout=25*3600) # Directories created in the DB
self.clientDirsCache = CacheManager(maxEntries=25000, timeout=2*3600) # File ingested in RXQ
self.feedNames = [] # source to feed
self.feeds = {}
if source is not None:
self.logger.info("Ingestor (source %s) can link files to clients: %s" % (source.name, self.allNames))
def setUp(self, logFile='log/PXManager.log'):
self.logger = Logger(logFile, 'DEBUG', 'Sub')
self.logger = self.logger.getLogger()
self.manager = PXManager()
self.manager.setLogger(self.logger)
self.manager.initNames()
self.manager.initPXPaths()
self.manager.initShouldRunNames()
self.manager.initRunningNames()
def __init__(self):
self.getOptionsParser()
#print SwitchoverCopier.DRBD
if SwitchoverCopier.SYSTEM == 'PX':
from PXManager import PXManager
manager = PXManager(None, SwitchoverCopier.DRBD + '/px/')
LOG_NAME = manager.LOG + 'PX_SwitchoverCopier.log' # Log's name
SwitchoverCopier.SWITCH_DIR = '/apps/px/switchover/'
elif SwitchoverCopier.SYSTEM == 'PDS':
from PDSManager import PDSManager
manager = PDSManager(None, SwitchoverCopier.DRBD + '/pds/')
LOG_NAME = manager.LOG + 'PDS_SwitchoverCopier.log' # Log's name
SwitchoverCopier.SWITCH_DIR = '/apps/pds/switchover/'
self.logger = Logger(LOG_NAME, SwitchoverCopier.LOG_LEVEL, "Copier")
self.logger = self.logger.getLogger()
manager.setLogger(self.logger)
manager.afterInit()
self.logger.info("Beginning program SwitchoverCopier")
self.rxPaths = manager.getRxPaths()
self.txPaths = manager.getTxPaths()
self.logger.info("Receivers paths: " + str(self.rxPaths))
self.logger.info("Senders paths: " + str(self.txPaths))
self.manager = manager
def __init__(self):
self.getOptionsParser()
#print SwitchoverCopier.DRBD
if SwitchoverCopier.SYSTEM == 'PX':
from PXManager import PXManager
manager = PXManager(None, SwitchoverCopier.DRBD + '/px/')
LOG_NAME = manager.LOG + 'PX_SwitchoverCopier.log' # Log's name
SwitchoverCopier.SWITCH_DIR = '/apps/px/switchover/'
elif SwitchoverCopier.SYSTEM == 'PDS':
from PDSManager import PDSManager
manager = PDSManager(None, SwitchoverCopier.DRBD + '/pds/')
LOG_NAME = manager.LOG + 'PDS_SwitchoverCopier.log' # Log's name
SwitchoverCopier.SWITCH_DIR = '/apps/pds/switchover/'
self.logger = Logger(LOG_NAME, SwitchoverCopier.LOG_LEVEL, "Copier")
self.logger = self.logger.getLogger()
manager.setLogger(self.logger)
manager.afterInit()
self.logger.info("Beginning program SwitchoverCopier")
self.rxPaths = manager.getRxPaths()
self.txPaths = manager.getTxPaths()
self.logger.info("Receivers paths: " + str(self.rxPaths))
self.logger.info("Senders paths: " + str(self.txPaths))
self.manager = manager
def __init__(self, request, addOn, sendOn, logger):
self.logger = logger
self.addOn = addOn
self.sendOn = sendOn
self.drp = DirectRoutingParser(PXPaths.ROUTING_TABLE, [], logger)
self.drp.printErrors = False
self.drp.parseAndShowErrors()
self.dbs = DBSearcher(request, False)
self.results = self.dbs.results
self.receiverName = 'request-reply'
self.pxManager = PXManager()
self.pxManager.setLogger(self.logger)
self.pxManager.initNames()
self.bulletin = ''
self.constructBulletin()
def __init__(self, source=None, logger=None):
# General Attributes
self.source = source
self.reader = None
self.drp = None
self.count = 0
self.Mcount = 99999
if source is not None:
self.ingestDir = PXPaths.RXQ + self.source.name
if self.source.type == 'filter' or self.source.type == 'filter-bulletin' :
self.ingestDir = PXPaths.FXQ + self.source.name
self.logger = source.logger
elif logger is not None:
self.logger = logger
self.pxManager = PXManager() # Create a manager
self.pxManager.setLogger(self.logger) # Give the logger to the the manager
self.pxManager.initNames() # Set rx and tx names
self.clientNames = self.pxManager.getTxNames() # Obtains the list of client's names (the ones to wich we can link files)
self.filterNames = self.pxManager.getFxNames() # Obtains the list of filter's names (the ones to wich we can link files)
if source is not None:
if self.source.name in self.filterNames : self.filterNames.remove(self.source.name)
self.sourlientNames = self.pxManager.getTRxNames() # Obtains the list of sourlient's names (the ones to wich we can link files)
self.allNames = self.clientNames + self.filterNames + self.sourlientNames # Clients + Sourlients names
self.clients = {} # All the Client/Filter/Sourlient objects
self.fileCache = None # product processed.
self.dbDirsCache = CacheManager(maxEntries=200000, timeout=25*3600) # Directories created in the DB
self.clientDirsCache = CacheManager(maxEntries=25000, timeout=2*3600) # File ingested in RXQ
self.feedNames = [] # source to feed
self.feeds = {}
if source is not None:
self.logger.info("Ingestor (source %s) can link files to clients: %s" % (source.name, self.allNames))
def setUp(self,logFile='log/PXManager.log'): self.logger = Logger(logFile, 'DEBUG', 'Sub') self.logger = self.logger.getLogger() self.manager = PXManager() self.manager.setLogger(self.logger) self.manager.initNames() self.manager.initPXPaths() self.manager.initShouldRunNames() self.manager.initRunningNames()
def __init__(self, nopull=False, keep=False, date=None, xstats=False):
PXPaths.normalPaths()
self.manager = PXManager()
#self.logger = logger.getLogger()
# Date for which we want to obtain stats
if date == None:
self.date = dateLib.getYesterdayFormatted() # ISO Date
else:
self.date = date
self.dateDashed = dateLib.getISODateDashed(self.date)
self.machines = [] # Machines were the logs can be found
self.sources = [
] # Sources for which we will check arrival time of the products
self.client = [
] # Client for which we will check delivery time of the products (ONLY ONE ENTRY in the list)
self.messages = [] # FIXME: Special messages coming from weird results
self.nopull = nopull # Do not pull the necessary files (we suppose they are already downloaded)
self.keep = keep # Erase all the files present before downloading new files
self.xstats = xstats # Boolean that determine if we will use xferlog in making stats
self.goodRx = [] # Lines matching initial values
self.goodTx = [] # Lines matching initial values
self.goodXferlog = [] # Lines matching initial values
self.receivingInfos = {
} # Dict. addressed by filename and containing a tuple of (formatted date, date in seconds, machine)
self.sendingInfos = {
} # Dict. addressed by filename and containing a tuple of (formatted date, date in seconds, machine)
self.xferlogInfos = {
} # Dict. addressed by filename and containing a tuple of (formatted date, date in seconds, machine)
self.stats = {} # Final stats
self.sortedStats = [] # Final sorted stats
self.max = 0 # Maximum latency time in seconds
self.min = sys.maxint # Minimum latency time in seconds
self.mean = 0 # Mean latency time in seconds
self.latencyThreshold = 15 # We don't want to go over this threshold (in seconds)
self.overThreshold = 0 # Number of files with latency over threshold
self.underThresholdP = 0 # Percentage of files for which the latency is equal or under threshold
self.meanWaiting = 0 # Mean waiting time before being noticed by the PDS
self.random = str(
random.random()
)[2:] # Unique identificator permitting the program to be run in parallel
self.system = None # 'PDS' or 'PX'
self.rejected = 0 # Count of rejected files
self.maxInfos = ['NO FILE', ('00:00:00', 'No machine', 0)
] # Informations about the max.
