--- /dev/null	Thu Jan 01 00:00:00 1970 +0000
+++ b/commons/pyRepetUnit/components/AbstractClusterLauncher.py	Mon Apr 29 03:20:15 2013 -0400
@@ -0,0 +1,606 @@
+#!/usr/bin/env python
+
+# Copyright INRA (Institut National de la Recherche Agronomique)
+# http://www.inra.fr
+# http://urgi.versailles.inra.fr
+#
+# This software is governed by the CeCILL license under French law and
+# abiding by the rules of distribution of free software.  You can  use, 
+# modify and/ or redistribute the software under the terms of the CeCILL
+# license as circulated by CEA, CNRS and INRIA at the following URL
+# "http://www.cecill.info". 
+#
+# As a counterpart to the access to the source code and  rights to copy,
+# modify and redistribute granted by the license, users are provided only
+# with a limited warranty  and the software's author,  the holder of the
+# economic rights,  and the successive licensors  have only  limited
+# liability. 
+#
+# In this respect, the user's attention is drawn to the risks associated
+# with loading,  using,  modifying and/or developing or reproducing the
+# software by the user in light of its specific status of free software,
+# that may mean  that it is complicated to manipulate,  and  that  also
+# therefore means  that it is reserved for developers  and  experienced
+# professionals having in-depth computer knowledge. Users are therefore
+# encouraged to load and test the software's suitability as regards their
+# requirements in conditions enabling the security of their systems and/or 
+# data to be ensured and,  more generally, to use and operate it in the 
+# same conditions as regards security. 
+#
+# The fact that you are presently reading this means that you have had
+# knowledge of the CeCILL license and that you accept its terms.
+
+import getopt
+import time
+import glob
+import sys
+import os
+
+from commons.core.checker.CheckerException import CheckerException
+from commons.core.sql.RepetJob import RepetJob
+from commons.core.sql.Job import Job
+from commons.core.stat.Stat import Stat
+from pyRepet.launcher.AbstractProgramLauncher import AbstractProgramLauncher
+
+GENERIC_IN_FILE = "zDUMMYz"
+
+
+## Abstract class to launch a program in parallel on a cluster.
+#
+class AbstractClusterLauncher( object ):  #( IClusterLauncher )
+     
+    def __init__( self ):
+        """
+        Constructor.
+        """
+        self._inputDir = ""            # path to the directory with input files
+        self._queueName = ""           # name of the queue on the cluster
+        self._groupId = ""             # identifier of the group of jobs (groupid)
+        self._inputFileSuffix = "fa"   # suffix of the input files (default='fa')
+        self._prgAcronym = ""          # acronym of the program to launch
+        self._configFile = ""          # name of the configuration file (connect to MySQL)
+        self._currentDir = os.getcwd() # path to the current directory
+        self._tmpDir = ""              # path to the temporary directory
+        self._jobTable = "jobs"        # name of the table recording the jobs
+        self._catOutFiles = False      # concatenate output files of all jobs
+        self._clean = False            # clean job file, job stdout, job table...
+        self._verbose = 1              # verbosity level
+        self.jobdb = None              # RepetJob instance
+        self.job = Job()               # Job instance
+        
+        self._nbJobs = 0
+        self._cmdLineGenericOptions = "hi:Q:g:S:a:C:d:j:Zcv:"
+        self._cmdLineSpecificOptions = ""
+        
+        self._exeWrapper = "AbstractProgramLauncher.py"
+        self._prgLauncher = None
+        # list of instances derived from AbstractProgramLauncher()
+        # If several program are launched successively in the same job,
+        # 'lPrgLaunchers' has to be filled before run().
