def test_compute_seqs_per_file(self):
"""compute_seqs_per_file functions as expected
"""
temp_fasta_fp = get_tmp_filename(\
prefix='QiimeScriptUtilTests',suffix='.fasta')
temp_fasta = ['>seq', 'AAACCCCAAATTGG'] * 25
open(temp_fasta_fp, 'w').write('\n'.join(temp_fasta))
actual_25 = compute_seqs_per_file(temp_fasta_fp, 25)
actual_2 = compute_seqs_per_file(temp_fasta_fp, 2)
actual_10 = compute_seqs_per_file(temp_fasta_fp, 10)
actual_5 = compute_seqs_per_file(temp_fasta_fp, 5)
actual_40 = compute_seqs_per_file(temp_fasta_fp, 40)
remove(temp_fasta_fp)
self.assertEqual(actual_25, 1)
self.assertEqual(actual_2, 13)
self.assertEqual(actual_10, 3)
self.assertEqual(actual_5, 5)
self.assertEqual(actual_40, 1)
def test_compute_seqs_per_file(self):
"""compute_seqs_per_file functions as expected
"""
temp_fasta_fp = get_tmp_filename(\
prefix='QiimeScriptUtilTests',suffix='.fasta')
temp_fasta = ['>seq','AAACCCCAAATTGG'] * 25
open(temp_fasta_fp,'w').write('\n'.join(temp_fasta))
actual_25 = compute_seqs_per_file(temp_fasta_fp,25)
actual_2 = compute_seqs_per_file(temp_fasta_fp,2)
actual_10 = compute_seqs_per_file(temp_fasta_fp,10)
actual_5 = compute_seqs_per_file(temp_fasta_fp,5)
actual_40 = compute_seqs_per_file(temp_fasta_fp,40)
remove(temp_fasta_fp)
self.assertEqual(actual_25,1)
self.assertEqual(actual_2,13)
self.assertEqual(actual_10,3)
self.assertEqual(actual_5,5)
self.assertEqual(actual_40,1)
def _split_fasta(self,
input_fp,
jobs_to_start,
job_prefix,
output_dir):
# compute the number of sequences that should be included in
# each file after splitting the input fasta file
num_seqs_per_file = compute_seqs_per_file(input_fp,jobs_to_start)
# split the fasta files and get the list of resulting files
tmp_fasta_fps =\
split_fasta(open(input_fp),num_seqs_per_file,\
job_prefix,working_dir=output_dir)
return tmp_fasta_fps, True
def main():
option_parser, opts, args = parse_command_line_parameters(**script_info)
if opts.blast_db == None and opts.refseqs_fp == None:
option_parser.error('Either blast_db or refseqs_fp must be provided.')
# create local copies of command-line options
python_exe_fp = opts.python_exe_fp
pick_otus_fp = opts.pick_otus_fp
refseqs_fp = opts.refseqs_fp
cluster_jobs_fp = opts.cluster_jobs_fp
input_fasta_fp = opts.input_fasta_fp
jobs_to_start = opts.jobs_to_start
output_dir = opts.output_dir
poller_fp = opts.poller_fp
retain_temp_files = opts.retain_temp_files
suppress_polling = opts.suppress_polling
seconds_to_sleep = opts.seconds_to_sleep
max_e_value = opts.max_e_value
similarity = opts.similarity
poll_directly = opts.poll_directly
min_aligned_percent = opts.min_aligned_percent
created_temp_paths = []
if not opts.blast_db:
# Build the blast database from the reference_seqs_fp -- all procs
# will then access one db rather than create one per proc
blast_db, db_files_to_remove = \
build_blast_db_from_fasta_path(refseqs_fp)
created_temp_paths += db_files_to_remove
else:
blast_db = opts.blast_db
# split the input filepath into directory and filename, base filename and
# extension
input_dir, input_fasta_fn = split(input_fasta_fp)
input_file_basename, input_fasta_ext = splitext(input_fasta_fn)
# set the job_prefix either based on what the user passed in,
# or a random string beginning with RDP
job_prefix = opts.job_prefix or get_random_job_prefix('POTU')
