def _get_actual_file_if_queueable(
extracted_subfile: Dict,
original_file: OriginalFile,
samples: List[Sample]) -> OriginalFile:
"""Returns the actual file from the archive if it should be queued.
If the file has been processed or has an unstarted DownloaderJob,
None will be returned.
`extracted_subfile` should be a Dict containing metadata about the
file that was extracted from an archive.
`original_file` should be the file associated with the CURRENT
DownloaderJob.
`samples` are the samples that the actual file should be associated
with if it has to be created.
"""
# Check to see if we've made this original file before:
potential_existing_files = OriginalFile.objects.filter(
source_filename=original_file.source_filename,
filename=extracted_subfile['filename'],
is_archive=False
)
if potential_existing_files.count() > 0:
# We've already created this record, let's see if we actually
# needed to download it or if we just got it because we needed
# a file in the same archive.
actual_file = potential_existing_files[0]
if actual_file.needs_processing():
if not actual_file.is_downloaded:
actual_file.is_downloaded = True
actual_file.save()
return actual_file
else:
return None
else:
actual_file = OriginalFile()
actual_file.is_downloaded = True
actual_file.is_archive = False
actual_file.absolute_file_path = extracted_subfile['absolute_path']
actual_file.filename = extracted_subfile['filename']
actual_file.calculate_size()
actual_file.calculate_sha1()
actual_file.has_raw = True
actual_file.source_url = original_file.source_url
actual_file.source_filename = original_file.source_filename
actual_file.save()
for sample in samples:
original_file_sample_association = OriginalFileSampleAssociation()
original_file_sample_association.sample = sample
original_file_sample_association.original_file = actual_file
original_file_sample_association.save()
return actual_file
def download_geo(job_id: int) -> None:
"""The main function for the GEO Downloader.
Downloads a single tar file containing the files representing
samples relating to a single experiment stored in
GEO.
"""
job = utils.start_job(job_id)
accession_code = job.accession_code
original_file = job.original_files.first()
if not original_file:
job.failure_reason = "No files associated with the job."
logger.error("No files associated with the job.", downloader_job=job_id)
utils.end_downloader_job(job, success=False)
return
url = original_file.source_url
related_samples = original_file.samples.exclude(technology="RNA-SEQ")
# First, download the sample archive URL.
# Then, unpack all the ones downloaded.
# Then create processor jobs!
# The files for all of the samples are contained within the same zip file. Therefore only
# download the one.
os.makedirs(LOCAL_ROOT_DIR + "/" + accession_code, exist_ok=True)
dl_file_path = LOCAL_ROOT_DIR + "/" + accession_code + "/" + url.split("/")[-1]
logger.debug("Starting to download: " + url, job_id=job_id, accession_code=accession_code)
_download_file(url, dl_file_path, job)
original_file.absolute_file_path = dl_file_path
original_file.is_downloaded = True
original_file.save()
unpacked_sample_files = []
try:
# enumerate all files inside the archive
archived_files = list(ArchivedFile(dl_file_path).get_files())
except FileExtractionError as e:
job.failure_reason = e
logger.exception(
"Error occurred while extracting file.", path=dl_file_path, exception=str(e)
)
utils.end_downloader_job(job, success=False)
return
for og_file in archived_files:
sample = og_file.get_sample()
# We don't want RNA-Seq data from GEO:
# https://github.com/AlexsLemonade/refinebio/issues/966
if sample and sample.technology == "RNA-SEQ":
logger.warn("RNA-Seq sample found in GEO downloader job.", sample=sample)
continue
if not sample and (
not og_file.is_processable() or og_file.experiment_accession_code() != accession_code
):
# skip the files that we know are not processable and can't be associated with a sample
# also skip the files were we couldn't find a sample and they don't mention the current experiment
continue
potential_existing_file = OriginalFile.objects.filter(
source_filename=original_file.source_filename,
filename=og_file.filename,
is_archive=False,
).first()
if potential_existing_file:
