def processFilesInDirectory(dirName, anselUnicodeConverter = None, commitNonblocking =0, numThreads = 1, deleteAfterIndexing = 1):
"""processes MARC and PMS files in the indexer-queue directory. If numThreads > 1
it will try to parallelize the MARC record processing (but not PMS indexing -- no reason for that)
"""
pmsFiles = glob.glob( "%s/PMS*.DAT" % dirName )
updatedAnyRecords = 0
count= 0
for fileOn in pmsFiles:
print "processing PMS file %s" % fileOn
processedFilenameOn, deletedBibsOn = processPMSFile(fileOn)
if processedFilenameOn:
print "processing MARC file %s" % processedFilenameOn
indexerDriver.processFile( processedFilenameOn, anselUnicodeConverter )
# now that we are done processing the file, we delete it.
print "deleting MARC file %s " % processedFilenameOn
os.remove( processedFilenameOn )
print "deleting PMS file %s" % fileOn
os.remove( fileOn )
if deletedBibsOn:
print "processing deleted bibs from MARC file %s" % processedFilenameOn
for bibOn in deletedBibsOn:
print "deleting bib %s" % bibOn
indexerDriver.deleteRecord( bibOn )
updatedAnyRecords = 1
else:
print "no records to index"
os.remove( fileOn )
print "[%s] now checking for MARC files" % time.ctime()
_marcFiles = glob.glob( "%s/*.MARC" % dirName)
_marcFiles += glob.glob ("%s/*.marc" % dirName )
_marcFiles += glob.glob ("%s/*.dat" % dirName )
_marcFiles += glob.glob ("%s/*.DAT" % dirName )
_marcFiles += glob.glob( "%s/*scriblio*" % dirName )
# dedupe _marcFiles here incase a file matches more than one glob
# using a dictionary is the fastest way to dedupe a list with Jython
marcFileDict = {}
for fileOn in _marcFiles:
marcFileDict[fileOn] = None
marcFiles = marcFileDict.keys()
marcFiles.sort()
numMarcFiles = len(marcFiles)
print "[%s] found %d files to process." % (time.ctime(), numMarcFiles )
if numThreads == 1:
for fileOn in marcFiles:
print "processing MARC file %s" % fileOn
count = indexerDriver.processFile( fileOn, anselUnicodeConverter, nonblocking=1 ) # csdebug: added nonblocking here
updatedAnyRecords = 1
if deleteAfterIndexing:
os.remove( fileOn )
elif numThreads >= numMarcFiles:
# spin off a thread for each one
# was getting weird problems with multithreading (AttributeErrors when trying to iterate
# over all controlFields in MARC record -- trying separate anselUnicodeConverters to see if that's the issue.
threads = []
threadrefs = []
i = 0
for fileOn in marcFiles:
convOn = AnselToUnicode()
jobOn = indexerDriver.processFileJob( fileOn, convOn, nonblocking = 1, pid = i)# csdebug: handle nonblocking option
_threadOn = Thread( jobOn, "process file job %s" %i )
threads.append( _threadOn )
threadrefs.append( jobOn )
print "starting thread %s processing file %s" % (i, fileOn)
_threadOn.start()
i += 1
updatedAnyRecords = 1
print "joining threads"
for i in range( len(threads) ):
threads[i].join()
# TODO: make sure the thread was successful before nuking.
if deleteAfterIndexing:
print "deleting %s" % threadrefs[i].filename
os.remove( threadrefs[i].filename )
else:
# do work queue here.
print "not yet implemented"
# finally, do a commit here.
if updatedAnyRecords:
print "[%s] starting final commit" % time.ctime()
if commitNonblocking:
solrConnection.commitNonblocking()
else:
solrConnection.commit()
print "[%s] done committing" % time.ctime()
return count
fOut = open( fOutName, "w" )
fOut.write( data )
fOut.flush()
fOut.close()
else:
fOutName = None
return fOutName, deletedBibs
if __name__ == '__main__':
pmsFiles = glob.glob( "%s/PMS*.DAT" % HORIZON_BASE_DIR )
for fileOn in pmsFiles:
print "processing PMS file %s" % fileOn
processedFilenameOn, deletedBibsOn = processPMSFile(fileOn)
if processedFilenameOn:
print "processing MARC file %s" % processedFilenameOn
indexerDriver.processFile( processedFilenameOn )