# A temporary output directory is created in output_dir named
# job_prefix. Output files are then moved from the temporary
# directory to the output directory when they are complete, allowing
# a poller to detect when runs complete by the presence of their
# output files.
working_dir = '%s/%s' % (output_dir,job_prefix)
try:
makedirs(working_dir)
created_temp_paths.append(working_dir)
except OSError:
# working dir already exists
pass
# compute the number of sequences that should be included in
# each file after splitting the input fasta file
num_seqs_per_file = compute_seqs_per_file(input_fasta_fp,jobs_to_start)
# split the fasta files and get the list of resulting files
tmp_fasta_fps =\
split_fasta(open(input_fasta_fp),num_seqs_per_file,\
job_prefix,working_dir=output_dir)
created_temp_paths += tmp_fasta_fps
# build the filepath for the 'jobs script'
jobs_fp = '%s/%sjobs.txt' % (output_dir, job_prefix)
created_temp_paths.append(jobs_fp)
# generate the list of commands to be pushed out to nodes and the list of
# output files generated by each job
commands, job_result_filepaths = \
get_job_commands(python_exe_fp,pick_otus_fp,tmp_fasta_fps,
output_dir,blast_db,job_prefix,working_dir,max_e_value,similarity,
min_aligned_percent)
created_temp_paths += job_result_filepaths
# Set up poller apparatus if the user does not suppress polling
if not suppress_polling:
# Write the list of files which must exist for the jobs to be
# considered complete
expected_files_filepath = '%s/expected_out_files.txt' % working_dir
write_filepaths_to_file(job_result_filepaths,expected_files_filepath)
created_temp_paths.append(expected_files_filepath)
# Write the mapping file which described how the output files from
# each job should be merged into the final output files
merge_map_filepath = '%s/merge_map.txt' % working_dir
process_run_results_f =\
'qiime.parallel.pick_otus_blast.parallel_blast_process_run_results_f'
write_merge_map_file_pick_otus(job_result_filepaths,output_dir,\
merge_map_filepath,input_file_basename)
created_temp_paths.append(merge_map_filepath)
# Create the filepath listing the temporary files to be deleted,
# but don't write it yet
deletion_list_filepath = '%s/deletion_list.txt' % working_dir
created_temp_paths.append(deletion_list_filepath)
# Generate the command to run the poller, and the list of temp files
# created by the poller
if not poll_directly:
poller_command, poller_result_filepaths =\
get_poller_command(python_exe_fp,poller_fp,expected_files_filepath,\
merge_map_filepath,deletion_list_filepath,process_run_results_f,\
seconds_to_sleep=seconds_to_sleep)
created_temp_paths += poller_result_filepaths
# append the poller command to the list of job commands
commands.append(poller_command)
else:
poller_command, poller_result_filepaths =\
get_poller_command(python_exe_fp,poller_fp,expected_files_filepath,\
merge_map_filepath,deletion_list_filepath,process_run_results_f,\
seconds_to_sleep=seconds_to_sleep,command_prefix='',command_suffix='')
created_temp_paths += poller_result_filepaths
if not retain_temp_files:
# If the user wants temp files deleted, now write the list of
# temp files to be deleted
write_filepaths_to_file(created_temp_paths,deletion_list_filepath)
else:
# Otherwise just write an empty file
write_filepaths_to_file([],deletion_list_filepath)
# write the commands to the 'jobs files'
write_jobs_file(commands,job_prefix=job_prefix,jobs_fp=jobs_fp)
# submit the jobs file using cluster_jobs, if not suppressed by the
# user
if not opts.suppress_submit_jobs:
submit_jobs(cluster_jobs_fp,jobs_fp,job_prefix)
if poll_directly:
try:
check_call(poller_command.split())
except CalledProcessError, e:
print '**Error occuring when calling the poller directly. '+\
'Jobs may have been submitted, but are not being polled.'