# We've already created this record, let's see if we actually
# needed to download it or if we just got it because we needed
# a file in the same archive.
if potential_existing_file.needs_processing():
if not potential_existing_file.is_downloaded:
potential_existing_file.is_downloaded = True
potential_existing_file.save()
unpacked_sample_files.append(potential_existing_file)
continue
# Then this is a new file and we should create an original file for it
actual_file = OriginalFile()
actual_file.is_downloaded = True
actual_file.is_archive = False
actual_file.absolute_file_path = og_file.file_path
actual_file.filename = og_file.filename
actual_file.calculate_size()
actual_file.calculate_sha1()
actual_file.has_raw = True
actual_file.source_url = original_file.source_url
actual_file.source_filename = original_file.source_filename
actual_file.save()
# try to see if the file should be associated with a sample
if sample:
original_file_sample_association = OriginalFileSampleAssociation()
original_file_sample_association.sample = sample
original_file_sample_association.original_file = actual_file
original_file_sample_association.save()
else:
# if not, we can associate this file with all samples in the experiment
for sample in related_samples:
original_file_sample_association = OriginalFileSampleAssociation()
original_file_sample_association.sample = sample
original_file_sample_association.original_file = actual_file
original_file_sample_association.save()
unpacked_sample_files.append(actual_file)
if len(unpacked_sample_files) > 0:
success = True
logger.debug(
"File downloaded and extracted successfully.",
url=url,
dl_file_path=dl_file_path,
downloader_job=job_id,
)
else:
success = False
logger.info(
"Unable to extract any files.",
url=url,
dl_file_path=dl_file_path,
downloader_job=job_id,
)
job.failure_reason = "Failed to extract any downloaded files."
if success:
create_processor_jobs_for_original_files(unpacked_sample_files, job)
if original_file.is_archive:
original_file.delete_local_file()
utils.end_downloader_job(job, success)
return success
def download_transcriptome(job_id: int) -> None:
"""The main function for the Transcriptome Index Downloader.
Two files are needed for the Transcriptome Index Downloader: a
fasta file and a gtf file. However each pair need to be processed
into two different sized indices. (See the
processors.transcriptome_index._create_index function's docstring
for more info.) Therefore we only download each set once, then
let each processor find it in the same location.
"""
job = utils.start_job(job_id)
file_assocs = DownloaderJobOriginalFileAssociation.objects.filter(
downloader_job=job)
long_files_to_process = []
short_files_to_process = []
for assoc in file_assocs:
long_original_file = assoc.original_file
if long_original_file.is_archive:
filename_species = "".join(
long_original_file.source_filename.split(".")[:-2])
else:
# Does this ever happen?
filename_species = "".join(
long_original_file.source_filename.split(".")[:-1])
# First download the files and make the original files for the
# long transciptome index.
long_dir = os.path.join(LOCAL_ROOT_DIR, filename_species + "_long")
os.makedirs(long_dir, exist_ok=True)
long_dl_file_path = os.path.join(long_dir,
long_original_file.source_filename)
job = _download_file(long_original_file.source_url, long_dl_file_path,
job)
if not job.success:
break
long_original_file.is_downloaded = True
long_original_file.absolute_file_path = long_dl_file_path
long_original_file.filename = long_original_file.source_filename
long_original_file.has_raw = True
long_original_file.calculate_size()
long_original_file.calculate_sha1()
long_original_file.save()
long_files_to_process.append(long_original_file)
# Next copy the files to another directory and create the
# original files for the short transcriptome index.
short_dir = os.path.join(LOCAL_ROOT_DIR, filename_species + "_short")
os.makedirs(short_dir, exist_ok=True)
short_dl_file_path = os.path.join(short_dir,
long_original_file.source_filename)
shutil.copyfile(long_dl_file_path, short_dl_file_path)
short_original_file = OriginalFile(
source_filename=long_original_file.source_filename,
source_url=long_original_file.source_url,
is_downloaded=True,
absolute_file_path=short_dl_file_path,
filename=long_original_file.filename,
has_raw=True,
)
short_original_file.calculate_size()
short_original_file.calculate_sha1()
short_original_file.save()
short_files_to_process.append(short_original_file)
if job.success:
logger.debug("Files downloaded successfully.", downloader_job=job_id)
create_long_and_short_processor_jobs(job, long_files_to_process,
short_files_to_process)
utils.end_downloader_job(job, job.success)
def download_geo(job_id: int) -> None:
"""The main function for the GEO Downloader.