def __init__(self):
self.getOptionsParser()
if SwitchoverDeleter.SYSTEM == 'PX':
from PXManager import PXManager
manager = PXManager(None, '/apps/px/')
LOG_NAME = manager.LOG + 'PX_SwitchoverDeleter.log' # Log's name
SwitchoverDeleter.SWITCH_DIR = '/apps/px/switchover/'
elif SwitchoverDeleter.SYSTEM == 'PDS':
from PDSManager import PDSManager
manager = PDSManager(None, '/apps/pds/')
LOG_NAME = manager.LOG + 'PDS_SwitchoverDeleter.log' # Log's name
SwitchoverDeleter.SWITCH_DIR = '/apps/pds/switchover/'
self.logger = Logger(LOG_NAME, SwitchoverDeleter.LOG_LEVEL, "Deleter")
self.logger = self.logger.getLogger()
manager.setLogger(self.logger)
self.logger.info("Beginning program SwitchoverDeleter")
self.manager = manager
def __init__(self, request, addOn, sendOn, logger):
self.logger = logger
self.addOn = addOn
self.sendOn = sendOn
self.drp = DirectRoutingParser(PXPaths.ROUTING_TABLE, [], logger)
self.drp.printErrors = False
self.drp.parseAndShowErrors()
self.dbs = DBSearcher(request, False)
self.results = self.dbs.results
self.receiverName = 'request-reply'
self.pxManager = PXManager()
self.pxManager.setLogger(self.logger)
self.pxManager.initNames()
self.bulletin = ''
self.constructBulletin()
def __init__(self, type='impulses', interval=1, imageName=None):
PXPaths.normalPaths()
self.manager = PXManager()
#self.logger = logger
#self.manager.setLogger(self.logger)
self.latenciers = [] # Infos about a particular "latencier"
self.type = type # Type of graph must be in: ['linespoint', 'lines', 'boxes', 'impulses']
self.interval = interval * dateLib.MINUTE # Number of seconds between each point on the x-axis
self.imageName = imageName # Name of the image file
self.color = None
self.width = dateLib.DAY # Width of the x-axis in seconds
# With witdh=DAY and interval=MINUTE => len([60, 120, 180, ..., 86400]) = 1440
self.separators = dateLib.getSeparators(self.width, self.interval)
# '"0" 0, "1" 60, "2" 120, "3" 180, "4" 240, ... , "22" 1320, "23" 1380, "24" 1440'
self.xtics = self.getXTics(len(self.separators), self.interval)
self.graph = Gnuplot.Gnuplot()
class unittest_PXManager(unittest.TestCase):
def setUp(self,logFile='log/PXManager.log'):
self.logger = Logger(logFile, 'DEBUG', 'Sub')
self.logger = self.logger.getLogger()
self.manager = PXManager()
self.manager.setLogger(self.logger)
self.manager.initNames()
self.manager.initPXPaths()
self.manager.initShouldRunNames()
self.manager.initRunningNames()
def test_PXManager(self):
self.assertEqual(self.manager.getTxNames(), ['client-am', 'client-ftp', 'client-test'])
self.assertEqual(self.manager.getFxNames(),['filter-am'])
self.assertEqual(self.manager.getRxNames(),['source-am','source-ftp','source-test'])
self.assertEqual(self.manager.getFlowType('client-am'),('TX', ['client-am']))
self.assertEqual(self.manager.getDBName('MAPS_12Z:OCXCARTE:CMC:CHART:IMV6::20120611192817'),'./db/20120611/CHART/OCXCARTE/CMC/MAPS_12Z:OCXCARTE:CMC:CHART:IMV6::20120611192817')
self.assertEqual(self.manager.getShouldRunFxNames(),['filter-am'])
self.assertEqual(self.manager.getShouldRunRxNames(),['source-test'])
self.assertEqual(self.manager.getShouldRunTxNames(),['client-test'])
self.assertEqual(self.manager.getRunningFxNames(),['filter-am'])
self.assertEqual(self.manager.getRunningRxNames(),[])
self.assertEqual(self.manager.getRunningTxNames(),[])
self.assertEqual(self.manager.getNotRunningFxNames(),[])
self.assertEqual(self.manager.getNotRunningRxNames(),['source-test'])
self.assertEqual(self.manager.getNotRunningTxNames(),['client-test'])
class Ingestor(object):
"""
Normally, an Ingestor will be in a Source. It can also be used for the only reason that this object has
access to all the configuration options of the clients. For this particular case, source=None.
"""
def __init__(self, source=None, logger=None):
# General Attributes
self.source = source
self.reader = None
self.drp = None
self.count = 0
self.Mcount = 99999
if source is not None:
self.ingestDir = PXPaths.RXQ + self.source.name
if self.source.type == 'filter' or self.source.type == 'filter-bulletin':
self.ingestDir = PXPaths.FXQ + self.source.name
self.logger = source.logger
elif logger is not None:
self.logger = logger
self.pxManager = PXManager() # Create a manager
self.pxManager.setLogger(
self.logger) # Give the logger to the the manager
self.pxManager.initNames() # Set rx and tx names
self.clientNames = self.pxManager.getTxNames(
) # Obtains the list of client's names (the ones to wich we can link files)
self.filterNames = self.pxManager.getFxNames(
) # Obtains the list of filter's names (the ones to wich we can link files)
if source is not None:
if self.source.name in self.filterNames:
self.filterNames.remove(self.source.name)
self.sourlientNames = self.pxManager.getTRxNames(
) # Obtains the list of sourlient's names (the ones to wich we can link files)
self.allNames = self.clientNames + self.filterNames + self.sourlientNames # Clients + Sourlients names
self.clients = {} # All the Client/Filter/Sourlient objects
self.fileCache = None # product processed.
self.dbDirsCache = CacheManager(maxEntries=200000, timeout=25 *
3600) # Directories created in the DB
self.clientDirsCache = CacheManager(maxEntries=25000, timeout=2 *
3600) # File ingested in RXQ
self.feedNames = [] # source to feed
self.feeds = {}
if source is not None:
self.logger.info(
"Ingestor (source %s) can link files to clients: %s" %
(source.name, self.allNames))
def setFeeds(self, feedNames):
from Source import Source
sources = self.pxManager.getRxNames()
for name in feedNames:
if not name in sources: continue
instant = Source(name, self.logger, False)
if instant.type == 'am' or instant.type == 'amqp' or instant.type == 'wmo':
self.logger.warning(
"Feed (source %s) will be ignored (type %s)" %
(name, instant.type))
continue
self.feedNames.append(name)
self.feeds[name] = instant
self.logger.info(
"Ingestor (source %s) can link files to receiver: %s" %
(self.source.name, self.feedNames))
def createDir(self, dir, cacheManager):
if cacheManager.find(dir) == None:
try:
os.makedirs(dir, 01775)
except OSError:
(type, value, tb) = sys.exc_info()
self.logger.debug(
"Problem when creating dir (%s) => Type: %s, Value: %s" %
(dir, type, value))
def setClients(self):
""""
Set a dictionnary of Clients. Main usage will be to access value of
configuration options (mainly masks) of the Client objects.
"""
from Source import Source
from Sourlient import Sourlient
for name in self.clientNames:
self.clients[name] = Client(name, self.logger)
for name in self.filterNames:
self.clients[name] = Source(name, self.logger, False, True)
for name in self.sourlientNames:
self.clients[name] = Sourlient(name, self.logger, False)
#self.clients[name].readConfig()
#print self.clients[name].masks
def getIngestName(self, receptionName):
"""
Map reception name to ingest name, based on the source configuration.
This just inserts missing fields, like whattopds. DUMB!
FIXME: Have a library of functions, configurable per source, to
perform the mapping, perhaps using rmasks ? & other args.
"""
receptionNameParts = receptionName.split(':')
extensionParts = self.source.extension.split(':')
if len(receptionNameParts) > 6:
receptionNameParts = receptionNameParts[:6]
self.logger.warning("File %s truncated to %s" %
(receptionName, ':'.join(receptionNameParts)))
for i in range(1, 6):
if len(receptionNameParts) == i:
receptionNameParts = receptionNameParts + [extensionParts[i]]
elif receptionNameParts[i] == '':
receptionNameParts[i] = extensionParts[i]
receptionNameParts = receptionNameParts + [
time.strftime("%Y%m%d%H%M%S", time.gmtime())
]
return string.join(receptionNameParts, ':')
def getClientQueueName(self, clientName, ingestName, priority=None):
"""
Return the directory into which a file of a given priority should be placed.
Layout used is: /apps/px/txq/<client>/<priority>/YYYYmmddhh
"""
parts = ingestName.split(':')
if not priority:
priority = parts[4].split('.')[0]
clientpathName = PXPaths.TXQ + clientName + '/' + str(
priority) + '/' + time.strftime("%Y%m%d%H",
time.gmtime()) + '/' + ingestName
if clientName in self.filterNames:
clientpathName = PXPaths.FXQ + clientName + '/' + ingestName
return clientpathName
def getDBName(self, ingestName):
"""
Given an ingest name, return a relative database name
Given a file name of the form:
what : ori_system : ori_site : data_type : format :
link it to:
db/<today>/data_type/ori_system/ori_site/ingestName
(same pattern as PDS)
NB: see notes/tests.txt for why the date/time is recalculated everytime.
