def extract_seq(args):
# Download the file from Shock to the working directory.
if args['nodeId'] is not None:
shockClient = ShockClient(args['shockUrl'], args['auth'])
shockClient.download_to_path(args['nodeId'], args['sourceFile'])
# Extract the sequences from the source file.
numReads = 0
with open(args['destFile'], 'w') as f:
if args['sequenceLen'] > 0: # A length to trim to was specified
for seqRecord in SeqIO.parse(args['sourceFile'], args['format']):
seq = str(seqRecord.seq)
if len(seq) < args['sequenceLen']:
continue
if len(seq) > args['sequenceLen']:
seq = seq[:args['sequenceLen']]
f.write(str(seq) + '\n')
numReads += 1
if numReads == args['maxReads']:
break
elif args['maxReads'] > 0:
for seqRecord in SeqIO.parse(args['sourceFile'], args['format']):
f.write(str(seqRecord.seq) + '\n')
numReads += 1
if numReads == args['maxReads']:
break
else:
for seqRecord in SeqIO.parse(args['sourceFile'], args['format']):
f.write(str(seqRecord.seq) + '\n')
# Delete the file if it does not have enough reads.
if args['minReads'] > 0 and numReads < args['minReads']:
os.remove(args['destFile'])
return 0
def extract_seq(args):
# Download the file from Shock to the working directory.
if args['nodeId'] is not None:
shockClient = ShockClient(args['shockUrl'], args['auth'])
shockClient.download_to_path(args['nodeId'], args['sourceFile'])
# Extract the sequences from the source file.
numReads = 0
with open(args['destFile'], 'w') as f:
if args['sequenceLen'] > 0: # A length to trim to was specified
for seqRecord in SeqIO.parse(args['sourceFile'], args['format']):
seq = str(seqRecord.seq)
if len(seq) < args['sequenceLen']:
continue
if len(seq) > args['sequenceLen']:
seq = seq[:args['sequenceLen']]
f.write(str(seq) + '\n')
numReads += 1
if numReads == args['maxReads']:
break
elif args['maxReads'] > 0:
for seqRecord in SeqIO.parse(args['sourceFile'], args['format']):
f.write(str(seqRecord.seq) + '\n')
numReads += 1
if numReads == args['maxReads']:
break
else:
for seqRecord in SeqIO.parse(args['sourceFile'], args['format']):
f.write(str(seqRecord.seq) + '\n')
# Delete the file if it does not have enough reads.
if args['minReads'] > 0 and numReads < args['minReads']:
os.remove(args['destFile'])
return 0
def loadDatabaseFiles(self, mylog):
''' Load the static database files from Shock.
The static database files are stored in the directory specified by the
data_folder_path configuration variable. A file is only downloaded if
the file is not available on this system or the file has been updated
in Shock.
@param mylog Log object for messages
@return Nothing
@raise MissingFileError when database file is not found in Shock
'''
# Get the current info about the static database files from the cache file.
cacheFilename = self.StatusFiles['cache_file']
if os.path.exists(cacheFilename):
fileCache = json.load(open(cacheFilename, "r"))
else:
fileCache = dict()
# Create a shock client.
shockClient = ShockClient(self.shockURL)
# See if the static database files on this system are up-to-date with files stored in Shock.
shockFiles = dict(self.DataFiles.items() + self.SearchFiles.items())
for key in shockFiles:
# Get info about the file stored in Shock.
localPath = shockFiles[key]
name = os.path.basename(localPath)
nodelist = shockClient.query_node(
{'lookupname': 'ProbAnnoData/' + name})
if len(nodelist) == 0:
message = "Database file %s is not available from %s\n" % (
name, self.shockURL)
mylog.log_message(log.ERR, message) # MBM
raise MissingFileError(message)
node = nodelist[0]
# Download the file if the checksum does not match or the file is not available on this system.
download = False
if key in fileCache:
if node['file']['checksum']['md5'] != fileCache[key]['file'][
'checksum']['md5']:
download = True
else:
download = True
if os.path.exists(localPath) == False:
download = True
if download:
sys.stderr.write("Downloading %s to %s\n" % (key, localPath))
shockClient.download_to_path(node["id"], localPath)
fileCache[key] = node
mylog.log_message(log.INFO,
'Downloaded %s to %s' % (key, localPath))
# Save the updated cache file.
json.dump(fileCache, open(cacheFilename, "w"), indent=4)
return
def loadDatabaseFiles(self, mylog):
''' Load the static database files from Shock.
The static database files are stored in the directory specified by the
data_folder_path configuration variable. A file is only downloaded if
the file is not available on this system or the file has been updated
in Shock.
@param mylog Log object for messages
@return Nothing
@raise MissingFileError when database file is not found in Shock
'''
# Get the current info about the static database files from the cache file.
cacheFilename = self.StatusFiles['cache_file']
if os.path.exists(cacheFilename):
fileCache = json.load(open(cacheFilename, "r"))
else:
fileCache = dict()
# Create a shock client.
shockClient = ShockClient(self.shockURL)
# See if the static database files on this system are up-to-date with files stored in Shock.
shockFiles = dict(self.DataFiles.items() + self.SearchFiles.items())
for key in shockFiles:
# Get info about the file stored in Shock.
localPath = shockFiles[key]
name = os.path.basename(localPath)
nodelist = shockClient.query_node( { 'lookupname': 'ProbAnnoData/'+name } )
if len(nodelist) == 0:
message = "Database file %s is not available from %s\n" %(name, self.shockURL)
mylog.log_message(log.ERR, message) # MBM
raise MissingFileError(message)
node = nodelist[0]
# Download the file if the checksum does not match or the file is not available on this system.
download = False
if key in fileCache:
if node['file']['checksum']['md5'] != fileCache[key]['file']['checksum']['md5']:
download = True
else:
download = True
if os.path.exists(localPath) == False:
download = True
if download:
sys.stderr.write("Downloading %s to %s\n" %(key, localPath))
shockClient.download_to_path(node["id"], localPath)
fileCache[key] = node
mylog.log_message(log.INFO, 'Downloaded %s to %s' %(key, localPath))
# Save the updated cache file.
json.dump(fileCache, open(cacheFilename, "w"), indent=4)
return
def test_buildmatrix(self):
''' Run build_matrix() with four simple sequence files and verify the returned distance matrix.'''