+        self.lPrgLaunchers = []
+    
+    def setProgramLauncherAttributeFromCmdLine(self, o, a=""):
+        self.getProgramLauncherInstance().setASpecificAttributeFromCmdLine(o, a)
+      
+    def setClusterLauncherAttributeFromCmdLine(self, o, a=""):
+        if o == "-h":
+            print self.getHelpAsString()
+            sys.exit(0)
+        elif o == "-i":
+            self.setInputDirectory(a)
+        elif o == "-Q":
+            self.setQueueName(a)
+        elif o == "-g":
+            self.setGroupIdentifier(a)
+        elif o == "-S":
+            self.setInputFileSuffix(a)
+        elif o == "-a":
+            self.setAcronym(a)
+        elif o == "-C":
+            self.setConfigFile(a)
+        elif o == "-d":
+            self.setTemporaryDirectory(a)
+        elif o == "-j":
+            self.setJobTableName(a)
+        elif o == "-Z":
+            self.setCatOutputFiles()
+        elif o == "-c":
+            self.setClean()
+        elif o == "-v":
+            self.setVerbosityLevel(a)
+
+    def setAttributesFromCmdLine(self):
+        try:
+            opts, args = getopt.getopt(sys.argv[1:], self.getCmdLineOptions())
+        except getopt.GetoptError, err:
+            print str(err);
+            print self.getHelpAsString()
+            sys.exit(1)
+        for o, a in opts:
+            self.setClusterLauncherAttributeFromCmdLine(o, a)
+            self.setProgramLauncherAttributeFromCmdLine(o, a)
+    
+    def setAGenericAttributeFromCmdLine( self, o, a="" ):
+        self.setClusterLauncherAttributeFromCmdLine(o, a)
+        
+    def setASpecificAttributeFromCmdLine( self, o, a="" ):
+        self.setProgramLauncherAttributeFromCmdLine(o, a)
+        
+    def setInputDirectory( self, arg ):
+        self._inputDir = arg
+        
+    def setQueueName( self, arg ):
+        self._queueName = arg
+        
+    def setGroupIdentifier( self, arg ):
+        self._groupId = arg
+        
+    def setInputFileSuffix( self, arg ):
+        self._inputFileSuffix = arg
+        
+    def setAcronym( self, arg ):
+        self._prgAcronym = arg
+        
+    def setConfigFile( self, arg ):
+        if os.path.dirname( arg ) == "":
+            self._configFile = "%s/%s" % ( os.getcwd(), arg )
+        else:
+            self._configFile = arg
+            
+    def setCurrentDirectory( self, arg =os.getcwd()):
+        self._currentDir = arg
+        
+    def setTemporaryDirectory( self, arg ):
+        self._tmpDir = arg
+        
+    def setJobTableName( self, arg ):
+        self._jobTable = arg
+
+    def setCatOutputFiles( self ):
+        self._catOutFiles = True
+   
+    def setClean( self):
+        self._clean = True
+        
+    def setVerbosityLevel( self, arg ):
+        self._verbose = int(arg)
+        
+    def setExecutableWrapper( self, arg = "AbstractProgramLauncher.py" ):
+        self._exeWrapper = arg
+
+    def setSingleProgramLauncher( self ):
+        """
+        Set the wrapper and program command-lines of the program launcher.
+        Append the program launcher to 'self.lPrgLaunchers'.
+        """
+        self.getProgramLauncherInstance().setWrapperCommandLine()
+        self.getProgramLauncherInstance().setProgramCommandLine()
+        self.lPrgLaunchers.append( self.getProgramLauncherInstance() )      
+        
+    def getGenericHelpAsString( self ):
+        string = ""
+        string += "usage: %s.py [options]" % (type(self).__name__ )
+        string += "\ngeneric options:"
+        string += "\n     -h: this help"
+        string += "\n     -i: directory with input files (absolute path)"
+        string += "\n     -Q: name of the queue on the cluster"
+        string += "\n     -g: identifier of the group of jobs (groupid)"
+        string += "\n     -S: suffix of the input files (default='fa')"
+        string += "\n     -a: acronym of the program to be launched (default='%s')" % ( self.getAcronym() )
+        string += "\n     -C: configuration file to connect to MySQL (absolute path or in current dir)"
+        string += "\n     -d: temporary directory (absolute path, default=None)"
+        string += "\n     -j: table recording the jobs (default='jobs')"