"""
if ingestName.count(':') >= 4:
today = time.strftime("%Y%m%d", time.gmtime())
dirs = ingestName.split(':')
return PXPaths.DB + today + '/' + dirs[3] + '/' + dirs[
1] + '/' + dirs[2] + '/' + ingestName
else:
return ''
def getRouteKey(self, filename):
"""
Given an ingest name, return a route key based on the imask given
"""
# check against the masks
for mask in self.source.masks:
# no match
if not mask[3].match(filename): continue
# reject
if not mask[4]: return None
# accept... so key generation
parts = re.findall(mask[0], filename)
if len(parts) == 2 and parts[1] == '': parts.pop(1)
if len(parts) != 1: continue
key = parts[0]
if isinstance(parts[0], tuple): key = '_'.join(parts[0])
self.logger.debug("RouteKey Key = %s Mask = %s Filename = %s" %
(key, mask[0], filename))
return key
# fallback behavior return filename
return filename
def isMatching(self, client, ingestName):
"""
Verify if ingestName is matching one mask of a client
"""
from Source import Source
if len(client.masks_deprecated) > 0:
for mask in client.masks_deprecated:
if fnmatch.fnmatch(ingestName, mask[0]):
try:
if mask[2]:
return True
except:
return False
for mask in client.masks:
if mask[3].match(ingestName): return mask[4]
if isinstance(client, Source): return True
return False
def getMatchingClientNamesFromMasks(self, ingestName,
potentialClientNames):
matchingClientNames = []
for name in potentialClientNames:
try:
if self.isMatching(self.clients[name], ingestName):
matchingClientNames.append(name)
except KeyError:
pass
return matchingClientNames
def getMatchingFeedNamesFromMasks(self, ingestName, potentialFeedNames):
matchingFeedNames = []
for name in potentialFeedNames:
try:
if self.feeds[name].fileMatchMask(ingestName):
matchingFeedNames.append(name)
except KeyError:
pass
return matchingFeedNames
def ingest(self, receptionName, ingestName, clientNames, priority=None):
self.logger.debug("Reception Name: %s" % receptionName)
dbName = self.getDBName(ingestName)
if dbName == '':
self.logger.warning('Bad ingest name (%s) => No dbName' %
ingestName)
return 0
self.createDir(os.path.dirname(dbName), self.dbDirsCache)
#try:
# os.link(receptionName, dbName)
#except OSError:
# (type, value, tb) = sys.exc_info()
# self.logger.error("Unable to link %s %s, Type: %s, Value: %s" % (receptionName, dbName, type, value))
os.link(receptionName, dbName)
nbBytes = os.stat(receptionName)[stat.ST_SIZE]
if self.source.debug:
self.logger.info("DBDirsCache: %s" % self.dbDirsCache.cache)
stats, cached, total = self.dbDirsCache.getStats()
if total:
percentage = "%2.2f %% of the last %i requests were cached" % (
cached / total * 100, total)
else:
percentage = "No entries in the cache"
self.logger.info("DB Caching stats: %s => %s" %
(str(stats), percentage))
self.logger.debug("ClientDirsCache: %s" %
self.clientDirsCache.cache)
stats, cached, total = self.clientDirsCache.getStats()
if total:
percentage = "%2.2f %% of the last %i requests were cached" % (
cached / total * 100, total)
else:
percentage = "No entries in the cache"
self.logger.debug("Client Caching stats: %s => %s" %
(str(stats), percentage))
self.logger.info("Ingestion Name: %s" % ingestName)
self.logger.info("(%i Bytes) Ingested in DB as %s" % (nbBytes, dbName))
# Problem bulletins are databased, but not sent to clients
if ingestName.find("PROBLEM_BULLETIN") is not -1:
return 1
for name in clientNames:
clientQueueName = self.getClientQueueName(name, ingestName,
priority)
self.createDir(os.path.dirname(clientQueueName),
self.clientDirsCache)
#try:
# os.link(dbName, clientQueueName)
#except OSError:
# (type, value, tb) = sys.exc_info()
# self.logger.error("Unable to link %s %s, Type: %s, Value: %s" % (dbName, clientQueueName, type, value))
os.link(dbName, clientQueueName)
feedNames = []
if len(self.feedNames) > 0:
feedNames = self.getMatchingFeedNamesFromMasks(
ingestName, self.feedNames)
self.logger.debug("Matching (from patterns) feed names: %s" %
feedNames)
for name in feedNames:
if name in clientNames: continue
sourceQueueName = PXPaths.RXQ + name + '/' + ingestName
self.createDir(os.path.dirname(sourceQueueName),
self.clientDirsCache)
os.link(dbName, sourceQueueName)
self.logger.info("Queued for: %s" % string.join(clientNames) + ' ' +
string.join(feedNames))
return 1
def run(self):
if self.source.type == 'single-file' or self.source.type == 'pull-file':
self.ingestSingleFile()
elif self.source.type == 'bulletin-file' or self.source.type == 'pull-bulletin':
self.ingestBulletinFile()
elif self.source.type == 'collector':
self.ingestCollection()
def ingestSingleFile(self, igniter):
from DiskReader import DiskReader
from DirectRoutingParser import DirectRoutingParser
from PullFTP import PullFTP
if self.source.routemask:
self.drp = DirectRoutingParser(self.source.routingTable,
self.allNames, self.logger,
self.source.routing_version)
self.drp.parse()
if self.source.nodups:
self.fileCache = CacheManager(maxEntries=self.source.cache_size,
timeout=8 * 3600)
reader = DiskReader(self.ingestDir, self.source.batch,
self.source.validation,
self.source.patternMatching, self.source.mtime,
False, self.source.logger, self.source.sorter,
self.source)
sleep_sec = 1
if self.source.type == 'pull-file' or self.source.pull_script != None:
sleep_sec = self.source.pull_sleep
while True:
if igniter.reloadMode == True:
# We assign the defaults, reread configuration file for the source
# and reread all configuration file for the clients (all this in __init__)
if self.source.type == 'filter':
self.source.__init__(self.source.name, self.source.logger,
True, True)
else:
self.source.__init__(self.source.name, self.source.logger)
if self.source.routemask:
self.drp = DirectRoutingParser(self.source.routingTable,
self.allNames, self.logger)
self.drp.parse()
if self.source.nodups:
self.fileCache = CacheManager(
maxEntries=self.source.cache_size, timeout=8 * 3600)
reader = DiskReader(self.ingestDir, self.source.batch,
self.source.validation,
self.source.patternMatching,
self.source.mtime, False,
self.source.logger, self.source.sorter,
self.source)
self.logger.info("Receiver has been reloaded")
igniter.reloadMode = False
# pull files in rxq directory if in pull mode
if self.source.type == 'pull-file' or self.source.pull_script != None:
files = []
sleeping = os.path.isfile(PXPaths.RXQ + self.source.name +
'/.sleep')
if self.source.type == 'pull-file':
puller = PullFTP(self.source, self.logger, sleeping)
files = puller.get()
puller.close()
elif self.source.pull_script != None:
files = self.source.pull_script(self.source, self.logger,
sleeping)
if not sleeping:
self.logger.debug("Number of files pulled = %s" %
len(files))
else:
self.logger.info("This pull is sleeping")
# normal diskreader call for files
reader.read()
if len(reader.sortedFiles) <= 0:
time.sleep(sleep_sec)
continue
sortedFiles = reader.sortedFiles[:self.source.batch]
# processing the list if necessary...
if self.source.lx_execfile != None:
sfiles = []
sfiles.extend(sortedFiles)
self.logger.info("%d files process with lx_script" %
len(sfiles))
sortedFiles = self.source.run_lx_script(
sfiles, self.source.logger)
self.logger.info("%d files will be ingested" % len(sortedFiles))
for file in sortedFiles:
self.ingestFile(file)
def ingestFile(self, file):
# check for duplicated if user requieres
if self.source.nodups:
# get md5 key from file...
md5_key = self.fileCache.get_md5_from_file(file)
# If data is already in cache, we don't send it
if self.fileCache.find(md5_key, 'standard') is not None:
os.unlink(file)
self.logger.info("suppressed duplicate file %s",
os.path.basename(file))
return
# converting the file if necessary
if self.source.fx_execfile != None:
fxfile = self.source.run_fx_script(file, self.source.logger)
# convertion did not work
if fxfile == None:
self.logger.warning("FX script ignored the file : %s" %
os.path.basename(file))
os.unlink(file)
return
# file already in proper format
elif fxfile == file:
self.logger.warning("FX script kept the file as is : %s" %
os.path.basename(file))
