def system_call(cmd):
"""Call cmd and return (stdout, stderr, return_value).
cmd: can be either a string containing the command to be run, or a
sequence of strings that are the tokens of the command.
This function is ported from QIIME (http://www.qiime.org), previously
named qiime_system_call. QIIME is a GPL project, but we obtained permission
from the authors of this function to port it to pyqi (and keep it under
pyqi's BSD license).
"""
proc = Popen(cmd,
universal_newlines=True,
shell=True,
stdout=PIPE,
stderr=PIPE)
# communicate pulls all stdout/stderr from the PIPEs to
# avoid blocking -- don't remove this line!
stdout, stderr = proc.communicate()
return_value = proc.returncode
if return_value != 0:
raise ComputeError("Failed to execute: %s\nstdout: %s\nstderr: %s" %
(cmd, stdout, stderr))
return stdout, stderr, return_value
def submit_EBI(preprocessed_data_id, action, send, fastq_dir_fp=None):
"""Submit a preprocessed data to EBI
Parameters
----------
preprocessed_data_id : int
The preprocesssed data id
action : %s
The action to perform with this data
send : bool
True to actually send the files
fastq_dir_fp : str, optional
The fastq filepath
"""
preprocessed_data = PreprocessedData(preprocessed_data_id)
preprocessed_data_id_str = str(preprocessed_data_id)
study = Study(preprocessed_data.study)
sample_template = SampleTemplate(study.sample_template)
prep_template = PrepTemplate(preprocessed_data.prep_template)
investigation_type = None
new_investigation_type = None
status = preprocessed_data.submitted_to_insdc_status()
if status in ('submitting', 'success'):
raise ValueError("Cannot resubmit! Current status is: %s" % status)
if send:
# If we intend actually to send the files, then change the status in
# the database
preprocessed_data.update_insdc_status('submitting')
# we need to figure out whether the investigation type is a known one
# or if we have to submit a "new_investigation_type" to EBI
current_type = prep_template.investigation_type
ena_ontology = Ontology(convert_to_id('ENA', 'ontology'))
if current_type in ena_ontology.terms:
investigation_type = current_type
elif current_type in ena_ontology.user_defined_terms:
investigation_type = 'Other'
new_investigation_type = current_type
else:
# This should never happen
raise ValueError("Unrecognized investigation type: '%s'. This term "
"is neither one of the official terms nor one of the "
"user-defined terms in the ENA ontology")
if fastq_dir_fp is not None:
# If the user specifies a FASTQ directory, use it
# Set demux_samples to None so that MetadataTemplate.to_file will put
# all samples in the template files
demux_samples = None
else:
# If the user does not specify a FASTQ directory, create one and
# re-serialize the per-sample FASTQs from the demux file
fastq_dir_fp = mkdtemp(prefix=qiita_config.working_dir)
demux = [
path for _, path, ftype in preprocessed_data.get_filepaths()
if ftype == 'preprocessed_demux'
][0]
# Keep track of which files were actually in the demux file so that we
# can write those rows to the prep and samples templates
demux_samples = set()
with open_file(demux) as demux_fh:
for samp, iterator in to_per_sample_ascii(demux_fh,
list(sample_template)):
demux_samples.add(samp)
sample_fp = join(fastq_dir_fp, "%s.fastq.gz" % samp)
with gzopen(sample_fp, 'w') as fh:
for record in iterator:
fh.write(record)
output_dir = fastq_dir_fp + '_submission'
samp_fp = join(fastq_dir_fp, 'sample_metadata.txt')
prep_fp = join(fastq_dir_fp, 'prep_metadata.txt')
sample_template.to_file(samp_fp, demux_samples)
prep_template.to_file(prep_fp, demux_samples)
# Get specific output directory and set filepaths
get_output_fp = partial(join, output_dir)
study_fp = get_output_fp('study.xml')
sample_fp = get_output_fp('sample.xml')
experiment_fp = get_output_fp('experiment.xml')
run_fp = get_output_fp('run.xml')
submission_fp = get_output_fp('submission.xml')
if not isdir(output_dir):
makedirs(output_dir)
else:
raise IOError('The output folder already exists: %s' % output_dir)
with open(samp_fp, 'U') as st, open(prep_fp, 'U') as pt:
submission = EBISubmission.from_templates_and_per_sample_fastqs(
preprocessed_data_id_str,
study.title,
study.info['study_abstract'],
investigation_type,
st,
pt,
fastq_dir_fp,
new_investigation_type=new_investigation_type,
pmids=study.pmids)
submission.write_all_xml_files(study_fp, sample_fp, experiment_fp, run_fp,
submission_fp, action)
if send:
submission.send_sequences()
study_accession, submission_accession = submission.send_xml()
if study_accession is None or submission_accession is None:
preprocessed_data.update_insdc_status('failed')
raise ComputeError("EBI Submission failed!")