print str(e)
exit(-1)
def main():
option_parser, opts, args = parse_command_line_parameters(**script_info)
# create local copies of command-line options
python_exe_fp = opts.python_exe_fp
assign_taxonomy_fp = opts.assign_taxonomy_fp
confidence = opts.confidence
rdp_classifier_fp = opts.rdp_classifier_fp
id_to_taxonomy_fp = opts.id_to_taxonomy_fp
reference_seqs_fp = opts.reference_seqs_fp
cluster_jobs_fp = opts.cluster_jobs_fp
input_fasta_fp = opts.input_fasta_fp
jobs_to_start = opts.jobs_to_start
output_dir = opts.output_dir
poller_fp = opts.poller_fp
retain_temp_files = opts.retain_temp_files
suppress_polling = opts.suppress_polling
seconds_to_sleep = opts.seconds_to_sleep
poll_directly = opts.poll_directly
if not isfile(input_fasta_fp):
raise ValueError('This file does not exists: %s' % input_fasta_fp)
if id_to_taxonomy_fp or reference_seqs_fp:
if not id_to_taxonomy_fp or not isfile(id_to_taxonomy_fp):
raise ValueError('This file does not exits: %s' % id_to_taxonomy_fp)
if not reference_seqs_fp or not isfile(reference_seqs_fp):
raise ValueError('This file does not exits: %s' % reference_seqs_fp)
try:
makedirs(output_dir)
except OSError:
# output dir already exists
pass
created_temp_paths = []
# split the input filepath into directory and filename, base filename and
# extension
input_dir, input_fasta_fn = split(input_fasta_fp)
input_file_basename, input_fasta_ext = splitext(input_fasta_fn)
# set the job_prefix either based on what the user passed in,
# or a random string beginning with RDP
job_prefix = opts.job_prefix or get_random_job_prefix('RDP')
# A temporary output directory is created in output_dir named
# job_prefix. Output files are then moved from the temporary
# directory to the output directory when they are complete, allowing
# a poller to detect when runs complete by the presence of their
# output files.
working_dir = '%s/%s' % (output_dir,job_prefix)
try:
mkdir(working_dir)
created_temp_paths.append(working_dir)
except OSError:
# working dir already exists
pass
# compute the number of sequences that should be included in
# each file after splitting the input fasta file
num_seqs_per_file = compute_seqs_per_file(input_fasta_fp,jobs_to_start)
# split the fasta files and get the list of resulting files
tmp_fasta_fps =\
split_fasta(open(input_fasta_fp),num_seqs_per_file,job_prefix,output_dir)
created_temp_paths += tmp_fasta_fps
# build the filepath for the 'jobs script'
jobs_fp = '%s/%sjobs.txt' % (output_dir, job_prefix)
created_temp_paths.append(jobs_fp)
# generate the list of commands to be pushed out to nodes
commands, job_result_filepaths = \
get_commands(python_exe_fp,assign_taxonomy_fp,confidence,job_prefix,\
tmp_fasta_fps,rdp_classifier_fp,output_dir,working_dir,\
id_to_taxonomy_fp=id_to_taxonomy_fp,reference_seqs_fp=reference_seqs_fp)
created_temp_paths += job_result_filepaths
# Set up poller apparatus if the user does not suppress polling
if not suppress_polling:
# Write the list of files which must exist for the jobs to be
# considered complete
expected_files_filepath = '%s/expected_out_files.txt' % working_dir
write_filepaths_to_file(job_result_filepaths,expected_files_filepath)
created_temp_paths.append(expected_files_filepath)