# file converted...
else:
self.logger.info(
"FX script modified %s to %s " %
(os.path.basename(file), os.path.basename(fxfile)))
os.unlink(file)
file = fxfile
# filename to ingest
ingestName = self.getIngestName(os.path.basename(file))
# make sure we do not have a dbName already in there
# if it happens add a second to the postfix datetime stamp
# until the dbName is not found
dbName = self.getDBName(ingestName)
if dbName != '' and os.path.exists(dbName):
ingestOrig = ingestName
self.logger.warning(
'dbName %s found in db, attempt to modify suffix ' %
ingestName)
sec = int(ingestName[-2:]) + 1
while (sec <= 59):
ingestName = ingestName[:-2] + "%.2d" % sec
dbName = self.getDBName(ingestName)
if not os.path.exists(dbName): break
sec = sec + 1
# our trick did not work... process will bumb
if sec == 60: ingestName = ingestOrig
# usual clients
potentials = self.clientNames + self.filterNames
# routing clients and priority (accept mask + routing info)
priority = None
if self.source.routemask:
# ingestBase is the ingestName without the postfix reception date
lst = ingestName.split(':')
pos = -1
if lst[-2] == '': pos = -2
ingestBase = ':'.join(lst[:pos])
# build the key
key = self.getRouteKey(ingestBase)
# get the clients for that key (if possible)
lst = None
potentials = []
if key != None:
lst = self.drp.getClients(key)
if lst != None:
potentials = lst
priority = self.drp.getHeaderPriority(key)
else:
self.logger.warning(
"Key not in routing table (%s from %s)" %
(key, ingestName))
else:
self.logger.warning(
"Key not generated (no accept match) with %s" % ingestName)
# ingesting the file
matchingClients = self.getMatchingClientNamesFromMasks(
ingestName, potentials)
self.logger.debug("Matching (from patterns) client names: %s" %
matchingClients)
self.ingest(file, ingestName, matchingClients, priority)
os.unlink(file)
def ingestBulletinFile(self, igniter):
from DiskReader import DiskReader
import bulletinManager
import bulletinManagerAm
from PullFTP import PullFTP
sleep_sec = 1
if self.source.type == 'pull-bulletin' or self.source.pull_script != None:
sleep_sec = self.source.pull_sleep
bullManager = bulletinManager.bulletinManager(
self.ingestDir, self.logger, self.ingestDir, 99999, '\n',
self.source.extension, self.source.routingTable,
self.source.mapEnteteDelai, self.source,
self.source.addStationInFilename)
if self.source.bulletin_type == 'am':
bullManager = bulletinManagerAm.bulletinManagerAm(
self.ingestDir, self.logger, self.ingestDir, 99999, '\n',
self.source.extension, self.source.routingTable,
self.source.addSMHeader, PXPaths.STATION_TABLE,
self.source.mapEnteteDelai, self.source,
self.source.addStationInFilename)
if self.source.nodups:
self.fileCache = CacheManager(maxEntries=self.source.cache_size,
timeout=8 * 3600)
reader = DiskReader(bullManager.pathSource, self.source.batch,
self.source.validation,
self.source.patternMatching, self.source.mtime,
False, self.source.logger, self.source.sorter,
self.source)
while True:
# If a SIGHUP signal is received ...
if igniter.reloadMode == True:
# We assign the defaults, reread configuration file for the source
# and reread all configuration file for the clients (all this in __init__)
if self.source.type == 'filter-bulletin':
self.source.__init__(self.source.name, self.source.logger,
True, True)
else:
self.source.__init__(self.source.name, self.source.logger)
bullManager = bulletinManager.bulletinManager(
self.ingestDir, self.logger, self.ingestDir, 99999, '\n',
self.source.extension, self.source.routingTable,
self.source.mapEnteteDelai, self.source,
self.source.addStationInFilename)
if self.source.bulletin_type == 'am':
bullManager = bulletinManagerAm.bulletinManagerAm(
self.ingestDir, self.logger, self.ingestDir, 99999,
'\n', self.source.extension, self.source.routingTable,
self.source.addSMHeader, PXPaths.STATION_TABLE,
self.source.mapEnteteDelai, self.source,
self.source.addStationInFilename)
if self.source.nodups:
self.fileCache = CacheManager(
maxEntries=self.source.cache_size, timeout=8 * 3600)
reader = DiskReader(bullManager.pathSource, self.source.batch,
self.source.validation,
self.source.patternMatching,
self.source.mtime, False,
self.source.logger, self.source.sorter,
self.source)
self.logger.info("Receiver has been reloaded")
igniter.reloadMode = False
# pull files in rxq directory if in pull mode
if self.source.type == 'pull-bulletin' or self.source.pull_script != None:
files = []
sleeping = os.path.isfile(PXPaths.RXQ + self.source.name +
'/.sleep')
if self.source.type == 'pull-bulletin':
puller = PullFTP(self.source, self.logger, sleeping)
files = puller.get()
puller.close()
elif self.source.pull_script != None:
files = self.source.pull_script(self.source, self.logger,
sleeping)
if not sleeping:
self.logger.debug("Number of files pulled = %s" %
len(files))
else:
self.logger.info("This pull is sleeping")
# normal diskreader call for files
reader.read()
# processing the list if necessary...
if self.source.lx_execfile != None and len(reader.sortedFiles) > 0:
sfiles = []
sfiles.extend(reader.sortedFiles)
self.logger.info("%d files process with lx_script" %
len(sfiles))
sortedFiles = self.source.run_lx_script(
sfiles, self.source.logger)
reader.sortedFiles = sortedFiles
# continue normally
data = reader.getFilesContent(reader.batch)
if len(data) == 0:
time.sleep(sleep_sec)
continue
else:
self.logger.info("%d bulletins will be ingested", len(data))
# Write (and name correctly) the bulletins to disk, erase them after
for index in range(len(data)):
# ignore duplicate if requiered
duplicate = self.source.nodups and self.fileCache.find(
data[index], 'md5') is not None
#nb_bytes = len(data[index])
#self.logger.info("Lecture de %s: %d bytes" % (reader.sortedFiles[index], nb_bytes))
if not duplicate:
# converting the file if necessary
if self.source.fx_execfile != None:
file = reader.sortedFiles[index]
fxfile = self.source.run_fx_script(
file, self.source.logger)
# convertion did not work
if fxfile == None:
self.logger.warning(
"FX script ignored the file : %s" %
os.path.basename(file))
os.unlink(file)
continue
# file already in proper format
elif fxfile == file:
self.logger.warning(
"FX script kept the file as is : %s" %
os.path.basename(file))
# file converted...
else:
self.logger.info("FX script modified %s to %s " %
(os.path.basename(file),
os.path.basename(fxfile)))
os.unlink(file)
fp = open(fxfile, 'r')
dx = fp.read()
fp.close()
reader.sortedFiles[index] = fxfile
data[index] = dx
# writing/ingesting the bulletin
if isinstance(bullManager,
bulletinManagerAm.bulletinManagerAm):
bullManager.writeBulletinToDisk(data[index], True)
else:
bullManager.writeBulletinToDisk(
data[index], True, True)
try:
file = reader.sortedFiles[index]
os.unlink(file)
if duplicate:
self.logger.info("suppressed duplicate file %s",
os.path.basename(file))
self.logger.debug("%s has been erased",
os.path.basename(file))
except OSError, e:
(type, value, tb) = sys.exc_info()
self.logger.error(
"Unable to unlink %s ! Type: %s, Value: %s" %
(reader.sortedFiles[index], type, value))
sys.path.insert(1,sys.path[0] + '/../importedLibs')
sys.path.append(sys.path[0] + "/../")
sys.path.append("/apps/pds/tools/Columbo/lib")
# Local imports
import PXPaths; PXPaths.normalPaths()
from PXManager import PXManager
from Logger import Logger
from DirectRoutingParser import DirectRoutingParser
import ColumboPath
#config = ConfigParser.ConfigParser()
#config.readfp(open(ColumboPath.FULL_MAIN_CONF))
#logname = config.get('SEARCH_AND_RESEND', 'logname')
logname = "/tmp/tmpFlow.log"
logger = Logger(logname, 'INFO', 'SAS').getLogger()
manager = PXManager()
manager.setLogger(logger)
manager.initNames()
drp = DirectRoutingParser(PXPaths.ROUTING_TABLE, [], logger)
drp.printErrors = False
drp.parseAlias()
flows = manager.getAllFlowNames(False, drp)
print " ".join(flows)
os.remove(logname)
def getAllFlows(self, noPrint=True):
if noPrint:
iprint = lambda *x: None
else:
iprint = lambda *x:sys.stdout.write(" ".join(map(str, x)) + '\n')
allSources = []
allClients = []
allSourlients = []
allFlows = []
if not os.path.isdir(self.rootPath):
return 1
for cluster in self.clusters:
pxm = PXManager(self.rootPath + cluster + '/')
if pxm.initNames():
#print (self.rootPath + cluster + " inexistant!")