# Create a client.
cbdClient = CompressionBasedDistance(self._config['cbd_url'], user_id=self._config['test_user'], password=self._config['test_pwd'])
token = cbdClient._headers['AUTHORIZATION']
# Create the input parameters.
input = dict()
input['format'] = 'fasta'
input['scale'] = 'std'
input['sequence_length'] = 0
input['min_reads'] = 0
input['max_reads'] = 0
input['extreme'] = 1
input['node_ids'] = list()
# Upload the files to Shock.
shockClient = ShockClient(self._config['shock_url'], token)
for filename in InputFiles:
node = shockClient.create_node(filename, '')
input['node_ids'].append(node['id'])
# Run the buildmatrix() function to generate a distance matrix.
jobid = cbdClient.build_matrix(input)
# Wait for the distance matrix to be built.
time.sleep(30)
# Get the distance matrix and save to a file.
outputPath = 'client-tests/unittest.csv'
args = [ os.path.join(os.environ['KB_TOP'], 'bin/cbd-getmatrix'), jobid, outputPath ]
proc = subprocess.Popen(args, stdout = subprocess.PIPE, stderr = subprocess.PIPE)
(so, se) = proc.communicate()
if proc.returncode != 0:
print so
print se
self.assertEqual(proc.returncode, 0)
# Confirm the returned distance matrix matches the saved valid output.
vf = open('client-tests/output.csv', 'r')
tf = open(outputPath, 'r')
for vline in vf:
tline = tf.readline()
self.assertEqual(vline, tline)
self.assertEqual(tf.readline(), '')
vf.close()
tf.close()
os.remove(outputPath)
def storeDatabaseFiles(self, token):
''' Store the static database files to Shock.
@param token: Authorization token for authenticating to shock
@return Nothing
'''
# Create a shock client.
shockClient = ShockClient(self.shockURL, token=token)
# Upload all of the static database files to shock.
fileCache = dict()
shockFiles = dict(self.DataFiles.items() + self.SearchFiles.items())
for key in shockFiles:
localPath = shockFiles[key]
name = os.path.basename(localPath)
if os.path.exists(localPath):
sys.stderr.write('Saving "%s"...' %(localPath))
# See if the file already exists in Shock.
query = { 'lookupname': 'ProbAnnoData/'+name }
nodelist = shockClient.query_node(query)
# Remove all instances of the file in Shock.
if nodelist != None:
for node in nodelist:
shockClient.delete_node(node['id'])
# Build the attributes for this file and store as json in a separate file.
moddate = time.ctime(os.path.getmtime(localPath))
attr = { 'lookupname': 'ProbAnnoData/'+name, 'moddate': moddate }
attrFilename = os.path.join(self.dataFolderPath, name+'.attr')
attrFid = open(attrFilename, 'w')
json.dump(attr, attrFid, indent=4)
attrFid.close()
# Upload the file to Shock.
metadata = shockClient.create_node(localPath, attrFilename)
fileCache[key] = metadata
os.remove(attrFilename)
# Remove the list of users from the read ACL to give the file public read permission.
# Note this needs to change for Shock version 0.9.5 but not sure how to set public ACLs.
readacl = shockClient.get_acl(metadata['id'])
shockClient.delete_acl(metadata['id'], 'read', readacl['read'][0])
sys.stderr.write('done\n')
else:
sys.stderr.write('Could not find "%s" so it was not saved\n' %(localPath))
# Save the metadata on all of the database files.
cacheFilename = os.path.join(self.dataFolderPath, StatusFiles['cache_file'])
json.dump(fileCache, open(cacheFilename, 'w'), indent=4)
return
def storeDatabaseFiles(self, token):
''' Store the static database files to Shock.
@param token: Authorization token for authenticating to shock
@return Nothing
'''
# Create a shock client.
shockClient = ShockClient(self.shockURL, token=token)
# Upload all of the static database files to shock.
fileCache = dict()
shockFiles = dict(self.DataFiles.items() + self.SearchFiles.items())
for key in shockFiles:
localPath = shockFiles[key]
name = os.path.basename(localPath)
if os.path.exists(localPath):
sys.stderr.write('Saving "%s"...' % (localPath))
# See if the file already exists in Shock.
query = {'lookupname': LOOKUP_NAME_PREFIX + '/' + name}
nodelist = shockClient.query_node(query)
# Remove all instances of the file in Shock.
if nodelist != None:
for node in nodelist:
shockClient.delete_node(node['id'])
# Build the attributes for this file and store as json in a separate file.
moddate = time.ctime(os.path.getmtime(localPath))
attr = {
'lookupname': LOOKUP_NAME_PREFIX + '/' + name,
'moddate': moddate
}
attrFilename = os.path.join(self.dataFolderPath,
name + '.attr')
attrFid = open(attrFilename, 'w')
json.dump(attr, attrFid, indent=4)
attrFid.close()
# Upload the file to Shock.
metadata = shockClient.create_node(localPath, attrFilename)
fileCache[key] = metadata
os.remove(attrFilename)
# Remove the list of users from the read ACL to give the file public read permission.
# Note this needs to change for Shock version 0.9.5 but not sure how to set public ACLs.
readacl = shockClient.get_acl(metadata['id'])
shockClient.delete_acl(metadata['id'], 'read',
readacl['read'][0])
sys.stderr.write('done\n')
else:
sys.stderr.write('Could not find "%s" so it was not saved\n' %
(localPath))
# Save the metadata on all of the database files.
cacheFilename = os.path.join(self.dataFolderPath,
self.StatusFiles['cache_file'])
json.dump(fileCache, open(cacheFilename, 'w'), indent=4)
return
input['sequence_length'] = args.sequenceLen
input['min_reads'] = args.minReads
input['max_reads'] = args.maxReads
if args.extreme:
input['extreme'] = 1
else:
input['extreme'] = 0
input['node_ids'] = list()
# Create a cbd client (which must be authenticated).
if args.url is None:
args.url = get_url()
cbdClient = CompressionBasedDistance(url=args.url)
# Create a shock client.
shockClient = ShockClient(args.shockurl,
cbdClient._headers['AUTHORIZATION'])
# Parse the input file with the list of sequence files.
(fileList, extensions, numMissingFiles) = parse_input_file(args.inputPath)
if numMissingFiles > 0:
exit(1)
# Set the format based on the sequence file extension if the format argument was not specified.
if args.format is None:
if len(extensions) == 1:
input['format'] = extensions.keys()[0]
else:
print "The format of the sequence files could not be determined. Set the format with the --format argument."
exit(1)
else:
input['format'] = args.format
class CompressionBasedDistance:
''' Calculate the compression based distance metric and save distance matrix to a file.
