def update_job_metadata(self, total_bytes_to_transfer,
total_resource_count):
job = get_current_job()
if job:
job.extra_metadata["file_size"] = total_bytes_to_transfer
job.extra_metadata["total_resources"] = total_resource_count
job.save_meta()
def _ping(started, server, checkrate):
try:
ping_once(started, server=server)
connection.close()
return
except ConnectionError:
logger.warn(
"Ping failed (could not connect). Trying again in {} minutes.".format(
checkrate
)
)
except Timeout:
logger.warn(
"Ping failed (connection timed out). Trying again in {} minutes.".format(
checkrate
)
)
except RequestException as e:
logger.warn(
"Ping failed ({})! Trying again in {} minutes.".format(e, checkrate)
)
connection.close()
job = get_current_job()
if job and job in scheduler:
scheduler.change_execution_time(
job, datetime.datetime.now() + datetime.timedelta(seconds=checkrate * 60)
)
def handle_async(self, *args, **options):
channel_id = options["channel_id"]
node_ids = options["node_ids"]
exclude_node_ids = options["exclude_node_ids"]
force_delete = options["force_delete"]
try:
channel = ChannelMetadata.objects.get(pk=channel_id)
except ChannelMetadata.DoesNotExist:
raise CommandError(
"Channel matching id {id} does not exist".format(id=channel_id)
)
delete_all_metadata = delete_metadata(
channel, node_ids, exclude_node_ids, force_delete
)
unused_files = LocalFile.objects.get_unused_files()
# Get orphan files that are being deleted
total_file_deletion_operations = unused_files.count()
job = get_current_job()
if job:
total_file_deletion_size = unused_files.aggregate(Sum("file_size")).get(
"file_size__sum", 0
)
job.extra_metadata["file_size"] = total_file_deletion_size
job.extra_metadata["total_resources"] = total_file_deletion_operations
job.save_meta()
progress_extra_data = {"channel_id": channel_id}
additional_progress = sum((1, bool(delete_all_metadata)))
with self.start_progress(
total=total_file_deletion_operations + additional_progress
) as progress_update:
for file in LocalFile.objects.delete_unused_files():
progress_update(1, progress_extra_data)
with db_task_write_lock:
LocalFile.objects.delete_orphan_file_objects()
progress_update(1, progress_extra_data)
if delete_all_metadata:
try:
os.remove(get_content_database_file_path(channel_id))
except OSError:
pass
progress_update(1, progress_extra_data)
def handle_async(self, *args, **options):
# set language for the translation of the messages
locale = settings.LANGUAGE_CODE if not options["locale"] else options["locale"]
translation.activate(locale)
self.overall_error = ""
job = get_current_job()
facility = self.get_facility(options)
if not facility:
self.overall_error = str(MESSAGES[NO_FACILITY])
else:
log_type = options["log_type"]
log_info = classes_info[log_type]
if options["output_file"] is None:
filename = log_info["filename"].format(facility.name, facility.id[:4])
else:
filename = options["output_file"]
filepath = os.path.join(os.getcwd(), filename)
queryset = log_info["queryset"]
total_rows = queryset.count()
with self.start_progress(total=total_rows) as progress_update:
try:
for row in csv_file_generator(
facility, log_type, filepath, overwrite=options["overwrite"]
):
progress_update(1)
except (ValueError, IOError) as e:
self.overall_error = str(MESSAGES[FILE_WRITE_ERROR].format(e))
if job:
job.extra_metadata["overall_error"] = self.overall_error
self.job.extra_metadata["filename"] = ntpath.basename(filepath)
job.save_meta()
else:
if self.overall_error:
raise CommandError(self.overall_error)
else:
logger.info(
"Created csv file {} with {} lines".format(filepath, total_rows)
)
translation.deactivate()
def _diskimport(
channel_id,
directory,
drive_id=None,
update_progress=None,
check_for_cancel=None,
node_ids=None,
is_updating=False,
exclude_node_ids=None,
extra_metadata=None,
):
call_command(
"importchannel",
"disk",
channel_id,
directory,
update_progress=update_progress,
check_for_cancel=check_for_cancel,
)
# Make some real-time updates to the metadata
job = get_current_job()
# Signal to UI that the DB-downloading step is done so it knows to display
# progress correctly
job.update_progress(0, 1.0)
job.extra_metadata["database_ready"] = True
# Add the channel name if it wasn't added initially
if job and job.extra_metadata.get("channel_name", "") == "":
job.extra_metadata["channel_name"] = get_channel_name(channel_id)
job.save_meta()
# Skip importcontent step if updating and no nodes have changed
if is_updating and (node_ids is not None) and len(node_ids) == 0:
pass
else:
call_command(
"importcontent",
"disk",
channel_id,
directory,
drive_id=drive_id,
node_ids=node_ids,
exclude_node_ids=exclude_node_ids,
update_progress=update_progress,
check_for_cancel=check_for_cancel,
)
def cancelable_job():
"""
Test function for checking if a job is cancelable. Meant to be used in a job cancel
test case.