# Write the mapping file which described how the output files from
# each job should be merged into the final output files
merge_map_filepath = '%s/merge_map.txt' % working_dir
write_merge_map_file_assign_taxonomy(job_result_filepaths,output_dir,\
merge_map_filepath,input_file_basename)
created_temp_paths.append(merge_map_filepath)
# Create the filepath listing the temporary files to be deleted,
# but don't write it yet
deletion_list_filepath = '%s/deletion_list.txt' % working_dir
created_temp_paths.append(deletion_list_filepath)
# Generate the command to run the poller, and the list of temp files
# created by the poller
if not poll_directly:
poller_command, poller_result_filepaths =\
get_poller_command(python_exe_fp,poller_fp,expected_files_filepath,\
merge_map_filepath,deletion_list_filepath,\
seconds_to_sleep=seconds_to_sleep)
created_temp_paths += poller_result_filepaths
# append the poller command to the list of job commands
commands.append(poller_command)
else:
poller_command, poller_result_filepaths =\
get_poller_command(python_exe_fp,poller_fp,\
expected_files_filepath,merge_map_filepath,\
deletion_list_filepath,seconds_to_sleep=seconds_to_sleep,\
command_prefix='',command_suffix='')
created_temp_paths += poller_result_filepaths
if not retain_temp_files:
# If the user wants temp files deleted, now write the list of
# temp files to be deleted
write_filepaths_to_file(created_temp_paths,deletion_list_filepath)
else:
# Otherwise just write an empty file
write_filepaths_to_file([],deletion_list_filepath)
# write the commands to the 'jobs files'
write_jobs_file(commands,job_prefix=job_prefix,jobs_fp=jobs_fp)
# submit the jobs file using cluster_jobs, if not suppressed by the
# user
if not opts.suppress_submit_jobs:
submit_jobs(cluster_jobs_fp,jobs_fp,job_prefix)
if poll_directly:
try:
check_call(poller_command.split())
except CalledProcessError, e:
print '**Error occuring when calling the poller directly. '+\
'Jobs may have been submitted, but are not being polled.'
print str(e)
exit(-1)
def main():
option_parser, opts, args = parse_command_line_parameters(**script_info)
# create local copies of command-line options
python_exe_fp = opts.python_exe_fp
pick_otus_fp = opts.pick_otus_fp
refseqs_fp = opts.refseqs_fp
cluster_jobs_fp = opts.cluster_jobs_fp
input_fasta_fp = opts.input_fasta_fp
jobs_to_start = opts.jobs_to_start
output_dir = opts.output_dir
poller_fp = opts.poller_fp
retain_temp_files = opts.retain_temp_files
suppress_polling = opts.suppress_polling
seconds_to_sleep = opts.seconds_to_sleep
similarity = opts.similarity
poll_directly = opts.poll_directly
uclust_stable_sort = not opts.suppress_uclust_stable_sort
save_uc_files = opts.save_uc_files
enable_rev_strand_match = opts.enable_rev_strand_match
optimal_uclust = opts.optimal_uclust
exact_uclust = opts.exact_uclust
max_accepts = opts.max_accepts
max_rejects = opts.max_rejects
stepwords = opts.stepwords
word_length = opts.word_length
created_temp_paths = []
# split the input filepath into directory and filename, base filename and
# extension
input_dir, input_fasta_fn = split(input_fasta_fp)
input_file_basename, input_fasta_ext = splitext(input_fasta_fn)
# set the job_prefix either based on what the user passed in,
# or a random string beginning with RDP
job_prefix = opts.job_prefix or get_random_job_prefix('POTU')