@param fileList List of paths to compressed files
@param scale Scale of distance values, 'std' for 0 to 1, 'inf' for 0 to infinity
@param outputFile Path to file with output distance matrix
@return Nothing
'''
def _cbdCalculator(self, fileList, scale, outputFile):
# Parse the files.
single_sizes = dict()
pair_sizes = dict()
for sourceFile in fileList:
# Should strip prefix too
fbase = os.path.basename(sourceFile)
# This works as long as '.sorted.xz' only occurs at the end of the path.
fname = fbase.replace('.sorted.xz', '')
if PairSeparator in fname:
pair_sizes[fname] = os.path.getsize(sourceFile)
else:
single_sizes[fname] = os.path.getsize(sourceFile)
# Map file names to indices.
fnames = single_sizes.keys()
fnames.sort()
indices = dict()
for name, i in zip(fnames, range(len(fnames))):
indices[name] = i
# Compute the distance scores.
pair_names = pair_sizes.keys()
cbd_array = numpy.zeros((len(fnames), len(fnames)), dtype=float)
for pair in pair_names:
name1, name2 = pair.split(PairSeparator)
c1 = float(single_sizes[name1])
c2 = float(single_sizes[name2])
c12 = float(pair_sizes[pair])
distance = 1.0 - (2.0 * ((c1 + c2 - c12) / (c1 + c2)))
if distance > 1.0:
part1 = "Distance %f is greater than 1.0. " % (distance)
part2 = "Check sequence read lengths and relative number of sequence reads. "
part3 = "(c1=%f %s, c2=%f %s c12=%f %s)" % (c1, name1, c2,
name2, c12, pair)
raise ValueError(part1 + part2 + part3)
if scale == 'inf':
distance = distance / (1.0 - distance)
cbd_array[indices[name1], indices[name2]] = distance
cbd_array[indices[name2], indices[name1]] = distance
# Build the output file in CSV format.
outf = open(outputFile, 'w')
outf.write('ID,' + ','.join(fnames) + '\n')
for i in range(len(fnames)):
outf.write(fnames[i] + ',' +
','.join(['{0:g}'.format(x)
for x in cbd_array[i, :]]) + '\n')
outf.close()
return
''' Cleanup after running a job.
@note All temporary files are removed even when there is an error.
@return Nothing
'''
def _cleanup(self):
# Delete input fasta files from Shock.
for nodeId in self.input['node_ids']:
try:
self.shockClient.delete_node(nodeId)
except Exception as e:
self._log(
log.ERR, 'Error deleting node %s from Shock: %s' %
(nodeId, e.message))
# Remove the work directory.
shutil.rmtree(self.jobDirectory)
# Stop the process pool.
self.pool.close()
self.pool.join()
return
''' Log a message to the system log.
@param level Message level (INFO, WARNING, etc.)
@param message Message text
@return Nothing
'''
def _log(self, level, message):
# Create a logger if this is the first time the method has been called.
if self.logger is None:
submod = os.environ.get('KB_SERVICE_NAME',
'CompressionBasedDistance')
self.logger = log.log(submod,
ip_address=True,
authuser=True,
module=True,
method=True,
call_id=True,
config=os.getenv('KB_DEPLOYMENT_CONFIG'))
# Log the message.
self.logger.log_message(level, message, self.context['client_ip'],
self.context['user_id'],
self.context['module'], self.context['method'],
self.context['call_id'])
return
def __init__(self):
self.logger = None
''' Run a job to build a distance matrix.
When successful the distance matrix csv file is stored in Shock.
@param job Dictionary with configuration variables, context variables, and input variables for job
@raise ExtractError: Error extracting sequences from input sequence file
@raise SeqLenError: Error with lengths of sequences in input sequence file
@raise SortError: Error sorting a raw sequence file
@raise MergeError: Error merging a raw sequence file
@raise CompressError: Error compressing a raw sequence file
@raise ShockError: Error saving file to Shock
@return Nothing
'''
def runJob(self, job):
self.config = job['config']
self.context = job['context']
self.input = job['input']
# Create a shock client and authenticate as the user.
self.shockClient = ShockClient(self.config['shock_url'],
self.context['token'])
# Create a user and job state client and authenticate as the user.
ujsClient = UserAndJobState(self.config['userandjobstate_url'],
token=self.context['token'])
# Create a process pool.
self.pool = Pool(processes=int(self.config['num_pool_processes']))
# Create a work directory for storing intermediate files.
self.jobDirectory = make_job_dir(self.config['work_folder_path'],
job['id'])
self._log(
log.INFO, 'Job ' + job['id'] + ' running with work folder ' +
self.jobDirectory)
# Download input fasta files from Shock and extract sequences to work directory.
try:
ujsClient.update_job_progress(job['id'], self.context['token'],
'extracting sequence files', 1,
timestamp(3600))
except:
pass
resultList = []
sequenceList = []
for nodeId in self.input['node_ids']:
node = self.shockClient.get_node(nodeId)
sourceFile = os.path.join(self.jobDirectory, node['file']['name'])
destFile = '%s.sequence' % (os.path.splitext(sourceFile)[0])
if PairSeparator in destFile: # Check for pair separator string in file name and replace as needed.
destFile = destFile.replace(PairSeparator, '-')
sequenceList.append(destFile)
args = dict(
) # Needs to be scoped here so each process gets its own copy
args['format'] = self.input['format']
args['shockUrl'] = self.config['shock_url']
args['auth'] = self.context['token']
args['sequenceLen'] = self.input['sequence_length']
args['minReads'] = self.input['min_reads']
args['maxReads'] = self.input['max_reads']
args['nodeId'] = nodeId
args['sourceFile'] = sourceFile
args['destFile'] = destFile
result = self.pool.apply_async(extract_seq, (args, ))
resultList.append(result)
for result in resultList:
if result.get() != 0:
self._cleanup()
raise ExtractError(
"Error extracting sequences from input sequence file, result: %d"
% (result.get()))
for path in self.input['file_paths']:
sourceFile = os.path.basename(path)
destFile = '%s/%s.sequence' % (self.jobDirectory,
os.path.splitext(sourceFile)[0])
if PairSeparator in destFile: # Check for pair separator string in file name and replace as needed.
destFile = destFile.replace(PairSeparator, '-')
sequenceList.append(destFile)
args = dict(
) # Needs to be scoped here so each process gets its own copy
args['format'] = self.input['format']
args['shockUrl'] = self.config['shock_url']
args['auth'] = self.context['token']
args['sequenceLen'] = self.input['sequence_length']
args['minReads'] = self.input['min_reads']
args['maxReads'] = self.input['max_reads']
args['nodeId'] = None
args['sourceFile'] = path
args['destFile'] = destFile
result = self.pool.apply_async(extract_seq, (args, ))
resultList.append(result)
for result in resultList:
try:
result.get()
except Exception as e:
self._cleanup()
raise ExtractError(
"Error extracting sequences from input sequence file: %s" %
(e.message))
# Confirm that each file met the criteria for sequence length and number of sequences.
filesToRemove = list()
for index in range(len(sequenceList)):
# See if the file did not have the minimum number of sequences.
if not os.path.exists(sequenceList[index]):
filesToRemove.append(index)
continue
# See if the file has no data.
if os.path.getsize(sequenceList[index]) == 0:
self._cleanup()
raise SeqLenError("Sequence file '%s' has no sequences" %
(sequenceList[index]))
filteredList = list()
for index in range(len(sequenceList)):
if index not in filesToRemove:
filteredList.append(sequenceList[index])
if len(filteredList) < 2:
self._cleanup()
raise SeqLenError(
"There are not enough sequence files that meet the sequence length or number of sequences criteria."