else:
preprocessed_data.update_insdc_status('success', study_accession,
submission_accession)
else:
study_accession, submission_accession = None, None
return study_accession, submission_accession
def submit_EBI(artifact_id, action, send, test=False):
"""Submit an artifact to EBI
Parameters
----------
artifact_id : int
The artifact id
action : %s
The action to perform with this data
send : bool
True to actually send the files
test : bool
If True some restrictions will be ignored, only used in parse_EBI_reply
"""
# step 1: init and validate
ebi_submission = EBISubmission(artifact_id, action)
# step 2: generate demux fastq files
try:
ebi_submission.generate_demultiplexed_fastq()
except Exception:
error_msg = format_exc()
if isdir(ebi_submission.full_ebi_dir):
rmtree(ebi_submission.full_ebi_dir)
LogEntry.create('Runtime',
error_msg,
info={'ebi_submission': artifact_id})
raise
# step 3: generate and write xml files
ebi_submission.generate_xml_files()
if send:
# getting aspera's password
old_ascp_pass = environ.get('ASPERA_SCP_PASS', '')
if old_ascp_pass == '':
environ['ASPERA_SCP_PASS'] = qiita_config.ebi_seq_xfer_pass
ascp_passwd = environ['ASPERA_SCP_PASS']
LogEntry.create('Runtime',
('Submission of sequences of pre_processed_id: '
'%d completed successfully' % artifact_id))
# step 4: sending sequences
if action != 'MODIFY':
LogEntry.create('Runtime',
("Submitting sequences for pre_processed_id: "
"%d" % artifact_id))
for cmd in ebi_submission.generate_send_sequences_cmd():
stdout, stderr, rv = system_call(cmd)
if rv != 0:
error_msg = ("ASCP Error:\nStd output:%s\nStd error:%s" %
(stdout, stderr))
environ['ASPERA_SCP_PASS'] = old_ascp_pass
raise ComputeError(error_msg)
open(ebi_submission.ascp_reply,
'a').write('stdout:\n%s\n\nstderr: %s' % (stdout, stderr))
environ['ASPERA_SCP_PASS'] = old_ascp_pass
# step 5: sending xml and parsing answer
xmls_cmds = ebi_submission.generate_curl_command(
ebi_seq_xfer_pass=ascp_passwd)
LogEntry.create('Runtime', ("Submitting XMLs for pre_processed_id: "
"%d" % artifact_id))
xml_content, stderr, rv = system_call(xmls_cmds)
if rv != 0:
error_msg = ("Error:\nStd output:%s\nStd error:%s" %
(xml_content, stderr))
raise ComputeError(error_msg)
else:
LogEntry.create('Runtime',
('Submission of sequences of pre_processed_id: '
'%d completed successfully' % artifact_id))
open(ebi_submission.curl_reply,
'w').write('stdout:\n%s\n\nstderr: %s' % (xml_content, stderr))
try:
st_acc, sa_acc, bio_acc, ex_acc, run_acc = \
ebi_submission.parse_EBI_reply(xml_content, test=test)
except EBISubmissionError as e:
error = str(e)
le = LogEntry.create('Fatal',
"Command: %s\nError: %s\n" %
(xml_content, error),
info={'ebi_submission': artifact_id})
raise ComputeError("EBI Submission failed! Log id: %d\n%s" %
(le.id, error))
if action == 'ADD' or test:
if st_acc:
ebi_submission.study.ebi_study_accession = st_acc
if sa_acc:
ebi_submission.sample_template.ebi_sample_accessions = sa_acc
if bio_acc:
ebi_submission.sample_template.biosample_accessions = bio_acc
if ex_acc:
ebi_submission.prep_template.ebi_experiment_accessions = ex_acc
ebi_submission.artifact.ebi_run_accessions = run_acc
else:
st_acc, sa_acc, bio_acc, ex_acc, run_acc = None, None, None, None, None
return st_acc, sa_acc, bio_acc, ex_acc, run_acc
def submit_VAMPS(artifact_id):
"""Submit artifact to VAMPS
Parameters
----------
artifact_id : int
The artifact id