+        string += "\n     -c: clean the temporary data"
+        string += "\n     -v: verbosity level (default=0/1/2)"
+        return string
+        
+    def getSpecificHelpAsString( self ):
+        pass
+    
+    def getHelpAsString(self):
+        return self.getGenericHelpAsString() + self.getSpecificHelpAsString()
+    
+    def getInputDirectory( self ):
+        return self._inputDir
+        
+    def getQueueName( self ):
+        return self._queueName
+        
+    def getGroupIdentifier( self ):
+        return self._groupId
+        
+    def getInputFileSuffix( self ):
+        return self._inputFileSuffix
+    
+    def getAcronym( self ):
+        return self._prgAcronym
+        
+    def getConfigFile( self ):
+        return self._configFile
+        
+    def getCurrentDirectory( self ):
+        return self._currentDir
+        
+    def getTemporaryDirectory( self ):
+        return self._tmpDir
+        
+    def getJobTableName( self ):
+        return self._jobTable
+    
+    def getCatOutputFiles( self ):
+        return self._catOutFiles
+    
+    def getClean( self ):
+        return self._clean
+        
+    def getVerbosityLevel( self ):
+        return self._verbose
+    
+    def getWrapperName( self ):
+        return self.getProgramLauncherInstance().getWrapperName()
+    
+    def getProgramName( self ):
+        return self.getProgramLauncherInstance().getProgramName()
+    
+    def getPatternToConcatenate( self ):
+        return self.getProgramLauncherInstance().getOutputFile().replace( GENERIC_IN_FILE, "*" )
+    
+    def getProgramLauncherInstance( self ):
+        if self._prgLauncher == None:
+            self._prgLauncher = AbstractProgramLauncher()
+        return self._prgLauncher
+    
+    def getInputFilesList(self):
+        lInFiles = glob.glob("%s/*.%s" % (self._inputDir, self._inputFileSuffix))
+        return lInFiles
+    
+    def getCmdLineOptions(self):
+        return "hi:Q:g:S:a:C:d:j:Zcv:"
+
+                        
+    def getProgramCommandLineAsString( self ):
+        """
+        Return the command-line to launch in each job.
+        Specified in each wrapper.
+        """
+        pass
+    
+    
+    def getListFilesToKeep( self ):
+        """
+        Return the list of files to keep at the end of each job.
+        Specified in each wrapper.
+        """
+        pass
+    
+    
+    def getListFilesToRemove( self ):
+        """
+        Return the list of files to remove at the end of each job.
+        Specified in each wrapper.
+        """
+        pass
+    
+    
+    def getJobFileNameAsString( self, count ):
+        """
+        Return the name of the job file as a string.
+        @param count: job number (e.g. '1') or '*'
+        @type count: integer or string
+        """
+        jobFile = "ClusterLauncher"
+        jobFile += "_groupid%s" % ( self.getGroupIdentifier() )
+        if count != "*":
+            jobFile += "_job%i" % ( count )
+            jobFile += "_time%s" % ( time.strftime("%Y-%m-%d-%H-%M-%S") )
+        else:
+            jobFile += "_job*"
+            jobFile += "_time%s-*" % ( time.strftime("%Y-%m") )
+        jobFile += ".py"
+        return jobFile
+    
+    
+    def getCmdUpdateJobStatusAsString( self, newStatus ):
+        """
+        Return the command to update the job status in the table.
+        """
+        prg = os.environ["REPET_PATH"] + "/bin/srptChangeJobStatus.py"
+        cmd = prg
+        cmd += " -t %s" % ( self.job.tablename )
+        if str(self.job.jobid).isdigit():
+            cmd += " -j %s" % ( self.job.jobname )
+        else:
+            cmd += " -j %s" % ( self.job.jobid )
+        cmd += " -g %s" % ( self.job.groupid )
+        if self.job.queue != "":
+            cmd += " -q %s" % ( self.job.queue )
+        cmd += " -s %s" % ( newStatus )
+        cmd += " -c %s" % ( self.getConfigFile() )
+        cmd += " -v %i" % ( self._verbose )
+        return "os.system( \"%s\" )\n" % ( cmd )
+    
+    
+    def getCmdToLaunchWrapper( self, fileName, genericCmd, exeWrapper ):
+        """
+        Return the launching command as a string.
+        Launch the wrapper, retrieve its exit status, update status if error.