It then calls the check_for_cancel, followed by a time.sleep function, 3 times.
:param check_for_cancel: A function that the BBQ framework passes in when a job is set to be cancellable.
Calling this function makes the thread check if a cancellation has been requested, and then exits early if true.
:return: None
"""
job = get_current_job()
for _ in range(10):
time.sleep(0.5)
if job.check_for_cancel():
return
def _remoteimport(
channel_id,
baseurl,
peer_id=None,
update_progress=None,
check_for_cancel=None,
node_ids=None,
is_updating=False,
exclude_node_ids=None,
extra_metadata=None,
):
call_command(
"importchannel",
"network",
channel_id,
baseurl=baseurl,
update_progress=update_progress,
check_for_cancel=check_for_cancel,
)
# Make some real-time updates to the metadata
job = get_current_job()
# Signal to UI that the DB-downloading step is done so it knows to display
# progress correctly
job.update_progress(0, 1.0)
job.extra_metadata["database_ready"] = True
# Add the channel name if it wasn't added initially
if job and job.extra_metadata.get("channel_name", "") == "":
job.extra_metadata["channel_name"] = get_channel_name(channel_id)
job.save_meta()
call_command(
"importcontent",
"network",
channel_id,
baseurl=baseurl,
peer_id=peer_id,
node_ids=node_ids,
exclude_node_ids=exclude_node_ids,
import_updates=is_updating,
update_progress=update_progress,
check_for_cancel=check_for_cancel,
)
def update_progress_cancelable_job():
"""
Test function for checking if a job is cancelable when it updates progress.
Meant to be used in a job cancel with progress update test case.
It then calls the check_for_cancel, followed by a time.sleep function, 10 times.
:param update_progress: A function that is called to update progress
:param check_for_cancel: A function that the iceqube framework passes in when a job is set to be cancellable.
Calling this function makes the thread check if a cancellation has been requested, and then exits early if true.
:return: None
"""
job = get_current_job()
for i in range(10):
time.sleep(0.5)
job.update_progress(i, 9)
if job.check_for_cancel():
return
def _diskimport(
channel_id,
directory,
drive_id=None,
update_progress=None,
check_for_cancel=None,
node_ids=None,
is_updating=False,
exclude_node_ids=None,
extra_metadata=None,
):
call_command(
"importchannel",
"disk",
channel_id,
directory,
update_progress=update_progress,
check_for_cancel=check_for_cancel,
)
# Add the channel name if it wasn't added initially
job = get_current_job()
if job and job.extra_metadata.get("channel_name", "") == "":
job.extra_metadata["channel_name"] = get_channel_name(channel_id)
job.save_meta()
# Skip importcontent step if updating and no nodes have changed
if is_updating and (node_ids is not None) and len(node_ids) == 0:
pass
else:
call_command(
"importcontent",
"disk",
channel_id,
directory,
drive_id=drive_id,
node_ids=node_ids,
exclude_node_ids=exclude_node_ids,
update_progress=update_progress,
check_for_cancel=check_for_cancel,
)
def handle_async(self, *args, **options):
# set language for the translation of the messages
locale = settings.LANGUAGE_CODE if not options["locale"] else options[
"locale"]
translation.activate(locale)
translate_labels()
self.overall_error = []
filepath = self.get_filepath(options)
facility = self.get_facility(options)
job = get_current_job()
total_rows = FacilityUser.objects.filter(facility=facility).count()
with self.start_progress(total=total_rows) as progress_update:
try:
for row in csv_file_generator(
facility,
filepath,
overwrite=options["overwrite"],
):
progress_update(1)
except (ValueError, IOError) as e:
self.overall_error.append(MESSAGES[FILE_WRITE_ERROR].format(e))
raise CommandError(self.overall_error[-1])
# freeze error messages translations:
self.overall_error = [str(msg) for msg in self.overall_error]
if job:
job.extra_metadata["overall_error"] = self.overall_error
job.extra_metadata["users"] = total_rows
job.extra_metadata["filename"] = ntpath.basename(filepath)
job.save_meta()
else:
logger.info("Created csv file {} with {} lines".format(
filepath, total_rows))