)
# Sort the sequences.
try:
ujsClient.update_job_progress(job['id'], self.context['token'],
'sorting sequence files', 1,
timestamp(3600))
except:
pass
resultList = []
sortedList = []
for sourceFile in filteredList:
destFile = '%s.sorted' % (os.path.splitext(sourceFile)[0])
sortedList.append(destFile)
args = ['/usr/bin/sort', '--output=%s' % (destFile), sourceFile]
result = self.pool.apply_async(run_command, (args, ))
resultList.append(result)
for result in resultList:
try:
result.get()
except CommandError as e:
self._cleanup()
raise SortError(
"Error sorting sequence file: %s\nCommand: '%s'\nStdout: '%s'\nStderr: '%s'"
% (e.message, e.cmd, e.stdout, e.stderr))
# Create combined and sorted files.
try:
ujsClient.update_job_progress(
job['id'], self.context['token'],
'merging all pairs of sequence files', 1, timestamp(3600))
except:
pass
resultList = []
for p, q in combinations(sortedList, 2):
pbase = os.path.basename(p)
qbase = os.path.basename(q)
dbase = '%s%s%s.sorted' % (os.path.splitext(pbase)[0],
PairSeparator,
os.path.splitext(qbase)[0])
destFile = os.path.join(self.jobDirectory, dbase)
sortedList.append(destFile)
args = ['/usr/bin/sort', '-m', '--output=%s' % (destFile), p, q]
result = self.pool.apply_async(run_command, (args, ))
resultList.append(result)
for result in resultList:
try:
result.get()
except CommandError as e:
self._cleanup()
raise MergeError(
"Error merging sequence file: %s\nCommand: '%s'\nStdout: '%s'\nStderr: '%s'"
% (e.message, e.cmd, e.stdout, e.stderr))
# Compress all sorted files.
try:
ujsClient.update_job_progress(job['id'], self.context['token'],
'compressing sequence files', 1,
timestamp(3600))
except:
pass
resultList = []
compressedList = []
for sourceFile in sortedList:
compressedList.append(sourceFile + '.xz')
if self.input['extreme']:
level = '-9e'
else:
level = '-9'
args = ['/usr/bin/xz', '--keep', level, '--no-warn', sourceFile]
result = self.pool.apply_async(run_command, (args, ))
resultList.append(result)
for result in resultList:
try:
result.get()
except CommandError as e:
self._cleanup()
raise CompressError(
"Error compressing sequence file: %s\nCommand: '%s'\nStdout: '%s'\nStderr: '%s'"
% (e.message, e.cmd, e.stdout, e.stderr))
# Calculate the distance matrix.
try:
ujsClient.update_job_progress(job['id'], self.context['token'],
'calculating distance matrix', 1,
timestamp(3600))
except:
pass
csvFile = os.path.join(self.jobDirectory, '%s.csv' % (job['id']))
self._cbdCalculator(compressedList, self.input['scale'], csvFile)
# Store the output file in shock.
try:
ujsClient.update_job_progress(job['id'], self.context['token'],
'storing output file in shock', 1,
timestamp(3600))
except:
pass
node = self.shockClient.create_node(csvFile, '')
if not node['id']:
# Shock let us down. Save the distance matrix in the work directory for possible recovery.
os.rename(
csvFile,
'%s/%s.csv' % (self.config['work_folder_path'], job['id']))
self._cleanup()
raise ShockError(
"Error saving distance matrix file to Shock. A Shock node was not created."
)
# Mark the job as complete.
results = {
'shocknodes': [node['id']],
'shockurl': self.config['shock_url']
}
ujsClient.complete_job(job['id'], self.context['token'], 'done', None,
results)
self._log(log.INFO, 'Job ' + job['id'] + ' completed successfully')
# Cleanup after ourselves.
self._cleanup()
return
def calculate(self, listFilePath, scale, csvFile):
# Each line of the list file is a path to a compressed file.
compressedList = list()
listFile = open(listFilePath, 'r')
for line in listFile:
compressedList.append(line.strip())
listFile.close()
# Calculate the distance matrix.
self._cbdCalculator(compressedList, scale, csvFile)
return
def runJob(self, job):
self.config = job['config']
self.context = job['context']
self.input = job['input']
# Create a shock client and authenticate as the user.
self.shockClient = ShockClient(self.config['shock_url'],
self.context['token'])
# Create a user and job state client and authenticate as the user.
ujsClient = UserAndJobState(self.config['userandjobstate_url'],
token=self.context['token'])
# Create a process pool.
self.pool = Pool(processes=int(self.config['num_pool_processes']))
# Create a work directory for storing intermediate files.
self.jobDirectory = make_job_dir(self.config['work_folder_path'],
job['id'])
self._log(
log.INFO, 'Job ' + job['id'] + ' running with work folder ' +
self.jobDirectory)
# Download input fasta files from Shock and extract sequences to work directory.
try:
ujsClient.update_job_progress(job['id'], self.context['token'],
'extracting sequence files', 1,
timestamp(3600))
except:
pass
resultList = []
sequenceList = []
for nodeId in self.input['node_ids']:
node = self.shockClient.get_node(nodeId)
sourceFile = os.path.join(self.jobDirectory, node['file']['name'])
destFile = '%s.sequence' % (os.path.splitext(sourceFile)[0])
if PairSeparator in destFile: # Check for pair separator string in file name and replace as needed.
destFile = destFile.replace(PairSeparator, '-')
sequenceList.append(destFile)
args = dict(
) # Needs to be scoped here so each process gets its own copy
args['format'] = self.input['format']
args['shockUrl'] = self.config['shock_url']
args['auth'] = self.context['token']
args['sequenceLen'] = self.input['sequence_length']
args['minReads'] = self.input['min_reads']
args['maxReads'] = self.input['max_reads']
args['nodeId'] = nodeId
args['sourceFile'] = sourceFile
args['destFile'] = destFile
result = self.pool.apply_async(extract_seq, (args, ))
resultList.append(result)
for result in resultList:
if result.get() != 0:
self._cleanup()
raise ExtractError(
"Error extracting sequences from input sequence file, result: %d"
% (result.get()))
for path in self.input['file_paths']:
sourceFile = os.path.basename(path)
destFile = '%s/%s.sequence' % (self.jobDirectory,
os.path.splitext(sourceFile)[0])
if PairSeparator in destFile: # Check for pair separator string in file name and replace as needed.