continue
clients, sourlients, sources, aliases = pxm.getFlowNames(tuple=True)
# Populate flowCluster for current cluster
pxm.getFlowDict(self.sourceCluster, sources, 'source', cluster)
pxm.getFlowDict(self.clientCluster, clients, 'client', cluster)
pxm.getFlowDict(self.sourlientCluster, sourlients, 'sourlient', cluster)
allSources.extend(sources)
allClients.extend(clients)
allSourlients.extend(sourlients)
iprint("%s" % (80*'#'))
iprint("CLUSTER %s" % cluster.upper())
iprint("%s" % (80*'#'))
iprint("sources (%s): %s" % (len(sources), sources))
iprint("clients (%s): %s" % (len(clients), clients))
iprint("sourlients (%s): %s" % (len(sourlients), sourlients))
#print "aliases: %s" % aliases
iprint()
pxm = PXManager()
pxm.initNames()
self.flowCluster = self.createFlowDict()
self.dupSources = pxm.identifyDuplicate(allSources)
self.dupClients = pxm.identifyDuplicate(allClients)
self.dupSourlients = pxm.identifyDuplicate(allSourlients)
self.allSources = pxm.removeDuplicate(allSources)
self.allClients = pxm.removeDuplicate(allClients)
self.allSourlients = pxm.removeDuplicate(allSourlients)
self.allFlows.extend(allSources)
self.allFlows.extend(allClients)
self.allFlows.extend(allSourlients)
self.dupFlows = pxm.identifyDuplicate(allFlows)
self.allFlows = pxm.removeDuplicate(allFlows)
iprint("Duplicate between sources from all clusters: %s" % self.dupSources)
iprint("Duplicate between clients from all clusters: %s" % self.dupClients)
iprint("Duplicate between sourlients from all clusters: %s" % self.dupSourlients)
iprint("Duplicate beetween flows (sources, clients, sourlients) from all clusters: %s" % self.dupFlows)
iprint()
keys = self.flowCluster.keys()
keys.sort()
for key in keys:
if len(self.flowCluster[key]) > 1:
iprint("%s: %s" % (key, self.flowCluster[key]))
iprint("source cluster(%s)" % len(self.sourceCluster))
iprint(self.sourceCluster)
iprint("client cluster(%s)" % len(self.clientCluster))
iprint(self.clientCluster)
iprint("sourlient cluster(%s)" % len(self.sourlientCluster))
iprint(self.sourlientCluster)
iprint("flow cluster(%s)" % len(self.flowCluster))
iprint()
class unittest_PXManager(unittest.TestCase):
def setUp(self, logFile='log/PXManager.log'):
self.logger = Logger(logFile, 'DEBUG', 'Sub')
self.logger = self.logger.getLogger()
self.manager = PXManager()
self.manager.setLogger(self.logger)
self.manager.initNames()
self.manager.initPXPaths()
self.manager.initShouldRunNames()
self.manager.initRunningNames()
def test_PXManager(self):
self.assertEqual(self.manager.getTxNames(),
['client-am', 'client-ftp', 'client-test'])
self.assertEqual(self.manager.getFxNames(), ['filter-am'])
self.assertEqual(self.manager.getRxNames(),
['source-am', 'source-ftp', 'source-test'])
self.assertEqual(self.manager.getFlowType('client-am'),
('TX', ['client-am']))
self.assertEqual(
self.manager.getDBName(
'MAPS_12Z:OCXCARTE:CMC:CHART:IMV6::20120611192817'),
'./db/20120611/CHART/OCXCARTE/CMC/MAPS_12Z:OCXCARTE:CMC:CHART:IMV6::20120611192817'
)
self.assertEqual(self.manager.getShouldRunFxNames(), ['filter-am'])
self.assertEqual(self.manager.getShouldRunRxNames(), ['source-test'])
self.assertEqual(self.manager.getShouldRunTxNames(), ['client-test'])
self.assertEqual(self.manager.getRunningFxNames(), ['filter-am'])
self.assertEqual(self.manager.getRunningRxNames(), [])
self.assertEqual(self.manager.getRunningTxNames(), [])
self.assertEqual(self.manager.getNotRunningFxNames(), [])
self.assertEqual(self.manager.getNotRunningRxNames(), ['source-test'])
self.assertEqual(self.manager.getNotRunningTxNames(), ['client-test'])
class RequestReplyAFTN:
def __init__(self, request, addOn, sendOn, logger):
self.logger = logger
self.addOn = addOn
self.sendOn = sendOn
self.drp = DirectRoutingParser(PXPaths.ROUTING_TABLE, [], logger)
self.drp.printErrors = False
self.drp.parseAndShowErrors()
self.dbs = DBSearcher(request, False)
self.results = self.dbs.results
self.receiverName = 'request-reply'
self.pxManager = PXManager()
self.pxManager.setLogger(self.logger)
self.pxManager.initNames()
self.bulletin = ''
self.constructBulletin()
def putBulletinInQueue(self):
try:
filename = 'reply.' + dateLib.getTodayFormatted('%Y%m%d%H%M%S')
path = self.pxManager.getFlowQueueName(self.receiverName, self.drp)
file = open(path + '/' + filename, 'w')
file.write(self.bulletin)
file.close()
except:
(type, value, tb) = sys.exc_info()
self.logger.error('In putBulletinInQueue: Type = %s, Value = %s' % (type, value))
def constructBulletin(self):
self.bulletin = ''
speciResult = ''
if self.dbs.requestType == 1:
try:
file = open(self.results, 'r')
self.bulletin = file.read()
except:
(type, value, tb) = sys.exc_info()
self.logger.error('In constructBulletin(): Type = %s, Value = %s' % (type, value))
elif self.dbs.requestType == 2:
for result in self.results:
if self.dbs.type == 'SA':
station, theLine, bestHeaderTime, theFile, bestFileTime, speciLine, speciHeaderTime, speciFile, speciFileTime = result
else:
station, theLine, bestHeaderTime, theFile, bestFileTime = result
if theFile:
parts = os.path.basename(theFile).split('_')
header = parts[0] + ' ' + parts[1]
new_header = self.drp.getKeyFromHeader(header)
# We verify if we can include the result in the reply (consult the routing table)
try:
if self.sendOn not in self.drp.getHeaderClients(new_header):
self.logger.warning("%s has been excluded of reply for %s" % (new_header, self.sendOn))
continue
else:
pass
"""
clients = self.drp.getHeaderClients(header)
clients.sort()
self.logger.info("%s has been accepted in the reply for %s (in %s)" % (header, self.sendOn, str(clients)))
"""
except:
(type, value, tb) = sys.exc_info()
self.logger.error("Type: %s, Value: %s" % (type, value))
self.logger.warning("%s has been excluded of reply for %s" % (new_header, self.sendOn))
continue
self.bulletin = self.bulletin + header + ' ' + bestHeaderTime + '\n' + theLine.strip() + '\n\n'
#print repr(theLine)
if self.dbs.type == 'SA' and speciLine:
speciHeader = header[0] + 'P' + header[2:]
speciResult = speciHeader + ' ' + speciHeaderTime + '\n' + speciLine.strip() + '\n\n'
self.bulletin = speciResult + self.bulletin
if self.bulletin:
self.bulletin = self.addOn + self.bulletin
#print self.bulletin
return self.bulletin
class Ingestor(object):
"""
Normally, an Ingestor will be in a Source. It can also be used for the only reason that this object has
access to all the configuration options of the clients. For this particular case, source=None.