Downloads a single tar file containing the files representing
samples relating to a single experiement stored in
GEO.
"""
job = utils.start_job(job_id)
file_assocs = DownloaderJobOriginalFileAssociation.objects.filter(
downloader_job=job)
original_file = file_assocs[0].original_file
url = original_file.source_url
accession_code = job.accession_code
sample_assocs = OriginalFileSampleAssociation.objects.filter(
original_file=original_file)
related_samples = Sample.objects.filter(
id__in=sample_assocs.values('sample_id'))
# First, download the sample archive URL.
# Then, unpack all the ones downloaded.
# Then create processor jobs!
# The files for all of the samples are
# contained within the same zip file. Therefore only
# download the one.
os.makedirs(LOCAL_ROOT_DIR + '/' + accession_code, exist_ok=True)
dl_file_path = LOCAL_ROOT_DIR + '/' + accession_code + '/' + url.split(
'/')[-1]
logger.debug("Starting to download: " + url,
job_id=job_id,
accession_code=accession_code)
_download_file(url, dl_file_path, job)
original_file.absolute_file_path = dl_file_path
original_file.is_downloaded = True
original_file.save()
has_raw = True
unpacked_sample_files = []
# These files are tarred, and also subsequently gzipped
if '.tar' in dl_file_path:
try:
extracted_files = _extract_tar(dl_file_path, accession_code)
except Exception as e:
job.failure_reason = e
logger.exception("Error occured while extracting tar file.",
path=dl_file_path,
exception=str(e))
utils.end_downloader_job(job, success=False)
return
for og_file in extracted_files:
filename = og_file['filename']
if '_' in filename:
sample_id = filename.split('_')[0]
else:
sample_id = filename.split('.')[0]
try:
sample = Sample.objects.get(accession_code=sample_id)
except Exception as e:
# We don't have this sample, but it's not a total failure. This happens.
continue
try:
# Files from the GEO supplemental file are gzipped inside of the tarball. Great!
archive_file = OriginalFile.objects.get(
source_filename__contains=sample_id)
archive_file.is_downloaded = True
archive_file.is_archive = True
archive_file.absolute_file_path = og_file['absolute_path']
archive_file.calculate_size()
archive_file.calculate_sha1()
archive_file.save()
if '.gz' in og_file['filename']:
extracted_subfile = _extract_gz(og_file['absolute_path'],
accession_code)
else:
extracted_subfile = [og_file]
actual_file = OriginalFile()
actual_file.is_downloaded = True
actual_file.is_archive = False
actual_file.absolute_file_path = extracted_subfile[0][
'absolute_path']
actual_file.filename = extracted_subfile[0]['filename']
actual_file.calculate_size()
actual_file.calculate_sha1()
actual_file.has_raw = True
actual_file.source_url = original_file.source_url
actual_file.source_filename = original_file.source_filename
actual_file.save()
original_file_sample_association = OriginalFileSampleAssociation(
)
original_file_sample_association.sample = sample
original_file_sample_association.original_file = actual_file
original_file_sample_association.save()
archive_file.delete_local_file()
archive_file.is_downloaded = False
archive_file.save()
unpacked_sample_files.append(actual_file)
except Exception as e:
# TODO - is this worth failing a job for?
logger.debug(
"Found a file we didn't have an OriginalFile for! Why did this happen?: "
+ og_file['filename'],
exc_info=1,
file=og_file['filename'],
sample_id=sample_id,
accession_code=accession_code)
# If we don't know why we have it, get rid of it.
os.remove(og_file["absolute_path"])