Raises
------
ComputeError
- If the artifact cannot be submitted to VAMPS
- If the artifact is associated with more than one prep template
"""
artifact = Artifact(artifact_id)
if not artifact.can_be_submitted_to_vamps:
raise ComputeError("Artifact %d cannot be submitted to VAMPS" %
artifact_id)
study = artifact.study
sample_template = study.sample_template
prep_templates = artifact.prep_templates
if len(prep_templates) > 1:
raise ComputeError(
"Multiple prep templates associated with the artifact: %s" %
artifact_id)
prep_template = prep_templates[0]
# Also need to check that is not submitting (see item in #1523)
if artifact.is_submitted_to_vamps:
raise ValueError("Cannot resubmit artifact %s to VAMPS!" % artifact_id)
# Generating a tgz
targz_folder = mkdtemp(prefix=qiita_config.working_dir)
targz_fp = join(targz_folder,
'%d_%d_%d.tgz' % (study.id, prep_template.id, artifact_id))
targz = taropen(targz_fp, mode='w:gz')
# adding sample/prep
samp_fp = join(targz_folder, 'sample_metadata.txt')
sample_template.to_file(samp_fp)
targz.add(samp_fp, arcname='sample_metadata.txt')
prep_fp = join(targz_folder, 'prep_metadata.txt')
prep_template.to_file(prep_fp)
targz.add(prep_fp, arcname='prep_metadata.txt')
# adding preprocessed data
for _, fp, fp_type in artifact.filepaths:
if fp_type == 'preprocessed_fasta':
targz.add(fp, arcname='preprocessed_fasta.fna')
targz.close()
# submitting
cmd = ("curl -F user=%s -F pass='******' -F uploadFile=@%s -F "
"press=UploadFile %s" %
(qiita_config.vamps_user, qiita_config.vamps_pass, targz_fp,
qiita_config.vamps_url))
obs, stderr, rv = system_call(cmd)
if rv != 0:
error_msg = ("Error:\nStd output:%s\nStd error:%s" % (obs, stderr))
raise ComputeError(error_msg)
exp = ("<html>\n<head>\n<title>Process Uploaded File</title>\n</head>\n"
"<body>\n</body>\n</html>")
if obs != exp:
return False
else:
artifact.is_submitted_to_vamps = True
return True
def submit_EBI(artifact_id, action, send, test=False, test_size=False):
"""Submit an artifact to EBI
Parameters
----------
artifact_id : int
The artifact id
action : %s
The action to perform with this data
send : bool
True to actually send the files
test : bool
If True some restrictions will be ignored, only used in parse_EBI_reply
test_size : bool
If True the EBI-ENA restriction size will be changed to 6000
"""
# step 1: init and validate
ebi_submission = EBISubmission(artifact_id, action)
# step 2: generate demux fastq files
try:
ebi_submission.generate_demultiplexed_fastq()
except Exception:
error_msg = format_exc()
if isdir(ebi_submission.full_ebi_dir):
rmtree(ebi_submission.full_ebi_dir)
LogEntry.create('Runtime',
error_msg,
info={'ebi_submission': artifact_id})
raise
# step 3: generate and write xml files
ebi_submission.generate_xml_files()
# before we continue let's check the size of the submission
to_review = [
ebi_submission.study_xml_fp, ebi_submission.sample_xml_fp,
ebi_submission.experiment_xml_fp, ebi_submission.run_xml_fp,
ebi_submission.submission_xml_fp
]
total_size = sum([stat(tr).st_size for tr in to_review if tr is not None])
# note that the max for EBI is 10M but let's play it safe
max_size = 10e+6 if not test_size else 5000
if total_size > max_size:
LogEntry.create(
'Runtime', 'The submission: %d is larger than allowed (%d), will '
'try to fix: %d' % (artifact_id, max_size, total_size))
