def import_bedfiles(
cls, technique, targets_path, baits_path, assembly, species, description=None
):
"""
Register input_bed_path in technique's storage dir and update `data`.
Arguments:
technique (str): technique slug.
targets_path (str): path to targets bedfile.
baits_path (str): path to baits bedfile.
assembly (str): name of reference genome for bedfile.
species (str): name of genome species.
description (str): a description of the BED files.
Returns:
dict: updated technique instance as retrieved from API.
"""
utils.check_admin()
technique = api.get_instance("techniques", technique)
targets_key = f"{assembly}_targets_bedfile"
baits_key = f"{assembly}_baits_bedfile"
if targets_key in technique["reference_data"]:
raise click.UsageError(
f"Technique '{technique['slug']}' "
f"has registered BED files for '{assembly}':\n"
f'\n\t{technique["reference_data"][targets_key]}'
f'\n\t{technique["reference_data"][baits_key]}'
)
if not technique["storage_url"]:
technique = update_storage_url("techniques", technique["pk"])
api.create_instance("assemblies", name=assembly, species=species)
beds_dir = join(technique["storage_url"], "bed_files", assembly)
base_name = slugify(f'{technique["slug"]}.{assembly}')
targets_dst = join(beds_dir, f"{base_name}.targets.bed")
baits_dst = join(beds_dir, f"{base_name}.baits.bed")
os.makedirs(beds_dir, exist_ok=True)
for src, dst in [(targets_path, targets_dst), (baits_path, baits_dst)]:
cls.echo_src_dst("Copying", src, dst)
shutil.copy(src, dst)
click.secho(f"\nProcessing {basename(dst)}...", fg="blue")
cls.process_bedfile(dst)
click.secho(f'\nSuccess! patching {technique["slug"]}...', fg="green")
for i, j in [(targets_key, targets_dst), (baits_key, baits_dst)]:
technique["reference_data"][i] = {
"url": j + ".gz",
"description": description,
}
return api.patch_instance(
endpoint="techniques",
instance_id=technique["pk"],
storage_usage=utils.get_tree_size(technique["storage_url"]),
reference_data=technique["reference_data"],
)
def import_data(
cls,
identifier,
data_src,
data_id,
symlink,
description,
sub_dir=None,
model="assemblies",
):
"""
Register reference resources for a given assembly.
Arguments:
identifier (str): name of assembly or technique.
model (str): either `techniques` or `assemblies`.
data_src (str): path to reference data.
data_id (str): identifier that will be used for reference data.
symlink (str): symlink instead of move.
description (str): reference data description.
sub_dir (str): target sub dir for the resource, default is data_id.
Returns:
dict: updated assembly instance as retrieved from API.
"""
utils.check_admin()
data_id = slugify(data_id, separator="_")
click.echo(f'`data_id` set to: {click.style(data_id, fg="green")}')
instance = api.get_instance(model, identifier)
if data_id in instance["reference_data"]:
raise click.UsageError(
f"{instance['name']} has already reference data registered with id "
f'"{data_id}":\n\n\t{instance["reference_data"][data_id]}'
)
if not instance["storage_url"]:
instance = update_storage_url(model, instance["name"])
data_dir = join(instance["storage_url"], sub_dir or data_id)
data_dst = join(data_dir, basename(data_src))
os.makedirs(data_dir, exist_ok=True)
if symlink:
cls.echo_src_dst("Linking", data_src, data_dst)
cls.symlink(data_src, data_dst)
else:
cls.echo_src_dst("Moving", data_src, data_dst)
cls.move(data_src, data_dst)
click.secho(f'\nSuccess! patching {instance["name"]}...', fg="green")
instance["reference_data"][data_id] = {}
instance["reference_data"][data_id]["url"] = data_dst
instance["reference_data"][data_id]["description"] = description
return api.patch_instance(
endpoint=model,
instance_id=instance["pk"],
storage_usage=utils.get_tree_size(instance["storage_url"]),
reference_data=instance["reference_data"],
)
def process_finished(filters):
"""Process and update finished analyses."""
utils.check_admin()
filters.update(status="FINISHED")
for i in api.get_instances("analyses", verbose=True, **filters):
if i["status"] == "FINISHED":
api.patch_analysis_status(i, "SUCCEEDED")
def test_check_admin():
admin = _DEFAULTS["ADMIN_USER"]
_DEFAULTS["ADMIN_USER"] = "******"
with pytest.raises(PermissionError) as error:
utils.check_admin()
assert "not the admin" in str(error.value)
_DEFAULTS["ADMIN_USER"] = admin
def _set_analysis_permissions(analysis):
protect_results = analysis.status == "SUCCEEDED"
unique_analysis_per_individual = False
application_protect_results = True
chgrp_cmd = (
["false"]
if not system_settings.DEFAULT_LINUX_GROUP
else ["chgrp", "-R", system_settings.DEFAULT_LINUX_GROUP, analysis.storage_url]
)
try:
application = import_from_string(analysis.application.application_class)()
unique_analysis_per_individual = application.unique_analysis_per_individual
application_protect_results = application.application_protect_results
except ImportError:
pass
if (
# dont protect results if project level analysis
analysis.project_level_analysis
# dont protect results if individual level automerge
or (analysis.individual_level_analysis and not unique_analysis_per_individual)
# dont protect results if the application says so
or not application_protect_results
):
protect_results = False
if protect_results:
utils.check_admin()
if analysis.ran_by != system_settings.api_username:
src = analysis.storage_url + "__tmp"
shutil.move(analysis.storage_url, src)
cmd = utils.get_rsync_command(src, analysis.storage_url, chmod="a-w")
subprocess.check_call(cmd, shell=True)
else:
subprocess.check_call(["chmod", "-R", "a-w", analysis.storage_url])
try:
subprocess.check_output(chgrp_cmd, stderr=subprocess.STDOUT)
except subprocess.CalledProcessError:
pass
elif not protect_results or analysis.status in {"FAILED", "FINISHED"}:
for i in [chgrp_cmd, ["chmod", "-R", "g+rwX", analysis.storage_url]]:
try:
subprocess.check_output(i, stderr=subprocess.STDOUT)
except subprocess.CalledProcessError:
pass
def patch_results(filters, force):
"""Update the results field of many analyses."""