# This is a .tgz file.
elif '.tgz' in dl_file_path:
# If this is the MINiML file, it has been preprocessed
if '_family.xml.tgz' in dl_file_path:
has_raw = False
try:
extracted_files = _extract_tgz(dl_file_path, accession_code)
except Exception as e:
job.failure_reason = e
logger.exception("Error occured while extracting tgz file.",
path=dl_file_path,
exception=str(e))
utils.end_downloader_job(job, success=False)
return
for og_file in extracted_files:
if '.txt' in og_file['filename']:
try:
gsm_id = og_file['filename'].split('-')[0]
sample = Sample.objects.get(accession_code=gsm_id)
except Exception as e:
os.remove(og_file["absolute_path"])
continue
actual_file = OriginalFile()
actual_file.is_downloaded = True
actual_file.is_archive = False
actual_file.absolute_file_path = og_file['absolute_path']
actual_file.filename = og_file['filename']
actual_file.calculate_size()
actual_file.calculate_sha1()
actual_file.has_raw = has_raw
actual_file.source_url = original_file.source_url
actual_file.source_filename = original_file.source_filename
actual_file.save()
original_file_sample_association = OriginalFileSampleAssociation(
)
original_file_sample_association.sample = sample
original_file_sample_association.original_file = actual_file
original_file_sample_association.save()
unpacked_sample_files.append(actual_file)
# These files are only gzipped.
# These are generally the _actually_ raw (rather than the non-raw data in a RAW file) data
elif '.gz' in dl_file_path:
try:
extracted_files = _extract_gz(dl_file_path, accession_code)
except Exception as e:
job.failure_reason = e
logger.exception("Error occured while extracting gz file.",
path=dl_file_path,
exception=str(e))
utils.end_downloader_job(job, success=False)
return
for og_file in extracted_files:
filename = og_file['filename']
sample_id = filename.split('.')[0]
try:
# The archive we downloaded
archive_file = OriginalFile.objects.get(
source_filename__contains=filename)
archive_file.is_downloaded = True
archive_file.is_archive = True
archive_file.absolute_file_path = dl_file_path
archive_file.calculate_size()
archive_file.calculate_sha1()
archive_file.save()
actual_file = OriginalFile()
actual_file.is_downloaded = True
actual_file.is_archive = False
actual_file.absolute_file_path = og_file['absolute_path']
actual_file.filename = og_file['filename']
actual_file.calculate_size()
actual_file.calculate_sha1()
actual_file.has_raw = True
actual_file.source_url = original_file.source_url
actual_file.source_filename = original_file.source_filename
actual_file.save()
for sample in related_samples:
new_association = OriginalFileSampleAssociation()
new_association.original_file = actual_file
new_association.sample = sample
new_association.save()
archive_file.delete_local_file()
archive_file.is_downloaded = False
archive_file.save()
unpacked_sample_files.append(actual_file)
except Exception as e:
logger.debug(
"Found a file we didn't have an OriginalFile for! Why did this happen?: "
+ og_file['filename'],
exc_info=1,
file=og_file['filename'],
sample_id=sample_id,
accession_code=accession_code)
os.remove(og_file["absolute_path"])
# This is probably just a .txt file
else:
filename = dl_file_path.split('/')[-1]
sample_id = filename.split('_')[0]
actual_file = OriginalFile()
actual_file.is_downloaded = True
actual_file.is_archive = False
actual_file.absolute_file_path = dl_file_path
actual_file.filename = filename
actual_file.calculate_size()
actual_file.calculate_sha1()
actual_file.has_raw = True
actual_file.source_url = original_file.source_url
actual_file.source_filename = original_file.source_filename
actual_file.save()
for sample in related_samples:
new_association = OriginalFileSampleAssociation()
new_association.original_file = actual_file
new_association.sample = sample
new_association.save()
unpacked_sample_files.append(actual_file)
if len(unpacked_sample_files) > 0:
success = True
logger.debug("File downloaded and extracted successfully.",
url=url,
dl_file_path=dl_file_path,
downloader_job=job_id)
else:
success = False
logger.info("Unable to extract any files.",
url=url,
dl_file_path=dl_file_path,
downloader_job=job_id)
job.failure_reason = "Failed to extract any downloaded files."
if success:
utils.create_processor_jobs_for_original_files(unpacked_sample_files,
job)
if original_file.is_archive:
original_file.delete_local_file()
utils.end_downloader_job(job, success)
return success