# now that we are done processing the file, we delete it.
print "deleting MARC file %s " % processedFilenameOn
os.remove( processedFilenameOn )
print "deleting PMS file %s" % fileOn
os.remove( fileOn )
if deletedBibsOn:
print "processing deleted bibs from MARC file %s" % processedFilenameOn
for bibOn in deletedBibsOn:
print "deleting bib %s" % bibOn
indexerDriver.deleteRecord( bibOn )
else:
print "no records to index"
os.remove( fileOn )
def processFilesInDirectory(dirName,
anselUnicodeConverter=None,
commitNonblocking=0,
numThreads=1,
deleteAfterIndexing=1):
"""processes MARC and PMS files in the indexer-queue directory. If numThreads > 1
it will try to parallelize the MARC record processing (but not PMS indexing -- no reason for that)
"""
pmsFiles = glob.glob("%s/PMS*.DAT" % dirName)
updatedAnyRecords = 0
count = 0
for fileOn in pmsFiles:
print "processing PMS file %s" % fileOn
processedFilenameOn, deletedBibsOn = processPMSFile(fileOn)
if processedFilenameOn:
print "processing MARC file %s" % processedFilenameOn
indexerDriver.processFile(processedFilenameOn,
anselUnicodeConverter)
# now that we are done processing the file, we delete it.
print "deleting MARC file %s " % processedFilenameOn
os.remove(processedFilenameOn)
print "deleting PMS file %s" % fileOn
os.remove(fileOn)
if deletedBibsOn:
print "processing deleted bibs from MARC file %s" % processedFilenameOn
for bibOn in deletedBibsOn:
print "deleting bib %s" % bibOn
indexerDriver.deleteRecord(bibOn)
updatedAnyRecords = 1
else:
print "no records to index"
os.remove(fileOn)
print "[%s] now checking for MARC files" % time.ctime()
_marcFiles = glob.glob("%s/*.MARC" % dirName)
_marcFiles += glob.glob("%s/*.marc" % dirName)
_marcFiles += glob.glob("%s/*.dat" % dirName)
_marcFiles += glob.glob("%s/*.DAT" % dirName)
_marcFiles += glob.glob("%s/*scriblio*" % dirName)
# dedupe _marcFiles here incase a file matches more than one glob
# using a dictionary is the fastest way to dedupe a list with Jython
marcFileDict = {}
for fileOn in _marcFiles:
marcFileDict[fileOn] = None
marcFiles = marcFileDict.keys()
marcFiles.sort()
numMarcFiles = len(marcFiles)
print "[%s] found %d files to process." % (time.ctime(), numMarcFiles)
if numThreads == 1:
for fileOn in marcFiles:
print "processing MARC file %s" % fileOn
count = indexerDriver.processFile(
fileOn, anselUnicodeConverter,
nonblocking=1) # csdebug: added nonblocking here
updatedAnyRecords = 1
if deleteAfterIndexing:
os.remove(fileOn)
elif numThreads >= numMarcFiles:
# spin off a thread for each one
# was getting weird problems with multithreading (AttributeErrors when trying to iterate
# over all controlFields in MARC record -- trying separate anselUnicodeConverters to see if that's the issue.
threads = []
threadrefs = []
i = 0
for fileOn in marcFiles:
convOn = AnselToUnicode()
jobOn = indexerDriver.processFileJob(
fileOn, convOn, nonblocking=1,
pid=i) # csdebug: handle nonblocking option
_threadOn = Thread(jobOn, "process file job %s" % i)
threads.append(_threadOn)
threadrefs.append(jobOn)
print "starting thread %s processing file %s" % (i, fileOn)
_threadOn.start()
i += 1
updatedAnyRecords = 1
print "joining threads"
for i in range(len(threads)):
threads[i].join()
# TODO: make sure the thread was successful before nuking.
if deleteAfterIndexing:
print "deleting %s" % threadrefs[i].filename
os.remove(threadrefs[i].filename)
else:
# do work queue here.
print "not yet implemented"
# finally, do a commit here.
if updatedAnyRecords:
print "[%s] starting final commit" % time.ctime()
if commitNonblocking:
solrConnection.commitNonblocking()
else:
solrConnection.commit()
print "[%s] done committing" % time.ctime()
return count
fOut.write(data)
fOut.flush()
fOut.close()
else:
fOutName = None
return fOutName, deletedBibs
if __name__ == '__main__':
pmsFiles = glob.glob("%s/PMS*.DAT" % HORIZON_BASE_DIR)
for fileOn in pmsFiles:
print "processing PMS file %s" % fileOn
processedFilenameOn, deletedBibsOn = processPMSFile(fileOn)
if processedFilenameOn:
print "processing MARC file %s" % processedFilenameOn
indexerDriver.processFile(processedFilenameOn)
# now that we are done processing the file, we delete it.
print "deleting MARC file %s " % processedFilenameOn
os.remove(processedFilenameOn)
print "deleting PMS file %s" % fileOn
os.remove(fileOn)
if deletedBibsOn:
print "processing deleted bibs from MARC file %s" % processedFilenameOn
for bibOn in deletedBibsOn:
print "deleting bib %s" % bibOn
indexerDriver.deleteRecord(bibOn)
else:
print "no records to index"
os.remove(fileOn)