# transform current metadata to dataframe for easier curation
rows = {k: dict(v) for k, v in ebi_submission.samples.items()}
df = pd.DataFrame.from_dict(rows, orient='index')
# remove unique columns and same value in all columns
nunique = df.apply(pd.Series.nunique)
nsamples = len(df.index)
cols_to_drop = set(nunique[(nunique == 1) |
(nunique == nsamples)].index)
# maximize deletion by removing also columns that are almost all the
# same or almost all unique
cols_to_drop = set(nunique[(nunique <= int(nsamples * .01)) |
(nunique >= int(nsamples * .5))].index)
cols_to_drop = cols_to_drop - {
'taxon_id', 'scientific_name', 'description'
}
all_samples = ebi_submission.sample_template.ebi_sample_accessions
samples = [k for k in ebi_submission.samples if all_samples[k] is None]
if samples:
ebi_submission.write_xml_file(
ebi_submission.generate_sample_xml(samples, cols_to_drop),
ebi_submission.sample_xml_fp)
# now let's recalculate the size to make sure it's fine
new_total_size = sum(
[stat(tr).st_size for tr in to_review if tr is not None])
LogEntry.create(
'Runtime', 'The submission: %d after cleaning is %d and was %d' %
(artifact_id, total_size, new_total_size))
if new_total_size > max_size:
raise ComputeError(
'Even after cleaning the submission: %d is too large. Before '
'cleaning: %d, after: %d' %
(artifact_id, total_size, new_total_size))
st_acc, sa_acc, bio_acc, ex_acc, run_acc = None, None, None, None, None
if send:
# getting aspera's password
old_ascp_pass = environ.get('ASPERA_SCP_PASS', '')
if old_ascp_pass == '':
environ['ASPERA_SCP_PASS'] = qiita_config.ebi_seq_xfer_pass
ascp_passwd = environ['ASPERA_SCP_PASS']
LogEntry.create('Runtime',
('Submission of sequences of pre_processed_id: '
'%d completed successfully' % artifact_id))
# step 4: sending sequences
if action != 'MODIFY':
LogEntry.create('Runtime',
("Submitting sequences for pre_processed_id: "
"%d" % artifact_id))
for cmd in ebi_submission.generate_send_sequences_cmd():
stdout, stderr, rv = system_call(cmd)
if rv != 0:
error_msg = ("ASCP Error:\nStd output:%s\nStd error:%s" %
(stdout, stderr))
environ['ASPERA_SCP_PASS'] = old_ascp_pass
raise ComputeError(error_msg)
open(ebi_submission.ascp_reply,
'a').write('stdout:\n%s\n\nstderr: %s' % (stdout, stderr))
environ['ASPERA_SCP_PASS'] = old_ascp_pass
# step 5: sending xml
xmls_cmds = ebi_submission.generate_curl_command(
ebi_seq_xfer_pass=ascp_passwd)
LogEntry.create('Runtime', ("Submitting XMLs for pre_processed_id: "
"%d" % artifact_id))
xml_content, stderr, rv = system_call(xmls_cmds)
if rv != 0:
error_msg = ("Error:\nStd output:%s\nStd error:%s" %
(xml_content, stderr))
raise ComputeError(error_msg)
else:
LogEntry.create('Runtime',
('Submission of sequences of pre_processed_id: '
'%d completed successfully' % artifact_id))
open(ebi_submission.curl_reply,
'w').write('stdout:\n%s\n\nstderr: %s' % (xml_content, stderr))
# parsing answer / only if adding
if action == 'ADD' or test:
try:
st_acc, sa_acc, bio_acc, ex_acc, run_acc = \
ebi_submission.parse_EBI_reply(xml_content, test=test)
except EBISubmissionError as e:
error = str(e)
le = LogEntry.create('Fatal',
"Command: %s\nError: %s\n" %
(xml_content, error),
info={'ebi_submission': artifact_id})
raise ComputeError("EBI Submission failed! Log id: %d\n%s" %
(le.id, error))
if st_acc:
ebi_submission.study.ebi_study_accession = st_acc
if sa_acc:
ebi_submission.sample_template.ebi_sample_accessions = sa_acc