destFile = destFile.replace(PairSeparator, '-')
sequenceList.append(destFile)
args = dict(
) # Needs to be scoped here so each process gets its own copy
args['format'] = self.input['format']
args['shockUrl'] = self.config['shock_url']
args['auth'] = self.context['token']
args['sequenceLen'] = self.input['sequence_length']
args['minReads'] = self.input['min_reads']
args['maxReads'] = self.input['max_reads']
args['nodeId'] = None
args['sourceFile'] = path
args['destFile'] = destFile
result = self.pool.apply_async(extract_seq, (args, ))
resultList.append(result)
for result in resultList:
try:
result.get()
except Exception as e:
self._cleanup()
raise ExtractError(
"Error extracting sequences from input sequence file: %s" %
(e.message))
# Confirm that each file met the criteria for sequence length and number of sequences.
filesToRemove = list()
for index in range(len(sequenceList)):
# See if the file did not have the minimum number of sequences.
if not os.path.exists(sequenceList[index]):
filesToRemove.append(index)
continue
# See if the file has no data.
if os.path.getsize(sequenceList[index]) == 0:
self._cleanup()
raise SeqLenError("Sequence file '%s' has no sequences" %
(sequenceList[index]))
filteredList = list()
for index in range(len(sequenceList)):
if index not in filesToRemove:
filteredList.append(sequenceList[index])
if len(filteredList) < 2:
self._cleanup()
raise SeqLenError(
"There are not enough sequence files that meet the sequence length or number of sequences criteria."
)
# Sort the sequences.
try:
ujsClient.update_job_progress(job['id'], self.context['token'],
'sorting sequence files', 1,
timestamp(3600))
except:
pass
resultList = []
sortedList = []
for sourceFile in filteredList:
destFile = '%s.sorted' % (os.path.splitext(sourceFile)[0])
sortedList.append(destFile)
args = ['/usr/bin/sort', '--output=%s' % (destFile), sourceFile]
result = self.pool.apply_async(run_command, (args, ))
resultList.append(result)
for result in resultList:
try:
result.get()
except CommandError as e:
self._cleanup()
raise SortError(
"Error sorting sequence file: %s\nCommand: '%s'\nStdout: '%s'\nStderr: '%s'"
% (e.message, e.cmd, e.stdout, e.stderr))
# Create combined and sorted files.
try:
ujsClient.update_job_progress(
job['id'], self.context['token'],
'merging all pairs of sequence files', 1, timestamp(3600))
except:
pass
resultList = []
for p, q in combinations(sortedList, 2):
pbase = os.path.basename(p)
qbase = os.path.basename(q)
dbase = '%s%s%s.sorted' % (os.path.splitext(pbase)[0],
PairSeparator,
os.path.splitext(qbase)[0])
destFile = os.path.join(self.jobDirectory, dbase)
sortedList.append(destFile)
args = ['/usr/bin/sort', '-m', '--output=%s' % (destFile), p, q]
result = self.pool.apply_async(run_command, (args, ))
resultList.append(result)
for result in resultList:
try:
result.get()
except CommandError as e:
self._cleanup()
raise MergeError(
"Error merging sequence file: %s\nCommand: '%s'\nStdout: '%s'\nStderr: '%s'"
% (e.message, e.cmd, e.stdout, e.stderr))
# Compress all sorted files.
try:
ujsClient.update_job_progress(job['id'], self.context['token'],
'compressing sequence files', 1,
timestamp(3600))
except:
pass
resultList = []
compressedList = []
for sourceFile in sortedList:
compressedList.append(sourceFile + '.xz')
if self.input['extreme']:
level = '-9e'
else:
level = '-9'
args = ['/usr/bin/xz', '--keep', level, '--no-warn', sourceFile]
result = self.pool.apply_async(run_command, (args, ))
resultList.append(result)
for result in resultList:
try:
result.get()
except CommandError as e:
self._cleanup()
raise CompressError(
"Error compressing sequence file: %s\nCommand: '%s'\nStdout: '%s'\nStderr: '%s'"
% (e.message, e.cmd, e.stdout, e.stderr))
# Calculate the distance matrix.
try:
ujsClient.update_job_progress(job['id'], self.context['token'],
'calculating distance matrix', 1,
timestamp(3600))
except:
pass
csvFile = os.path.join(self.jobDirectory, '%s.csv' % (job['id']))
self._cbdCalculator(compressedList, self.input['scale'], csvFile)
# Store the output file in shock.
try:
ujsClient.update_job_progress(job['id'], self.context['token'],
'storing output file in shock', 1,
timestamp(3600))
except:
pass
node = self.shockClient.create_node(csvFile, '')
if not node['id']:
# Shock let us down. Save the distance matrix in the work directory for possible recovery.
os.rename(
csvFile,
'%s/%s.csv' % (self.config['work_folder_path'], job['id']))
self._cleanup()
raise ShockError(
"Error saving distance matrix file to Shock. A Shock node was not created."
)
# Mark the job as complete.
results = {
'shocknodes': [node['id']],
'shockurl': self.config['shock_url']
}
ujsClient.complete_job(job['id'], self.context['token'], 'done', None,
results)
self._log(log.INFO, 'Job ' + job['id'] + ' completed successfully')
# Cleanup after ourselves.
self._cleanup()
return
class CompressionBasedDistance:
''' Calculate the compression based distance metric and save distance matrix to a file.
@param fileList List of paths to compressed files
@param scale Scale of distance values, 'std' for 0 to 1, 'inf' for 0 to infinity
@param outputFile Path to file with output distance matrix
@return Nothing
'''
def _cbdCalculator(self, fileList, scale, outputFile):
# Parse the files.
single_sizes = dict()
pair_sizes = dict()
for sourceFile in fileList:
# Should strip prefix too
fbase = os.path.basename(sourceFile)
# This works as long as '.sorted.xz' only occurs at the end of the path.
fname = fbase.replace('.sorted.xz', '')
if PairSeparator in fname:
pair_sizes[fname] = os.path.getsize(sourceFile)
else:
single_sizes[fname] = os.path.getsize(sourceFile)
# Map file names to indices.
fnames = single_sizes.keys()
fnames.sort()
indices = dict()
for name,i in zip(fnames, range(len(fnames))):
indices[name] = i
# Compute the distance scores.
pair_names = pair_sizes.keys()
cbd_array = numpy.zeros((len(fnames), len(fnames)), dtype=float)
for pair in pair_names:
name1, name2 = pair.split(PairSeparator)
c1 = float(single_sizes[name1])
c2 = float(single_sizes[name2])
c12 = float(pair_sizes[pair])
distance = 1.0 - ( 2.0 * ( (c1 + c2 - c12) / (c1 + c2) ) )
if distance > 1.0:
part1 = "Distance %f is greater than 1.0. " %(distance)
part2 = "Check sequence read lengths and relative number of sequence reads. "
part3 = "(c1=%f %s, c2=%f %s c12=%f %s)" %(c1, name1, c2, name2, c12, pair)
raise ValueError(part1+part2+part3)
if scale == 'inf':
distance = distance/(1.0 - distance)
cbd_array[indices[name1],indices[name2]] = distance
cbd_array[indices[name2],indices[name1]] = distance
# Build the output file in CSV format.
outf = open(outputFile, 'w')
outf.write('ID,' + ','.join(fnames) + '\n')
for i in range(len(fnames)):
outf.write(fnames[i] + ',' + ','.join(['{0:g}'.format(x) for x in cbd_array[i,:]]) + '\n')
outf.close()
return
''' Cleanup after running a job.