translation.deactivate()
def handle_async(self, *args, **options):
# initialize stats data structures:
self.overall_error = []
db_new_classes = []
db_update_classes = []
classes_to_clear = []
db_new_users = []
db_update_users = []
users_to_delete = []
per_line_errors = []
# set language for the translation of the messages
locale = settings.LANGUAGE_CODE if not options["locale"] else options[
"locale"]
translation.activate(locale)
self.job = get_current_job()
filepath = options["filepath"]
self.default_facility = self.get_facility(options)
self.number_lines = self.get_number_lines(filepath)
self.exit_if_error()
with self.start_progress(total=100) as self.progress_update:
# validate csv headers:
has_header = self.csv_headers_validation(filepath)
if not has_header:
self.overall_error.append(MESSAGES[INVALID_HEADER])
self.exit_if_error()
self.progress_update(1) # state=csv_headers
try:
with open(filepath) as f:
reader = csv.DictReader(f, strict=True)
per_line_errors, classes, users, roles = self.csv_values_validation(
reader, self.header_translation)
except (ValueError, FileNotFoundError, csv.Error) as e:
self.overall_error.append(MESSAGES[FILE_READ_ERROR].format(e))
self.exit_if_error()
(
db_new_users,
db_update_users,
keeping_users,
more_line_errors,
) = self.build_users_objects(users)
per_line_errors += more_line_errors
(
db_new_classes,
db_update_classes,
fixed_classes,
) = self.build_classes_objects(classes)
classes = fixed_classes
users_to_delete, classes_to_clear = self.get_delete(
options, keeping_users, db_update_classes)
per_line_errors += self.db_validate_list(db_new_users, users=True)
per_line_errors += self.db_validate_list(db_update_users,
users=True)
# progress = 91%
per_line_errors += self.db_validate_list(db_new_classes)
per_line_errors += self.db_validate_list(db_update_classes)
if not options["dryrun"]:
self.delete_users(users_to_delete)
# clear users from classes not included in the csv:
Membership.objects.filter(
collection__in=classes_to_clear).delete()
# bulk_create and bulk_update are not possible with current Morango:
db_users = db_new_users + db_update_users
for user in db_users:
user.save()
# assign roles to users:
users_data = {u.username: u for u in db_users}
self.add_roles(users_data, roles)
db_created_classes = []
for classroom in db_new_classes:
created_class = Classroom.objects.create(
name=classroom.name, parent=classroom.parent)
db_created_classes.append(created_class)
# hack to get ids created by Morango:
db_new_classes = db_created_classes
self.add_classes_memberships(
classes, users_data, db_new_classes + db_update_classes)
self.remove_memberships(keeping_users, classes[0], classes[1])
classes_report = {
"created": len(db_new_classes),
"updated": len(db_update_classes),
"cleared": len(classes_to_clear),
}
users_report = {
"created": len(db_new_users),
"updated": len(db_update_users),
"deleted": len(users_to_delete),
}
self.output_messages(
per_line_errors,
classes_report,
users_report,
filepath,
options["errorlines"],
)
translation.deactivate()
def handle(self, *args, **options):
self.job = get_current_job()
return self.handle_async(*args, **options)
def _transfer( # noqa: max-complexity=16
self,
method,
channel_id,
path=None,
drive_id=None,
node_ids=None,
exclude_node_ids=None,
baseurl=None,
peer_id=None,
renderable_only=True,
import_updates=False,
):
try:
if not import_updates:
nodes_for_transfer = get_nodes_to_transfer(
channel_id,
node_ids,
exclude_node_ids,
False,
renderable_only=renderable_only,
drive_id=drive_id,
peer_id=peer_id,
)
total_resource_count = (nodes_for_transfer.exclude(
kind=content_kinds.TOPIC).values(
"content_id").distinct().count())
(
files_to_download,
total_bytes_to_transfer,
) = calculate_files_to_transfer(nodes_for_transfer, False)
else:
(
total_resource_count,
files_to_download,
total_bytes_to_transfer,
) = get_import_data_for_update(
channel_id,
renderable_only=renderable_only,
drive_id=drive_id,
peer_id=peer_id,
)
except LocationError:
if drive_id:
raise CommandError(
"The external drive with given drive id {} does not exist."