if bio_acc:
ebi_submission.sample_template.biosample_accessions = bio_acc
if ex_acc:
ebi_submission.prep_template.ebi_experiment_accessions = ex_acc
ebi_submission.artifact.ebi_run_accessions = run_acc
return st_acc, sa_acc, bio_acc, ex_acc, run_acc
def submit_EBI(preprocessed_data_id, action, send):
"""Submit a preprocessed data to EBI
Parameters
----------
preprocessed_data_id : int
The preprocesssed data id
action : %s
The action to perform with this data
send : bool
True to actually send the files
"""
# step 1: init and validate
ebi_submission = EBISubmission(preprocessed_data_id, action)
# step 2: generate demux fastq files
ebi_submission.study.ebi_submission_status = 'submitting'
try:
ebi_submission.generate_demultiplexed_fastq()
except:
error_msg = format_exc()
if isdir(ebi_submission.full_ebi_dir):
rmtree(ebi_submission.full_ebi_dir)
ebi_submission.study.ebi_submission_status = 'failed: %s' % error_msg
LogEntry.create('Runtime',
error_msg,
info={'ebi_submission': preprocessed_data_id})
raise
# step 3: generate and write xml files
ebi_submission.generate_xml_files()
if send:
# step 4: sending sequences
if action != 'MODIFY':
old_ascp_pass = environ.get('ASPERA_SCP_PASS', '')
environ['ASPERA_SCP_PASS'] = qiita_config.ebi_seq_xfer_pass
LogEntry.create('Runtime',
("Submitting sequences for pre_processed_id: "
"%d" % preprocessed_data_id))
for cmd in ebi_submission.generate_send_sequences_cmd():
stdout, stderr, rv = system_call(cmd)
if rv != 0:
error_msg = ("Error:\nStd output:%s\nStd error:%s" %
(stdout, stderr))
raise ComputeError(error_msg)
open(ebi_submission.ascp_reply,
'a').write('stdout:\n%s\n\nstderr: %s' % (stdout, stderr))
environ['ASPERA_SCP_PASS'] = old_ascp_pass
LogEntry.create(
'Runtime',
('Submission of sequences of pre_processed_id: '
'%d completed successfully' % preprocessed_data_id))
# step 5: sending xml and parsing answer
xmls_cmds = ebi_submission.generate_curl_command()
LogEntry.create('Runtime', ("Submitting XMLs for pre_processed_id: "
"%d" % preprocessed_data_id))
xml_content, stderr, rv = system_call(xmls_cmds)
if rv != 0:
error_msg = ("Error:\nStd output:%s\nStd error:%s" %
(xml_content, stderr))
raise ComputeError(error_msg)
else:
LogEntry.create(
'Runtime',
('Submission of sequences of pre_processed_id: '
'%d completed successfully' % preprocessed_data_id))
open(ebi_submission.curl_reply,
'w').write('stdout:\n%s\n\nstderr: %s' % (xml_content, stderr))
try:
st_acc, sa_acc, bio_acc, ex_acc, run_acc = \
ebi_submission.parse_EBI_reply(xml_content)
except EBISubmissionError as e:
le = LogEntry.create('Fatal',
"Command: %s\nError: %s\n" %
(xml_content, str(e)),
info={'ebi_submission': preprocessed_data_id})
ebi_submission.study.ebi_submission_status = (
"failed: XML parsing, log id: %d" % le.id)
raise ComputeError("EBI Submission failed! Log id: %d" % le.id)
ebi_submission.study.ebi_submission_status = 'submitted'
if action == 'ADD':
if st_acc:
ebi_submission.study.ebi_study_accession = st_acc
if sa_acc:
ebi_submission.sample_template.ebi_sample_accessions = sa_acc
if bio_acc:
ebi_submission.sample_template.biosample_accessions = bio_acc
if ex_acc:
ebi_submission.prep_template.ebi_experiment_accessions = ex_acc
ebi_submission.artifact.ebi_run_accessions = run_acc
else:
st_acc, sa_acc, bio_acc, ex_acc, run_acc = None, None, None, None, None
return st_acc, sa_acc, bio_acc, ex_acc, run_acc