# A temporary output directory is created in output_dir named
# job_prefix. Output files are then moved from the temporary
# directory to the output directory when they are complete, allowing
# a poller to detect when runs complete by the presence of their
# output files.
working_dir = '%s/%s' % (output_dir,job_prefix)
try:
makedirs(working_dir)
created_temp_paths.append(working_dir)
except OSError:
# working dir already exists
pass
# compute the number of sequences that should be included in
# each file after splitting the input fasta file
num_seqs_per_file = compute_seqs_per_file(input_fasta_fp,jobs_to_start)
# split the fasta files and get the list of resulting files
tmp_fasta_fps =\
split_fasta(open(input_fasta_fp),num_seqs_per_file,\
job_prefix,working_dir=output_dir)
created_temp_paths += tmp_fasta_fps
# build the filepath for the 'jobs script'
jobs_fp = '%s/%sjobs.txt' % (output_dir, job_prefix)
created_temp_paths.append(jobs_fp)
# generate the list of commands to be pushed out to nodes and the list of
# output files generated by each job
commands, job_result_filepaths = \
get_job_commands(python_exe_fp,pick_otus_fp,tmp_fasta_fps,
output_dir,refseqs_fp,job_prefix,working_dir,similarity,
enable_rev_strand_match,optimal_uclust,exact_uclust,max_accepts,max_rejects,
stepwords, word_length, uclust_stable_sort, save_uc_files)
if save_uc_files:
# keep any .uc files that get created
created_temp_paths +=\
[fp for fp in job_result_filepaths if not fp.endswith('.uc')]
else:
created_temp_paths += [job_result_filepaths]
# Set up poller apparatus if the user does not suppress polling
if not suppress_polling:
# Write the list of files which must exist for the jobs to be
# considered complete
expected_files_filepath = '%s/expected_out_files.txt' % working_dir
write_filepaths_to_file(job_result_filepaths,expected_files_filepath)
created_temp_paths.append(expected_files_filepath)
# Write the mapping file which described how the output files from
# each job should be merged into the final output files
merge_map_filepath = '%s/merge_map.txt' % working_dir
process_run_results_f =\
'qiime.parallel.pick_otus_uclust_ref.parallel_uclust_ref_process_run_results_f'
write_merge_map_file_pick_otus(job_result_filepaths,output_dir,\
merge_map_filepath,input_file_basename,failures=True)
created_temp_paths.append(merge_map_filepath)
# Create the filepath listing the temporary files to be deleted,
# but don't write it yet
deletion_list_filepath = '%s/deletion_list.txt' % working_dir
created_temp_paths.append(deletion_list_filepath)
# Generate the command to run the poller, and the list of temp files
# created by the poller
if not poll_directly:
poller_command, poller_result_filepaths =\
get_poller_command(python_exe_fp,poller_fp,expected_files_filepath,\
merge_map_filepath,deletion_list_filepath,process_run_results_f,\
seconds_to_sleep=seconds_to_sleep)
created_temp_paths += poller_result_filepaths
# append the poller command to the list of job commands
commands.append(poller_command)
else:
poller_command, poller_result_filepaths =\
get_poller_command(python_exe_fp,poller_fp,expected_files_filepath,\
merge_map_filepath,deletion_list_filepath,process_run_results_f,\
seconds_to_sleep=seconds_to_sleep,command_prefix='',command_suffix='')
created_temp_paths += poller_result_filepaths
if not retain_temp_files:
# If the user wants temp files deleted, now write the list of
# temp files to be deleted
write_filepaths_to_file(created_temp_paths,deletion_list_filepath)
else:
# Otherwise just write an empty file
write_filepaths_to_file([],deletion_list_filepath)
# write the commands to the 'jobs files'
write_jobs_file(commands,job_prefix=job_prefix,jobs_fp=jobs_fp)
# submit the jobs file using cluster_jobs, if not suppressed by the
# user
if not opts.suppress_submit_jobs:
submit_jobs(cluster_jobs_fp,jobs_fp,job_prefix)
if poll_directly:
try:
check_call(poller_command.split())
except CalledProcessError, e:
print '**Error occuring when calling the poller directly. '+\
'Jobs may have been submitted, but are not being polled.'
print str(e)
exit(-1)