@note All temporary files are removed even when there is an error.
@return Nothing
'''
def _cleanup(self):
# Delete input fasta files from Shock.
for nodeId in self.input['node_ids']:
try:
self.shockClient.delete_node(nodeId)
except Exception as e:
self._log(log.ERR, 'Error deleting node %s from Shock: %s' %(nodeId, e.message))
# Remove the work directory.
shutil.rmtree(self.jobDirectory)
# Stop the process pool.
self.pool.close()
self.pool.join()
return
''' Log a message to the system log.
@param level Message level (INFO, WARNING, etc.)
@param message Message text
@return Nothing
'''
def _log(self, level, message):
# Create a logger if this is the first time the method has been called.
if self.logger is None:
submod = os.environ.get('KB_SERVICE_NAME', 'CompressionBasedDistance')
self.logger = log.log(submod, ip_address=True, authuser=True, module=True, method=True,
call_id=True, config=os.getenv('KB_DEPLOYMENT_CONFIG'))
# Log the message.
self.logger.log_message(level, message, self.context['client_ip'], self.context['user_id'], self.context['module'],
self.context['method'], self.context['call_id'])
return
def __init__(self):
self.logger = None
''' Run a job to build a distance matrix.
When successful the distance matrix csv file is stored in Shock.
@param job Dictionary with configuration variables, context variables, and input variables for job
@raise ExtractError: Error extracting sequences from input sequence file
@raise SeqLenError: Error with lengths of sequences in input sequence file
@raise SortError: Error sorting a raw sequence file
@raise MergeError: Error merging a raw sequence file
@raise CompressError: Error compressing a raw sequence file
@raise ShockError: Error saving file to Shock
@return Nothing
'''
def runJob(self, job):
self.config = job['config']
self.context = job['context']
self.input = job['input']
# Create a shock client and authenticate as the user.
self.shockClient = ShockClient(self.config['shock_url'], self.context['token'])
# Create a user and job state client and authenticate as the user.
ujsClient = UserAndJobState(self.config['userandjobstate_url'], token=self.context['token'])
# Create a process pool.
self.pool = Pool(processes=int(self.config['num_pool_processes']))
# Create a work directory for storing intermediate files.
self.jobDirectory = make_job_dir(self.config['work_folder_path'], job['id'])
self._log(log.INFO, 'Job '+job['id']+' running with work folder '+self.jobDirectory)
# Download input fasta files from Shock and extract sequences to work directory.
try:
ujsClient.update_job_progress(job['id'], self.context['token'], 'extracting sequence files', 1, timestamp(3600))
except:
pass
resultList = []
sequenceList = []
for nodeId in self.input['node_ids']:
node = self.shockClient.get_node(nodeId)
sourceFile = os.path.join(self.jobDirectory, node['file']['name'])
destFile = '%s.sequence' %(os.path.splitext(sourceFile)[0])
if PairSeparator in destFile: # Check for pair separator string in file name and replace as needed.
destFile = destFile.replace(PairSeparator, '-')
sequenceList.append(destFile)
args = dict() # Needs to be scoped here so each process gets its own copy
args['format'] = self.input['format']
args['shockUrl'] = self.config['shock_url']
args['auth'] = self.context['token']
args['sequenceLen'] = self.input['sequence_length']
args['minReads'] = self.input['min_reads']
args['maxReads'] = self.input['max_reads']
args['nodeId'] = nodeId
args['sourceFile'] = sourceFile
args['destFile'] = destFile
result = self.pool.apply_async(extract_seq, (args,))
resultList.append(result)
for result in resultList:
if result.get() != 0:
self._cleanup()
raise ExtractError("Error extracting sequences from input sequence file, result: %d" %(result.get()))
for path in self.input['file_paths']:
sourceFile = os.path.basename(path)
destFile = '%s/%s.sequence' %(self.jobDirectory, os.path.splitext(sourceFile)[0])
if PairSeparator in destFile: # Check for pair separator string in file name and replace as needed.
destFile = destFile.replace(PairSeparator, '-')
sequenceList.append(destFile)
args = dict() # Needs to be scoped here so each process gets its own copy
args['format'] = self.input['format']
args['shockUrl'] = self.config['shock_url']
args['auth'] = self.context['token']
args['sequenceLen'] = self.input['sequence_length']
args['minReads'] = self.input['min_reads']
args['maxReads'] = self.input['max_reads']
args['nodeId'] = None
args['sourceFile'] = path
args['destFile'] = destFile
result = self.pool.apply_async(extract_seq, (args,))
resultList.append(result)
for result in resultList:
try:
result.get()
except Exception as e:
self._cleanup()
raise ExtractError("Error extracting sequences from input sequence file: %s" %(e.message))
# Confirm that each file met the criteria for sequence length and number of sequences.
filesToRemove = list()
for index in range(len(sequenceList)):
# See if the file did not have the minimum number of sequences.
if not os.path.exists(sequenceList[index]):
filesToRemove.append(index)
continue
# See if the file has no data.
if os.path.getsize(sequenceList[index]) == 0:
self._cleanup()
raise SeqLenError("Sequence file '%s' has no sequences" %(sequenceList[index]))
filteredList = list()
for index in range(len(sequenceList)):
if index not in filesToRemove:
filteredList.append(sequenceList[index])
if len(filteredList) < 2:
self._cleanup()
raise SeqLenError("There are not enough sequence files that meet the sequence length or number of sequences criteria.")