+        """
+        specificCmd = genericCmd.replace( GENERIC_IN_FILE, fileName )
+        cmd = ""
+        cmd += "print \"LAUNCH: %s\"\n" % ( specificCmd )
+        cmd += "sys.stdout.flush()\n"
+        cmd += "exitStatus = os.system ( \"%s\" )\n" % ( specificCmd )
+        cmd += "if exitStatus != 0:\n"
+        cmd += "\tprint \"ERROR: wrapper '%s'" % ( exeWrapper )
+        cmd += " returned exit status '%i'\" % ( exitStatus )\n"
+        cmd += "\tos.chdir( \"%s\" )\n" % ( self.getTemporaryDirectory() )
+        cmd += "\tshutil.move( newDir, '%s' )\n" % ( self.getCurrentDirectory() )
+        cmd += "\t%s" % ( self.getCmdUpdateJobStatusAsString( "error" ) )
+        cmd += "\tsys.exit(1)\n"
+        return cmd
+    
+    
+    def getCmdToKeepFiles( self, fileName, lFilesToKeep ):
+        """
+        Return the commands to keep the output files.
+        """
+        cmd = ""
+        for f in lFilesToKeep:
+            f = f.replace( GENERIC_IN_FILE, fileName )
+            cmd += "if not os.path.exists( \"%s\" ):\n" % ( f )
+            cmd += "\tprint \"ERROR: output file '%s' doesn't exist\"\n" % ( f )
+            cmd += "\t%s" % ( self.getCmdUpdateJobStatusAsString( "error" ) )
+            cmd += "\tsys.exit(1)\n"
+            cmd += "if not os.path.exists( \"%s/%s\" ):\n" \
+                   % ( self._currentDir, f )
+            cmd += "\tshutil.copy( \"%s\", \"%s/%s\" )\n" % ( f, self.getCurrentDirectory(), f )
+        return cmd
+    
+    
+    def getCmdToRemoveFiles( self, fileName, lFilesToRemove ):
+        """
+        Return the commands to remove the temporary files.
+        """
+        cmd = ""
+        if lFilesToRemove != []:
+            for f in lFilesToRemove:
+                f = f.replace( GENERIC_IN_FILE, fileName )
+                cmd += "if os.path.exists( \"%s\" ):\n" % ( f )
+                cmd += "\tos.remove( \"%s\" )\n" % ( f )
+        return cmd
+    
+    
+    def getJobCommandsAsString( self, fileName, jobName, minFreeGigaInTmpDir=1 ):
+        """
+        Return all the job commands as a string.
+        """
+        cmd = "#!/usr/bin/env python\n"
+        cmd += "\n"
+        cmd += "import os\n"
+        cmd += "import sys\n"
+        cmd += "import shutil\n"
+        cmd += "import time\n"
+        cmd += "\n"
+        cmd += "print \"system:\", os.uname()\n"
+        cmd += "beginTime = time.time()\n"
+        cmd += "print 'beginTime=%f' % ( beginTime )\n"
+        cmd += "\n"
+        cmd += self.getCmdUpdateJobStatusAsString( "running" )
+        cmd += "\n"
+        cmd += "if not os.path.exists( \"%s\" ):\n" % ( self.getTemporaryDirectory() )
+        cmd += "\tprint \"ERROR: working dir '%s' doesn't exist\"\n" % ( \
+            self.getTemporaryDirectory() )
+        cmd += "\t%s" % ( self.getCmdUpdateJobStatusAsString( "error" ) )
+        cmd += "\tsys.exit(1)\n"
+        cmd += "freeSpace = os.statvfs( \"%s\" )\n" % ( self.getTemporaryDirectory() )
+        cmd += "if ( freeSpace.f_bavail * freeSpace.f_frsize ) / 1073741824.0 < %i:\n" % ( minFreeGigaInTmpDir ) # nb blocs * bloc size in bytes > 1 GigaByte ?