"""
def __init__(self, source=None, logger=None):
# General Attributes
self.source = source
self.reader = None
self.drp = None
self.count = 0
self.Mcount = 99999
if source is not None:
self.ingestDir = PXPaths.RXQ + self.source.name
if self.source.type == 'filter' or self.source.type == 'filter-bulletin' :
self.ingestDir = PXPaths.FXQ + self.source.name
self.logger = source.logger
elif logger is not None:
self.logger = logger
self.pxManager = PXManager() # Create a manager
self.pxManager.setLogger(self.logger) # Give the logger to the the manager
self.pxManager.initNames() # Set rx and tx names
self.clientNames = self.pxManager.getTxNames() # Obtains the list of client's names (the ones to wich we can link files)
self.filterNames = self.pxManager.getFxNames() # Obtains the list of filter's names (the ones to wich we can link files)
if source is not None:
if self.source.name in self.filterNames : self.filterNames.remove(self.source.name)
self.sourlientNames = self.pxManager.getTRxNames() # Obtains the list of sourlient's names (the ones to wich we can link files)
self.allNames = self.clientNames + self.filterNames + self.sourlientNames # Clients + Sourlients names
self.clients = {} # All the Client/Filter/Sourlient objects
self.fileCache = None # product processed.
self.dbDirsCache = CacheManager(maxEntries=200000, timeout=25*3600) # Directories created in the DB
self.clientDirsCache = CacheManager(maxEntries=25000, timeout=2*3600) # File ingested in RXQ
self.feedNames = [] # source to feed
self.feeds = {}
if source is not None:
self.logger.info("Ingestor (source %s) can link files to clients: %s" % (source.name, self.allNames))
def setFeeds(self, feedNames ):
from Source import Source
sources = self.pxManager.getRxNames()
for name in feedNames :
if not name in sources : continue
instant = Source(name, self.logger, False)
if instant.type == 'am' or instant.type == 'amqp' or instant.type == 'wmo' :
self.logger.warning("Feed (source %s) will be ignored (type %s)" % (name, instant.type) )
continue
self.feedNames.append(name)
self.feeds[name] = instant
self.logger.info("Ingestor (source %s) can link files to receiver: %s" % (self.source.name, self.feedNames))
def createDir(self, dir, cacheManager):
if cacheManager.find(dir) == None:
try:
os.makedirs(dir, 01775)
except OSError:
(type, value, tb) = sys.exc_info()
self.logger.debug("Problem when creating dir (%s) => Type: %s, Value: %s" % (dir, type, value))
def setClients(self):
""""
Set a dictionnary of Clients. Main usage will be to access value of
configuration options (mainly masks) of the Client objects.
"""
from Source import Source
from Sourlient import Sourlient
for name in self.clientNames:
self.clients[name] = Client(name, self.logger)
for name in self.filterNames :
self.clients[name] = Source(name, self.logger, False, True)
for name in self.sourlientNames:
self.clients[name] = Sourlient(name, self.logger, False)
#self.clients[name].readConfig()
#print self.clients[name].masks
def getIngestName(self, receptionName):
"""
Map reception name to ingest name, based on the source configuration.
This just inserts missing fields, like whattopds. DUMB!
FIXME: Have a library of functions, configurable per source, to
perform the mapping, perhaps using rmasks ? & other args.
"""
receptionNameParts = receptionName.split(':')
extensionParts = self.source.extension.split(':')
if len(receptionNameParts) > 6 :
receptionNameParts = receptionNameParts[:6]
self.logger.warning("File %s truncated to %s" % (receptionName,':'.join(receptionNameParts) ) )
for i in range(1,6):
if len(receptionNameParts) == i :
receptionNameParts = receptionNameParts + [extensionParts[i]]
elif receptionNameParts[i] == '':
receptionNameParts[i] = extensionParts[i]
receptionNameParts = receptionNameParts + [time.strftime("%Y%m%d%H%M%S", time.gmtime())]
return string.join(receptionNameParts,':')
def getClientQueueName(self, clientName, ingestName, priority=None):
"""
Return the directory into which a file of a given priority should be placed.
Layout used is: /apps/px/txq/<client>/<priority>/YYYYmmddhh
"""
parts = ingestName.split(':')
if not priority:
priority = parts[4].split('.')[0]
clientpathName = PXPaths.TXQ + clientName + '/' + str(priority) + '/' + time.strftime("%Y%m%d%H", time.gmtime()) + '/' + ingestName
if clientName in self.filterNames :
clientpathName = PXPaths.FXQ + clientName + '/' + ingestName
return clientpathName
def getDBName(self, ingestName):
"""
Given an ingest name, return a relative database name
Given a file name of the form:
what : ori_system : ori_site : data_type : format :
link it to:
db/<today>/data_type/ori_system/ori_site/ingestName
(same pattern as PDS)
NB: see notes/tests.txt for why the date/time is recalculated everytime.
"""
if ingestName.count(':') >= 4:
today = time.strftime("%Y%m%d", time.gmtime())
dirs = ingestName.split(':')
return PXPaths.DB + today + '/' + dirs[3] + '/' + dirs[1] + '/' + dirs[2] + '/' + ingestName
else:
return ''
def getRouteKey(self, filename):
"""
Given an ingest name, return a route key based on the imask given
"""
# check against the masks
for mask in self.source.masks:
# no match
if not mask[3].match(filename) : continue
# reject
if not mask[4] : return None
# accept... so key generation
parts = re.findall( mask[0], filename )
if len(parts) == 2 and parts[1] == '' : parts.pop(1)
if len(parts) != 1 : continue
key = parts[0]
if isinstance(parts[0],tuple) : key = '_'.join(parts[0])
self.logger.debug("RouteKey Key = %s Mask = %s Filename = %s" % (key,mask[0],filename) )
return key
# fallback behavior return filename
return filename
def isMatching(self, client, ingestName):
"""
Verify if ingestName is matching one mask of a client
"""
from Source import Source
if len(client.masks_deprecated) > 0 :
for mask in client.masks_deprecated:
if fnmatch.fnmatch(ingestName, mask[0]):
try:
if mask[2]:
return True
except:
return False
for mask in client.masks:
if mask[3].match(ingestName ) : return mask[4]
if isinstance(client,Source) : return True
return False
def getMatchingClientNamesFromMasks(self, ingestName, potentialClientNames):
matchingClientNames = []
for name in potentialClientNames:
try:
if self.isMatching(self.clients[name], ingestName):
matchingClientNames.append(name)
except KeyError:
pass
return matchingClientNames
def getMatchingFeedNamesFromMasks(self, ingestName, potentialFeedNames):
matchingFeedNames = []
for name in potentialFeedNames:
try:
if self.feeds[name].fileMatchMask(ingestName):
matchingFeedNames.append(name)
except KeyError:
pass
return matchingFeedNames
def ingest(self, receptionName, ingestName, clientNames, priority=None ):
self.logger.debug("Reception Name: %s" % receptionName)
dbName = self.getDBName(ingestName)
if dbName == '':
self.logger.warning('Bad ingest name (%s) => No dbName' % ingestName)
return 0
self.createDir(os.path.dirname(dbName), self.dbDirsCache)
#try:
# os.link(receptionName, dbName)
#except OSError:
# (type, value, tb) = sys.exc_info()
# self.logger.error("Unable to link %s %s, Type: %s, Value: %s" % (receptionName, dbName, type, value))
os.link(receptionName, dbName)
nbBytes = os.stat(receptionName)[stat.ST_SIZE]
if self.source.debug:
self.logger.info("DBDirsCache: %s" % self.dbDirsCache.cache)
stats, cached, total = self.dbDirsCache.getStats()
if total:
percentage = "%2.2f %% of the last %i requests were cached" % (cached/total * 100, total)
else:
percentage = "No entries in the cache"
self.logger.info("DB Caching stats: %s => %s" % (str(stats), percentage))
self.logger.debug("ClientDirsCache: %s" % self.clientDirsCache.cache)
stats, cached, total = self.clientDirsCache.getStats()
if total:
percentage = "%2.2f %% of the last %i requests were cached" % (cached/total * 100, total)
else:
percentage = "No entries in the cache"
self.logger.debug("Client Caching stats: %s => %s" % (str(stats), percentage))
self.logger.info("Ingestion Name: %s" % ingestName)
self.logger.info("(%i Bytes) Ingested in DB as %s" % (nbBytes, dbName))
# Problem bulletins are databased, but not sent to clients
if ingestName.find("PROBLEM_BULLETIN") is not -1:
return 1
for name in clientNames:
clientQueueName = self.getClientQueueName(name, ingestName, priority)
self.createDir(os.path.dirname(clientQueueName), self.clientDirsCache)
#try:
# os.link(dbName, clientQueueName)
#except OSError:
# (type, value, tb) = sys.exc_info()
# self.logger.error("Unable to link %s %s, Type: %s, Value: %s" % (dbName, clientQueueName, type, value))
os.link(dbName, clientQueueName)
feedNames = []
if len(self.feedNames) > 0 :
feedNames = self.getMatchingFeedNamesFromMasks(ingestName, self.feedNames )
self.logger.debug("Matching (from patterns) feed names: %s" % feedNames )
for name in feedNames:
if name in clientNames : continue
sourceQueueName = PXPaths.RXQ + name + '/' + ingestName