# Sort the sequences.
try:
ujsClient.update_job_progress(job['id'], self.context['token'], 'sorting sequence files', 1, timestamp(3600))
except:
pass
resultList = []
sortedList = []
for sourceFile in filteredList:
destFile = '%s.sorted' %(os.path.splitext(sourceFile)[0])
sortedList.append(destFile)
args = [ '/usr/bin/sort', '--output=%s' %(destFile), sourceFile ]
result = self.pool.apply_async(run_command, (args,))
resultList.append(result)
for result in resultList:
try:
result.get()
except CommandError as e:
self._cleanup()
raise SortError("Error sorting sequence file: %s\nCommand: '%s'\nStdout: '%s'\nStderr: '%s'" %(e.message, e.cmd, e.stdout, e.stderr))
# Create combined and sorted files.
try:
ujsClient.update_job_progress(job['id'], self.context['token'], 'merging all pairs of sequence files', 1, timestamp(3600))
except:
pass
resultList = []
for p,q in combinations(sortedList, 2):
pbase = os.path.basename(p)
qbase = os.path.basename(q)
dbase = '%s%s%s.sorted' %(os.path.splitext(pbase)[0], PairSeparator, os.path.splitext(qbase)[0])
destFile = os.path.join(self.jobDirectory, dbase)
sortedList.append(destFile)
args = [ '/usr/bin/sort', '-m', '--output=%s' %(destFile), p, q ]
result = self.pool.apply_async(run_command, (args,))
resultList.append(result)
for result in resultList:
try:
result.get()
except CommandError as e:
self._cleanup()
raise MergeError("Error merging sequence file: %s\nCommand: '%s'\nStdout: '%s'\nStderr: '%s'" %(e.message, e.cmd, e.stdout, e.stderr))
# Compress all sorted files.
try:
ujsClient.update_job_progress(job['id'], self.context['token'], 'compressing sequence files', 1, timestamp(3600))
except:
pass
resultList = []
compressedList = []
for sourceFile in sortedList:
compressedList.append(sourceFile+'.xz')
if self.input['extreme']:
level = '-9e'
else:
level = '-9'
args = [ '/usr/bin/xz', '--keep', level, '--no-warn', sourceFile ]
result = self.pool.apply_async(run_command, (args,))
resultList.append(result)
for result in resultList:
try:
result.get()
except CommandError as e:
self._cleanup()
raise CompressError("Error compressing sequence file: %s\nCommand: '%s'\nStdout: '%s'\nStderr: '%s'" %(e.message, e.cmd, e.stdout, e.stderr))
# Calculate the distance matrix.
try:
ujsClient.update_job_progress(job['id'], self.context['token'], 'calculating distance matrix', 1, timestamp(3600))
except:
pass
csvFile = os.path.join(self.jobDirectory, '%s.csv' %(job['id']))
self._cbdCalculator(compressedList, self.input['scale'], csvFile)
# Store the output file in shock.
try:
ujsClient.update_job_progress(job['id'], self.context['token'], 'storing output file in shock', 1, timestamp(3600))
except:
pass
node = self.shockClient.create_node(csvFile, '')
if not node['id']:
# Shock let us down. Save the distance matrix in the work directory for possible recovery.
os.rename(csvFile, '%s/%s.csv' %(self.config['work_folder_path'], job['id']))
self._cleanup()
raise ShockError("Error saving distance matrix file to Shock. A Shock node was not created.")
# Mark the job as complete.
results = { 'shocknodes': [ node['id'] ], 'shockurl': self.config['shock_url'] }
ujsClient.complete_job(job['id'], self.context['token'], 'done', None, results)
self._log(log.INFO, 'Job '+job['id']+' completed successfully')
# Cleanup after ourselves.
self._cleanup()
return
def calculate(self, listFilePath, scale, csvFile):
# Each line of the list file is a path to a compressed file.
compressedList = list()
listFile = open(listFilePath, 'r')
for line in listFile:
compressedList.append(line.strip())
listFile.close()
# Calculate the distance matrix.
self._cbdCalculator(compressedList, scale, csvFile)
return
def runJob(self, job):
self.config = job['config']
self.context = job['context']
self.input = job['input']
# Create a shock client and authenticate as the user.
self.shockClient = ShockClient(self.config['shock_url'], self.context['token'])
# Create a user and job state client and authenticate as the user.
ujsClient = UserAndJobState(self.config['userandjobstate_url'], token=self.context['token'])
# Create a process pool.
self.pool = Pool(processes=int(self.config['num_pool_processes']))
# Create a work directory for storing intermediate files.
self.jobDirectory = make_job_dir(self.config['work_folder_path'], job['id'])
self._log(log.INFO, 'Job '+job['id']+' running with work folder '+self.jobDirectory)
# Download input fasta files from Shock and extract sequences to work directory.
try:
ujsClient.update_job_progress(job['id'], self.context['token'], 'extracting sequence files', 1, timestamp(3600))
except:
pass
resultList = []
sequenceList = []
for nodeId in self.input['node_ids']:
node = self.shockClient.get_node(nodeId)
sourceFile = os.path.join(self.jobDirectory, node['file']['name'])
destFile = '%s.sequence' %(os.path.splitext(sourceFile)[0])
if PairSeparator in destFile: # Check for pair separator string in file name and replace as needed.
destFile = destFile.replace(PairSeparator, '-')
sequenceList.append(destFile)
args = dict() # Needs to be scoped here so each process gets its own copy
args['format'] = self.input['format']
args['shockUrl'] = self.config['shock_url']
args['auth'] = self.context['token']
args['sequenceLen'] = self.input['sequence_length']
args['minReads'] = self.input['min_reads']
args['maxReads'] = self.input['max_reads']
args['nodeId'] = nodeId
args['sourceFile'] = sourceFile
args['destFile'] = destFile
result = self.pool.apply_async(extract_seq, (args,))
resultList.append(result)
for result in resultList:
if result.get() != 0:
self._cleanup()
raise ExtractError("Error extracting sequences from input sequence file, result: %d" %(result.get()))
for path in self.input['file_paths']:
sourceFile = os.path.basename(path)
destFile = '%s/%s.sequence' %(self.jobDirectory, os.path.splitext(sourceFile)[0])
if PairSeparator in destFile: # Check for pair separator string in file name and replace as needed.
destFile = destFile.replace(PairSeparator, '-')
sequenceList.append(destFile)
args = dict() # Needs to be scoped here so each process gets its own copy
args['format'] = self.input['format']
args['shockUrl'] = self.config['shock_url']
args['auth'] = self.context['token']
args['sequenceLen'] = self.input['sequence_length']
args['minReads'] = self.input['min_reads']
args['maxReads'] = self.input['max_reads']
args['nodeId'] = None
args['sourceFile'] = path
args['destFile'] = destFile
result = self.pool.apply_async(extract_seq, (args,))
resultList.append(result)
for result in resultList:
try:
result.get()
except Exception as e:
self._cleanup()
raise ExtractError("Error extracting sequences from input sequence file: %s" %(e.message))
# Confirm that each file met the criteria for sequence length and number of sequences.
filesToRemove = list()
for index in range(len(sequenceList)):
# See if the file did not have the minimum number of sequences.
if not os.path.exists(sequenceList[index]):
filesToRemove.append(index)
continue
# See if the file has no data.
if os.path.getsize(sequenceList[index]) == 0:
self._cleanup()
raise SeqLenError("Sequence file '%s' has no sequences" %(sequenceList[index]))
filteredList = list()
for index in range(len(sequenceList)):
if index not in filesToRemove:
filteredList.append(sequenceList[index])
if len(filteredList) < 2:
self._cleanup()
raise SeqLenError("There are not enough sequence files that meet the sequence length or number of sequences criteria.")