+        cmd += "\tprint \"ERROR: less than %iGb in '%s'\"\n" % ( minFreeGigaInTmpDir, self.getTemporaryDirectory() )
+        cmd += "\t%s" % ( self.getCmdUpdateJobStatusAsString( "error" ) )
+        cmd += "\tsys.exit(1)\n"
+        cmd += "print \"working dir: %s\"\n" % ( self.getTemporaryDirectory() )
+        cmd += "sys.stdout.flush()\n"
+        cmd += "os.chdir( \"%s\" )\n" % ( self.getTemporaryDirectory() )
+        cmd += "\n"
+        cmd += "newDir = \"%s_%s_%s\"\n" % ( self.getGroupIdentifier(), jobName, time.strftime("%Y%m%d-%H%M%S") )
+        cmd += "if os.path.exists( newDir ):\n"
+        cmd += "\tshutil.rmtree( newDir )\n"
+        cmd += "os.mkdir( newDir )\n"
+        cmd += "os.chdir( newDir )\n"
+        cmd += "\n"
+        cmd += "if not os.path.exists( \"%s\" ):\n" % ( fileName )
+        cmd += "\tos.symlink( \"%s/%s\", \"%s\" )\n" % \
+               ( self._inputDir, fileName, fileName )
+        cmd += "\n"
+        
+        for pL in self.lPrgLaunchers:
+            cmd += self.getCmdToLaunchWrapper( \
+                fileName, \
+                pL.getWrapperCommandLine(), \
+                "%s.py" % ( type(pL).__name__ ) )
+            cmd += "\n"
+            cmd += self.getCmdToKeepFiles( fileName, pL.getListFilesToKeep() )
+            cmd += "\n"
+            cmd += self.getCmdToRemoveFiles( fileName, \
+                                             pL.getListFilesToRemove() )
+            
+        cmd += "if os.path.exists( \"%s\" ):\n" % ( fileName )
+        cmd += "\tos.remove( \"%s\" )\n" % ( fileName )
+        cmd += "os.chdir( \"..\" )\n"
+        cmd += "shutil.rmtree( newDir )\n"
+        cmd += self.getCmdUpdateJobStatusAsString( "finished" )
+        cmd += "\n"
+        cmd += "endTime = time.time()\n"
+        cmd += "print 'endTime=%f' % ( endTime)\n"
+        cmd += "print 'executionTime=%f' % ( endTime - beginTime )\n"
+        cmd += "print \"system:\", os.uname()\n"
+        cmd += "sys.exit(0)\n"
+        return cmd
+    
+    def getStatsOfExecutionTime( self ):
+        """
+        Return a Stat object about the execution time of each job as a
+        float expressed in seconds since the epoch, in UTC.
+        """
+        stat = Stat()
+        pattern = "%s/%s*.o*" % ( self.getCurrentDirectory(), \
+                                  self.getAcronym() )
+        lJobFiles = glob.glob( pattern )
+        for f in lJobFiles:
+            fH = open( f, "r" )
+            while True:
+                line = fH.readline()
+                if line == "":
+                    break
+                if "executionTime" in line:
+                    stat.add( float(line[:-1].split("=")[1] ) )
+                    break
+            fH.close()
+        return stat
+    
+    
+    def formatGroupidAndTime(self):
+        return self.job.groupid + " " + time.strftime("%Y-%m-%d %H:%M:%S") + ""
+
+    def submitJob(self, lInFiles):
+        count = 0
+        for inFile in lInFiles:
+            count += 1
+            fileName = os.path.basename(inFile)
+            if self._verbose > 1:
+                print "processing '%s' # %i..." % (fileName, count)
+                sys.stdout.flush()
+            
+            self.initializeJob(fileName, count)
+            time.sleep(0.5)
+            exitStatus = self.jobdb.submitJob(self.job)
+            if exitStatus != 0:
+                print "ERROR while submitting job '%i' to the cluster" % (count)
+                sys.exit(1)
+
+    def checkClusterLauncherAttributes( self ):
+        if self.getInputDirectory() == "":
+            message = "ERROR: missing input directory"
+            raise CheckerException(message)
+        if not os.path.exists( self.getInputDirectory() ):
+            message = "ERROR: input directory '%s' doesn't exist" % ( self.getInputDirectory() )
+            raise CheckerException(message)
+        if self.getGroupIdentifier() == "":
+            message = "ERROR: missing group identifier"
+            raise CheckerException(message)
+        if self.getAcronym() == "":
+            message = "ERROR: missing program acronym"
+            raise CheckerException(message)
+        if self.getConfigFile() == "":
+            message = "ERROR: missing config file to access MySQL"
+            raise CheckerException(message)
+        if not os.path.exists( self.getConfigFile() ):
+            message = "ERROR: config file '%s' doesn't exist" % ( self.getConfigFile() )
+            raise CheckerException(message)   
+        if self.getTemporaryDirectory() == "":
+            self.setTemporaryDirectory(self._currentDir)
+    
+    def checkGenericAttributes( self ):
+        self.checkClusterLauncherAttributes()
+        
+    def checkProgramLauncherAttributes( self ):
+        self.getProgramLauncherInstance().checkSpecificAttributes()
+    
+    def checkSpecificAttributes( self ):
+        self.checkProgramLauncherAttributes()
+        
+    def start( self ):
+            
+        if self.lPrgLaunchers == []:
+            self.setSingleProgramLauncher()
+        for pL in self.lPrgLaunchers:
+            if pL.getWrapperCommandLine() == "":
+                string = "ERROR: wrapper command is empty !"