self.createDir(os.path.dirname(sourceQueueName), self.clientDirsCache)
os.link(dbName, sourceQueueName)
self.logger.info("Queued for: %s" % string.join(clientNames) + ' ' + string.join(feedNames) )
return 1
def run(self):
if self.source.type == 'single-file' or self.source.type == 'pull-file':
self.ingestSingleFile()
elif self.source.type == 'bulletin-file' or self.source.type == 'pull-bulletin':
self.ingestBulletinFile()
elif self.source.type == 'collector':
self.ingestCollection()
def ingestSingleFile(self, igniter):
from DiskReader import DiskReader
from DirectRoutingParser import DirectRoutingParser
from PullFTP import PullFTP
if self.source.routemask :
self.drp = DirectRoutingParser(self.source.routingTable, self.allNames, self.logger, self.source.routing_version)
self.drp.parse()
if self.source.nodups :
self.fileCache = CacheManager(maxEntries=self.source.cache_size, timeout=8*3600)
reader = DiskReader(self.ingestDir, self.source.batch, self.source.validation, self.source.patternMatching,
self.source.mtime, False, self.source.logger, self.source.sorter, self.source)
sleep_sec = 1
if self.source.type == 'pull-file' or self.source.pull_script != None : sleep_sec = self.source.pull_sleep
while True:
if igniter.reloadMode == True:
# We assign the defaults, reread configuration file for the source
# and reread all configuration file for the clients (all this in __init__)
if self.source.type == 'filter' :
self.source.__init__(self.source.name, self.source.logger, True, True)
else :
self.source.__init__(self.source.name, self.source.logger)
if self.source.routemask :
self.drp = DirectRoutingParser(self.source.routingTable, self.allNames, self.logger)
self.drp.parse()
if self.source.nodups :
self.fileCache = CacheManager(maxEntries=self.source.cache_size, timeout=8*3600)
reader = DiskReader(self.ingestDir, self.source.batch, self.source.validation, self.source.patternMatching,
self.source.mtime, False, self.source.logger, self.source.sorter, self.source)
self.logger.info("Receiver has been reloaded")
igniter.reloadMode = False
# pull files in rxq directory if in pull mode
if self.source.type == 'pull-file' or self.source.pull_script != None :
files = []
sleeping = os.path.isfile(PXPaths.RXQ + self.source.name + '/.sleep')
if self.source.type == 'pull-file' :
puller = PullFTP(self.source,self.logger,sleeping)
files = puller.get()
puller.close()
elif self.source.pull_script != None :
files = self.source.pull_script(self.source,self.logger,sleeping)
if not sleeping :
self.logger.debug("Number of files pulled = %s" % len(files) )
else :
self.logger.info("This pull is sleeping")
# normal diskreader call for files
reader.read()
if len(reader.sortedFiles) <= 0:
time.sleep(sleep_sec)
continue
sortedFiles = reader.sortedFiles[:self.source.batch]
# processing the list if necessary...
if self.source.lx_execfile != None :
sfiles = []
sfiles.extend(sortedFiles)
self.logger.info("%d files process with lx_script" % len(sfiles))
sortedFiles = self.source.run_lx_script(sfiles,self.source.logger)
self.logger.info("%d files will be ingested" % len(sortedFiles))
for file in sortedFiles:
self.ingestFile(file)
def ingestFile(self, file ):
# check for duplicated if user requieres
if self.source.nodups :
# get md5 key from file...
md5_key = self.fileCache.get_md5_from_file(file)
# If data is already in cache, we don't send it
if self.fileCache.find(md5_key, 'standard') is not None:
os.unlink(file)
self.logger.info("suppressed duplicate file %s", os.path.basename(file))
return
# converting the file if necessary
if self.source.fx_execfile != None :
fxfile = self.source.run_fx_script(file,self.source.logger)
# convertion did not work
if fxfile == None :
self.logger.warning("FX script ignored the file : %s" % os.path.basename(file) )
os.unlink(file)
return
# file already in proper format
elif fxfile == file :
self.logger.warning("FX script kept the file as is : %s" % os.path.basename(file) )
# file converted...
else :
self.logger.info("FX script modified %s to %s " % (os.path.basename(file),os.path.basename(fxfile)) )
os.unlink(file)
file = fxfile
# filename to ingest
ingestName = self.getIngestName(os.path.basename(file))
# make sure we do not have a dbName already in there
# if it happens add a second to the postfix datetime stamp
# until the dbName is not found
dbName = self.getDBName(ingestName)
if dbName != '' and os.path.exists(dbName) :
ingestOrig = ingestName
self.logger.warning('dbName %s found in db, attempt to modify suffix ' % ingestName)
sec = int(ingestName[-2:]) + 1
while( sec <= 59 ) :
ingestName = ingestName[:-2] + "%.2d" % sec
dbName = self.getDBName(ingestName)
if not os.path.exists(dbName) : break
sec = sec + 1
# our trick did not work... process will bumb
if sec == 60 : ingestName = ingestOrig
# usual clients
potentials = self.clientNames + self.filterNames
# routing clients and priority (accept mask + routing info)
priority = None
if self.source.routemask :
# ingestBase is the ingestName without the postfix reception date
lst = ingestName.split(':')
pos = -1
if lst[-2] == '' : pos = -2
ingestBase = ':'.join(lst[:pos])
# build the key
key = self.getRouteKey(ingestBase)
# get the clients for that key (if possible)
lst = None
potentials = []
if key != None :
lst = self.drp.getClients(key)
if lst != None :
potentials = lst
priority = self.drp.getHeaderPriority(key)
else :
self.logger.warning("Key not in routing table (%s from %s)" % (key,ingestName) )
else :
self.logger.warning("Key not generated (no accept match) with %s" % ingestName )
# ingesting the file
matchingClients = self.getMatchingClientNamesFromMasks(ingestName, potentials )
self.logger.debug("Matching (from patterns) client names: %s" % matchingClients)
self.ingest(file, ingestName, matchingClients, priority )
os.unlink(file)
def ingestBulletinFile(self, igniter):
from DiskReader import DiskReader
import bulletinManager
import bulletinManagerAm
from PullFTP import PullFTP
sleep_sec = 1
if self.source.type == 'pull-bulletin' or self.source.pull_script != None : sleep_sec = self.source.pull_sleep
bullManager = bulletinManager.bulletinManager(
self.ingestDir,
self.logger,
self.ingestDir,
99999,
'\n',
self.source.extension,
self.source.routingTable,
self.source.mapEnteteDelai,
self.source,
self.source.addStationInFilename)
if self.source.bulletin_type == 'am' :
bullManager = bulletinManagerAm.bulletinManagerAm(
self.ingestDir,
self.logger,
self.ingestDir,
99999,
'\n',
self.source.extension,
self.source.routingTable,
self.source.addSMHeader,
PXPaths.STATION_TABLE,
self.source.mapEnteteDelai,
self.source,
self.source.addStationInFilename)
if self.source.nodups :
self.fileCache = CacheManager(maxEntries=self.source.cache_size, timeout=8*3600)
reader = DiskReader(bullManager.pathSource, self.source.batch, self.source.validation, self.source.patternMatching,
self.source.mtime, False, self.source.logger, self.source.sorter, self.source)
while True:
# If a SIGHUP signal is received ...
if igniter.reloadMode == True:
# We assign the defaults, reread configuration file for the source
# and reread all configuration file for the clients (all this in __init__)
if self.source.type == 'filter-bulletin' :
self.source.__init__(self.source.name, self.source.logger, True, True)
else :
self.source.__init__(self.source.name, self.source.logger)
bullManager = bulletinManager.bulletinManager(
self.ingestDir,
self.logger,
self.ingestDir,
99999,
'\n',
self.source.extension,
self.source.routingTable,
self.source.mapEnteteDelai,
self.source,
self.source.addStationInFilename)
if self.source.bulletin_type == 'am' :
bullManager = bulletinManagerAm.bulletinManagerAm(
self.ingestDir,
self.logger,
self.ingestDir,
99999,
'\n',
self.source.extension,
self.source.routingTable,
self.source.addSMHeader,
PXPaths.STATION_TABLE,
self.source.mapEnteteDelai,
self.source,
self.source.addStationInFilename)
if self.source.nodups :
self.fileCache = CacheManager(maxEntries=self.source.cache_size, timeout=8*3600)
reader = DiskReader(bullManager.pathSource, self.source.batch, self.source.validation, self.source.patternMatching,
self.source.mtime, False, self.source.logger, self.source.sorter,self.source)
self.logger.info("Receiver has been reloaded")
igniter.reloadMode = False
# pull files in rxq directory if in pull mode
if self.source.type == 'pull-bulletin' or self.source.pull_script != None :
files = []
sleeping = os.path.isfile(PXPaths.RXQ + self.source.name + '/.sleep')
if self.source.type == 'pull-bulletin' :
puller = PullFTP(self.source,self.logger,sleeping)
files = puller.get()
puller.close()
elif self.source.pull_script != None :
files = self.source.pull_script(self.source,self.logger,sleeping)
if not sleeping :
self.logger.debug("Number of files pulled = %s" % len(files) )
else :
self.logger.info("This pull is sleeping")
# normal diskreader call for files
reader.read()