# Sort the sequences.
try:
ujsClient.update_job_progress(job['id'], self.context['token'], 'sorting sequence files', 1, timestamp(3600))
except:
pass
resultList = []
sortedList = []
for sourceFile in filteredList:
destFile = '%s.sorted' %(os.path.splitext(sourceFile)[0])
sortedList.append(destFile)
args = [ '/usr/bin/sort', '--output=%s' %(destFile), sourceFile ]
result = self.pool.apply_async(run_command, (args,))
resultList.append(result)
for result in resultList:
try:
result.get()
except CommandError as e:
self._cleanup()
raise SortError("Error sorting sequence file: %s\nCommand: '%s'\nStdout: '%s'\nStderr: '%s'" %(e.message, e.cmd, e.stdout, e.stderr))
# Create combined and sorted files.
try:
ujsClient.update_job_progress(job['id'], self.context['token'], 'merging all pairs of sequence files', 1, timestamp(3600))
except:
pass
resultList = []
for p,q in combinations(sortedList, 2):
pbase = os.path.basename(p)
qbase = os.path.basename(q)
dbase = '%s%s%s.sorted' %(os.path.splitext(pbase)[0], PairSeparator, os.path.splitext(qbase)[0])
destFile = os.path.join(self.jobDirectory, dbase)
sortedList.append(destFile)
args = [ '/usr/bin/sort', '-m', '--output=%s' %(destFile), p, q ]
result = self.pool.apply_async(run_command, (args,))
resultList.append(result)
for result in resultList:
try:
result.get()
except CommandError as e:
self._cleanup()
raise MergeError("Error merging sequence file: %s\nCommand: '%s'\nStdout: '%s'\nStderr: '%s'" %(e.message, e.cmd, e.stdout, e.stderr))
# Compress all sorted files.
try:
ujsClient.update_job_progress(job['id'], self.context['token'], 'compressing sequence files', 1, timestamp(3600))
except:
pass
resultList = []
compressedList = []
for sourceFile in sortedList:
compressedList.append(sourceFile+'.xz')
if self.input['extreme']:
level = '-9e'
else:
level = '-9'
args = [ '/usr/bin/xz', '--keep', level, '--no-warn', sourceFile ]
result = self.pool.apply_async(run_command, (args,))
resultList.append(result)
for result in resultList:
try:
result.get()
except CommandError as e:
self._cleanup()
raise CompressError("Error compressing sequence file: %s\nCommand: '%s'\nStdout: '%s'\nStderr: '%s'" %(e.message, e.cmd, e.stdout, e.stderr))
# Calculate the distance matrix.
try:
ujsClient.update_job_progress(job['id'], self.context['token'], 'calculating distance matrix', 1, timestamp(3600))
except:
pass
csvFile = os.path.join(self.jobDirectory, '%s.csv' %(job['id']))
self._cbdCalculator(compressedList, self.input['scale'], csvFile)
# Store the output file in shock.
try:
ujsClient.update_job_progress(job['id'], self.context['token'], 'storing output file in shock', 1, timestamp(3600))
except:
pass
node = self.shockClient.create_node(csvFile, '')
if not node['id']:
# Shock let us down. Save the distance matrix in the work directory for possible recovery.
os.rename(csvFile, '%s/%s.csv' %(self.config['work_folder_path'], job['id']))
self._cleanup()
raise ShockError("Error saving distance matrix file to Shock. A Shock node was not created.")
# Mark the job as complete.
results = { 'shocknodes': [ node['id'] ], 'shockurl': self.config['shock_url'] }
ujsClient.complete_job(job['id'], self.context['token'], 'done', None, results)
self._log(log.INFO, 'Job '+job['id']+' completed successfully')
# Cleanup after ourselves.
self._cleanup()
return
ujsClient.delete_job(args.jobID)
exit(1)
# Check if the job is complete.
if not info['complete']:
print "Job '%s' has status '%s' and is working on task %s of %s. Check again later." \
%(args.jobID, info['status'], info['total_progress'], info['max_progress'])
exit(1)
# Show job info.
if args.showTimes:
print 'Job started at %s and finished at %s' % (info['started'],
info['last_update'])
# Create a shock client.
shockClient = ShockClient(info['results']['shockurl'],
ujsClient._headers['AUTHORIZATION'])
# Download the output to the specified file and remove the file from shock.
try:
shockClient.download_to_path(info['results']['shocknodes'][0],
args.outputPath)
except Exception as e:
print 'Error downloading distance matrix from %s: %s' % (
info['results']['shockurl'], e.message)
traceback.print_exc(file=sys.stdout)
try:
shockClient.delete_node(info['results']['shocknodes'][0])
except Exception as e:
print 'Error deleting distance matrix file from %s: ' % (
+info['results']['shockurl'], e.message)
traceback.print_exc(file=sys.stdout)
input['sequence_length'] = args.sequenceLen
input['min_reads'] = args.minReads
input['max_reads'] = args.maxReads
if args.extreme:
input['extreme'] = 1
else:
input['extreme'] = 0
input['node_ids'] = list()
# Create a cbd client (which must be authenticated).
if args.url is None:
args.url = get_url()
cbdClient = CompressionBasedDistance(url=args.url)
# Create a shock client.
shockClient = ShockClient(args.shockurl, cbdClient._headers['AUTHORIZATION'])
# Parse the input file with the list of sequence files.
(fileList, extensions, numMissingFiles) = parse_input_file(args.inputPath)
if numMissingFiles > 0:
exit(1)
# Set the format based on the sequence file extension if the format argument was not specified.
if args.format is None:
if len(extensions) == 1:
input['format'] = extensions.keys()[0]
else:
print "The format of the sequence files could not be determined. Set the format with the --format argument."
exit(1)
else:
input['format'] = args.format
print ujsClient.get_detailed_error(args.jobID)
ujsClient.delete_job(args.jobID)
exit(1)
# Check if the job is complete.
if not info['complete']:
print "Job '%s' has status '%s' and is working on task %s of %s. Check again later." \
%(args.jobID, info['status'], info['total_progress'], info['max_progress'])
exit(1)
# Show job info.
if args.showTimes:
print 'Job started at %s and finished at %s' %(info['started'], info['last_update'])
# Create a shock client.
shockClient = ShockClient(info['results']['shockurl'], ujsClient._headers['AUTHORIZATION'])
# Download the output to the specified file and remove the file from shock.
try:
shockClient.download_to_path(info['results']['shocknodes'][0], args.outputPath)
except Exception as e:
print 'Error downloading distance matrix from %s: %s' %(info['results']['shockurl'], e.message)
traceback.print_exc(file=sys.stdout)
try:
shockClient.delete_node(info['results']['shocknodes'][0])
except Exception as e:
print 'Error deleting distance matrix file from %s: ' %(+info['results']['shockurl'], e.message)
traceback.print_exc(file=sys.stdout)
# Delete the job.
ujsClient.delete_job(args.jobID)