+                print string
+                sys.exit(1)
+            if pL.getProgramCommandLine() == "":
+                string = "ERROR: program command is empty !"
+                print string
+                sys.exit(1)
+        self.checkProgramAvailability()
+        
+        try:
+            self.checkProgramLauncherAttributes()
+        except CheckerException, msg:
+            print msg
+            print self.getHelpAsString()
+            sys.exit(1)
+        
+        if self.getVerbosityLevel() > 0:
+            string = "START %s" % ( type(self).__name__ )
+            print string
+        self.job.tablename = self.getJobTableName()
+        self.job.groupid = self.getGroupIdentifier()
+        tokens = self.getQueueName().replace("'","").split(" ")
+        self.job.setQueue( tokens[0] )
+        if len(tokens) > 1:
+            lResources = tokens[1:]
+            self.job.lResources = lResources
+        if self.getVerbosityLevel() > 0:
+            print "groupid: %s" % ( self.getGroupIdentifier() )
+        self.jobdb = RepetJob( cfgFileName=self.getConfigFile() )
+        if self.jobdb.hasUnfinishedJob( self.job.tablename, \
+                                          self.job.groupid ):
+            self.jobdb.waitJobGroup( self.job.tablename, self.job.groupid )
+            return
+        self.jobdb.cleanJobGroup( self.job.tablename, self.job.groupid )
+        sys.stdout.flush()
+   
+    def end( self ):
+        if self.getClean():
+            self.removeAllJobFiles()
+            self.removeAllJobStdouts()
+            self.removeAllJobStderrs()
+            
+        if self.getCatOutputFiles():
+            self.catOutputFiles()
+        
+        self.jobdb.close()
+        
+        if self.getVerbosityLevel() > 0:
+            string = "END %s" % ( type(self).__name__ )
+            print string
+        sys.stdout.flush()
+        
+    def run( self ):
+        try:
+            self.checkClusterLauncherAttributes()
+        except CheckerException, msg:
+            print msg
+            print self.getHelpAsString()
+            sys.exit(1)
+            
+        self.start()
+        
+        lInFiles = self.getInputFilesList()
+        self._nbJobs = len(lInFiles)
+        
+        if self._verbose > 0:
+            string = "submitting " + str(self._nbJobs) +  " jobs... " + self.formatGroupidAndTime() 
+            print string; sys.stdout.flush()
+        
+        self.submitJob(lInFiles)
+        
+        if self._verbose > 0:        
+            string = "waiting for jobs... " + self.formatGroupidAndTime() 
+            print string; sys.stdout.flush()
+        
+        self.jobdb.waitJobGroup( self.job.tablename, self.job.groupid )
+        
+        if self._verbose > 0:
+            string = "all jobs completed ! " + self.formatGroupidAndTime() 
+            print string; sys.stdout.flush()
+            statsExecutionTime = self.getStatsOfExecutionTime()
+            print "execution time of all jobs (seconds): %f" % statsExecutionTime.getSum()
+            print "execution time per job: %s" % statsExecutionTime.string()
+            sys.stdout.flush()
+            
+        self.jobdb.cleanJobGroup( self.job.tablename, self.job.groupid )
+        
+        self.end()
+            
+