# processing the list if necessary...
if self.source.lx_execfile != None and len(reader.sortedFiles) > 0:
sfiles = []
sfiles.extend(reader.sortedFiles)
self.logger.info("%d files process with lx_script" % len(sfiles))
sortedFiles = self.source.run_lx_script(sfiles,self.source.logger)
reader.sortedFiles = sortedFiles
# continue normally
data = reader.getFilesContent(reader.batch)
if len(data) == 0:
time.sleep(sleep_sec)
continue
else:
self.logger.info("%d bulletins will be ingested", len(data))
# Write (and name correctly) the bulletins to disk, erase them after
for index in range(len(data)):
# ignore duplicate if requiered
duplicate = self.source.nodups and self.fileCache.find(data[index], 'md5') is not None
#nb_bytes = len(data[index])
#self.logger.info("Lecture de %s: %d bytes" % (reader.sortedFiles[index], nb_bytes))
if not duplicate :
# converting the file if necessary
if self.source.fx_execfile != None :
file = reader.sortedFiles[index]
fxfile = self.source.run_fx_script(file,self.source.logger)
# convertion did not work
if fxfile == None :
self.logger.warning("FX script ignored the file : %s" % os.path.basename(file) )
os.unlink(file)
continue
# file already in proper format
elif fxfile == file :
self.logger.warning("FX script kept the file as is : %s" % os.path.basename(file) )
# file converted...
else :
self.logger.info("FX script modified %s to %s " % (os.path.basename(file),os.path.basename(fxfile)) )
os.unlink(file)
fp = open(fxfile,'r')
dx = fp.read()
fp.close()
reader.sortedFiles[index] = fxfile
data[index] = dx
# writing/ingesting the bulletin
if isinstance(bullManager,bulletinManagerAm.bulletinManagerAm):
bullManager.writeBulletinToDisk(data[index], True)
else :
bullManager.writeBulletinToDisk(data[index], True, True)
try:
file = reader.sortedFiles[index]
os.unlink(file)
if duplicate : self.logger.info("suppressed duplicate file %s", os.path.basename(file))
self.logger.debug("%s has been erased", os.path.basename(file))
except OSError, e:
(type, value, tb) = sys.exc_info()
self.logger.error("Unable to unlink %s ! Type: %s, Value: %s" % (reader.sortedFiles[index], type, value))
def getAllFlows(self, noPrint=True):
if noPrint:
iprint = lambda *x: None
else:
iprint = lambda *x: sys.stdout.write(" ".join(map(str, x)) + '\n')
allSources = []
allClients = []
allSourlients = []
allFlows = []
if not os.path.isdir(self.rootPath):
return 1
for cluster in self.clusters:
pxm = PXManager(self.rootPath + cluster + '/')
if pxm.initNames():
#print (self.rootPath + cluster + " inexistant!")
continue
clients, sourlients, sources, aliases = pxm.getFlowNames(
tuple=True)
# Populate flowCluster for current cluster
pxm.getFlowDict(self.sourceCluster, sources, 'source', cluster)
pxm.getFlowDict(self.clientCluster, clients, 'client', cluster)
pxm.getFlowDict(self.sourlientCluster, sourlients, 'sourlient',
cluster)
allSources.extend(sources)
allClients.extend(clients)
allSourlients.extend(sourlients)
iprint("%s" % (80 * '#'))
iprint("CLUSTER %s" % cluster.upper())
iprint("%s" % (80 * '#'))
iprint("sources (%s): %s" % (len(sources), sources))
iprint("clients (%s): %s" % (len(clients), clients))
iprint("sourlients (%s): %s" % (len(sourlients), sourlients))
#print "aliases: %s" % aliases
iprint()
pxm = PXManager()
pxm.initNames()
self.flowCluster = self.createFlowDict()
self.dupSources = pxm.identifyDuplicate(allSources)
self.dupClients = pxm.identifyDuplicate(allClients)
self.dupSourlients = pxm.identifyDuplicate(allSourlients)
self.allSources = pxm.removeDuplicate(allSources)
self.allClients = pxm.removeDuplicate(allClients)
self.allSourlients = pxm.removeDuplicate(allSourlients)
self.allFlows.extend(allSources)
self.allFlows.extend(allClients)
self.allFlows.extend(allSourlients)
self.dupFlows = pxm.identifyDuplicate(allFlows)
self.allFlows = pxm.removeDuplicate(allFlows)
iprint("Duplicate between sources from all clusters: %s" %
self.dupSources)
iprint("Duplicate between clients from all clusters: %s" %
self.dupClients)
iprint("Duplicate between sourlients from all clusters: %s" %
self.dupSourlients)
iprint(
"Duplicate beetween flows (sources, clients, sourlients) from all clusters: %s"
% self.dupFlows)
iprint()
keys = self.flowCluster.keys()
keys.sort()
for key in keys:
if len(self.flowCluster[key]) > 1:
iprint("%s: %s" % (key, self.flowCluster[key]))
iprint("source cluster(%s)" % len(self.sourceCluster))
iprint(self.sourceCluster)
iprint("client cluster(%s)" % len(self.clientCluster))
iprint(self.clientCluster)
iprint("sourlient cluster(%s)" % len(self.sourlientCluster))
iprint(self.sourlientCluster)
iprint("flow cluster(%s)" % len(self.flowCluster))
iprint()
class RequestReplyAFTN:
def __init__(self, request, addOn, sendOn, logger):
self.logger = logger
self.addOn = addOn
self.sendOn = sendOn
self.drp = DirectRoutingParser(PXPaths.ROUTING_TABLE, [], logger)
self.drp.printErrors = False
self.drp.parseAndShowErrors()
self.dbs = DBSearcher(request, False)
self.results = self.dbs.results
self.receiverName = 'request-reply'
self.pxManager = PXManager()
self.pxManager.setLogger(self.logger)
self.pxManager.initNames()
self.bulletin = ''
self.constructBulletin()
def putBulletinInQueue(self):
try:
filename = 'reply.' + dateLib.getTodayFormatted('%Y%m%d%H%M%S')
path = self.pxManager.getFlowQueueName(self.receiverName, self.drp)
file = open(path + '/' + filename, 'w')
file.write(self.bulletin)
file.close()
except:
(type, value, tb) = sys.exc_info()
self.logger.error('In putBulletinInQueue: Type = %s, Value = %s' %
(type, value))
def constructBulletin(self):
self.bulletin = ''
speciResult = ''
if self.dbs.requestType == 1:
try:
file = open(self.results, 'r')
self.bulletin = file.read()
except:
(type, value, tb) = sys.exc_info()
self.logger.error(
'In constructBulletin(): Type = %s, Value = %s' %
(type, value))
elif self.dbs.requestType == 2:
for result in self.results:
if self.dbs.type == 'SA':
station, theLine, bestHeaderTime, theFile, bestFileTime, speciLine, speciHeaderTime, speciFile, speciFileTime = result
else:
station, theLine, bestHeaderTime, theFile, bestFileTime = result
if theFile:
parts = os.path.basename(theFile).split('_')
header = parts[0] + ' ' + parts[1]
new_header = self.drp.getKeyFromHeader(header)
# We verify if we can include the result in the reply (consult the routing table)
try:
if self.sendOn not in self.drp.getHeaderClients(
new_header):
self.logger.warning(
"%s has been excluded of reply for %s" %
(new_header, self.sendOn))
continue
else:
pass
"""
clients = self.drp.getHeaderClients(header)
clients.sort()
self.logger.info("%s has been accepted in the reply for %s (in %s)" % (header, self.sendOn, str(clients)))
"""
except:
(type, value, tb) = sys.exc_info()
self.logger.error("Type: %s, Value: %s" %
(type, value))
self.logger.warning(
"%s has been excluded of reply for %s" %
(new_header, self.sendOn))
continue
self.bulletin = self.bulletin + header + ' ' + bestHeaderTime + '\n' + theLine.strip(
) + '\n\n'
#print repr(theLine)
if self.dbs.type == 'SA' and speciLine:
speciHeader = header[0] + 'P' + header[2:]
speciResult = speciHeader + ' ' + speciHeaderTime + '\n' + speciLine.strip(
) + '\n\n'
self.bulletin = speciResult + self.bulletin
if self.bulletin:
self.bulletin = self.addOn + self.bulletin
#print self.bulletin
return self.bulletin