utils.check_admin()
skipped = []
with click.progressbar(
api.get_instances("analyses", verbose=True, **filters),
label="Patching analyses...",
) as bar:
for i in bar:
if force or not i.results:
results = api._get_analysis_results(i, raise_error=False)
api.patch_instance("analyses", i.pk, results=results)
else: # pragma: no cover
skipped.append(i)
if skipped: # pragma: no cover
click.echo(
f"{len(skipped)} analyses had results, use --force to update...")
def update_experiment_bam_file(experiment, assembly_name, analysis_pk, bam_url):
"""
Update default bam for a experiment given the assembly.
Arguments:
experiment (dict): experiment dict.
assembly_name (str): assembly name.
analysis_pk (int): analysis primary key.
bam_url (str): bam url.
Returns:
dict: patched experiment instance
"""
utils.check_admin()
pk = experiment["pk"]
bam_files = experiment["bam_files"]
if bam_files.get(assembly_name, None): # pragma: no cover
raise click.UsageError(f"Experiment {pk} already has {assembly_name} bam")
bam_files[assembly_name] = {"url": bam_url, "analysis": analysis_pk}
return api.patch_instance("experiments", pk, bam_files=bam_files)
def import_data(
self,
directories,
symlink=False,
commit=False,
key=lambda x: x["system_id"],
files_data=None,
dtypes=None,
**filters,
):
"""
Import raw data for multiple experiments.
Experiments's `storage_url`, `storage_usage`, `raw_data` are
updated.
Arguments:
directories (list): list of directories to be recursively explored.
symlink (bool): if True symlink instead of moving.
commit (bool): if True perform import operation.
key (function): given a experiment dict returns id to match.
filters (dict): key value pairs to use as API query params.
dtypes (list): data types that should be matched (e.g. BAM, PNG. etc.).
files_data (dict): keys are files basenames and values are
dicts with extra annotations such as PL, LB, or any other,
see also annotate_file_data.
Raises:
click.UsageError: if `key` returns the same identifier for multiple
experiments. If a experiment matches both fastq and bam files.
if cant determine read 1 or read 2 from matched fastq files.
Returns:
tuple: list of experiments for which data has been matched and a
summary of the operation.
"""
utils.check_admin()
files_data = files_data or {}
experiments_matched = []
cache = defaultdict(dict)
patterns = []
identifiers = {}
dtypes = set(dtypes or [])
# validate files_data
for i, j in files_data.items():
if not isinstance(j, dict): # pragma: no cover
raise click.UsageError(f"Invalid file data, expected dict {i}: {j}")
# get experiments and load cache dictionary
for i in api.get_instances("experiments", verbose=True, **filters):
index = f"primary_key_{i['pk']}"
using_id = f"{i['system_id']} (Skipped, identifier is NULL)"
identifier = key(i)
if identifier in identifiers: # duplicated identifiers not valid
raise click.UsageError(
f"Can't use same identifier for {i['system_id']} "
f"and {identifiers[identifier]}: {identifier}"
)
if i["raw_data"] or i["bam_files"]:
using_id = f"{i['system_id']} (Skipped, experiment has raw data)"
elif identifier:
identifiers[identifier] = i["system_id"]
patterns.append(self.get_regex_pattern(index, identifier))
using_id = f"{i['system_id']} (using {identifier})"
cache[index]["using_id"] = using_id
cache[index]["instance"] = i
cache[index]["files"] = []
if patterns:
# see http://stackoverflow.com/questions/8888567 for pattern
pattern = re.compile("|".join(patterns))
data_storage_dir = system_settings.BASE_STORAGE_DIRECTORY
label = f"Exploring directories..."
# explore dirs
for directory in set(directories):
with click.progressbar(os.walk(directory), label=label) as bar:
for root, _, files in bar:
if not root.startswith(data_storage_dir):
for i in files:
if len(patterns) > 500: # pragma: no cover
click.echo(
f"Matching {i} against "
f"{len(patterns)} experiments..."
)
path = join(root, i)
match = self.match_path(path, pattern)
if match and (not dtypes or match["dtype"] in dtypes):
cache[match.pop("index")]["files"].append(match)
# process files if needed
label = "Processing..."
bar = sorted(cache.values(), key=lambda x: x["instance"]["pk"])
with click.progressbar(bar, label=label) as bar:
for i in bar:
if commit and i["files"]:
experiments_matched.append(
self.import_files(
instance=i["instance"],
files=i["files"],
symlink=symlink,
files_data=files_data,
)
)
elif i["files"]: # pragma: no cover
experiments_matched.append(i["instance"])
return experiments_matched, self.get_summary(cache)