.format(drive_id))
if peer_id:
raise CommandError(
"The network location with the id {} does not exist".
format(peer_id))
except ValueError:
if import_updates:
raise CommandError(
"Tried to perform an channel update import when update data was not available"
)
raise
job = get_current_job()
if job:
job.extra_metadata["file_size"] = total_bytes_to_transfer
job.extra_metadata["total_resources"] = total_resource_count
job.save_meta()
number_of_skipped_files = 0
transferred_file_size = 0
file_checksums_to_annotate = []
public = None
# If we're downloading, check listing status
if method == DOWNLOAD_METHOD:
public = lookup_channel_listing_status(channel_id=channel_id,
baseurl=baseurl)
resources_before_transfer = (ContentNode.objects.filter(
channel_id=channel_id,
available=True).exclude(kind=content_kinds.TOPIC).values(
"content_id").distinct().count())
dummy_bytes_for_annotation = annotation.calculate_dummy_progress_for_annotation(
node_ids, exclude_node_ids, total_bytes_to_transfer)
with self.start_progress(
total=total_bytes_to_transfer +
dummy_bytes_for_annotation) as overall_progress_update:
if method == DOWNLOAD_METHOD:
session = requests.Session()
file_transfers = []
for f in files_to_download:
if self.is_cancelled():
break
filename = f.get_filename()
try:
dest = paths.get_content_storage_file_path(filename)
except InvalidStorageFilenameError:
# If the destination file name is malformed, just stop now.
overall_progress_update(f.file_size)
continue
# if the file already exists, add its size to our overall progress, and skip
if os.path.isfile(dest) and os.path.getsize(
dest) == f.file_size:
overall_progress_update(f.file_size)
file_checksums_to_annotate.append(f.id)
transferred_file_size += f.file_size
continue
# determine where we're downloading/copying from, and create appropriate transfer object
if method == DOWNLOAD_METHOD:
url = paths.get_content_storage_remote_url(filename,
baseurl=baseurl)
filetransfer = transfer.FileDownload(
url,
dest,
session=session,
cancel_check=self.is_cancelled)
file_transfers.append((f, filetransfer))
elif method == COPY_METHOD:
try:
srcpath = paths.get_content_storage_file_path(
filename, datafolder=path)
except InvalidStorageFilenameError:
# If the source file name is malformed, just stop now.
overall_progress_update(f.file_size)
continue
filetransfer = transfer.FileCopy(
srcpath, dest, cancel_check=self.is_cancelled)
file_transfers.append((f, filetransfer))
with concurrent.futures.ThreadPoolExecutor(
max_workers=10) as executor:
batch_size = 100
# ThreadPoolExecutor allows us to download files concurrently,
# greatly reducing download time in most cases. However, loading
# all the downloads into the pool requires considerable memory,
# so we divide the downloads into batches to keep memory usage down.
# In batches of 100, total RAM usage doesn't exceed 250MB in testing.
while len(file_transfers) > 0:
future_file_transfers = {}
for i in range(batch_size):
if len(file_transfers) > 0:
f, filetransfer = file_transfers.pop()
future = executor.submit(
self._start_file_transfer,
f,
filetransfer,
overall_progress_update,
)
future_file_transfers[future] = (f, filetransfer)
for future in concurrent.futures.as_completed(
future_file_transfers):
f, filetransfer = future_file_transfers[future]
try:
status = future.result()
if self.is_cancelled():
break
if status == FILE_SKIPPED:
number_of_skipped_files += 1
else:
file_checksums_to_annotate.append(f.id)
transferred_file_size += f.file_size
except transfer.TransferCanceled:
break
except Exception as e:
logger.error(
"An error occurred during content import: {}".
format(e))
if (isinstance(e, requests.exceptions.HTTPError)
and e.response.status_code
== 404) or (isinstance(e, OSError)
and e.errno == 2):
# Continue file import when the current file is not found from the source and is skipped.
overall_progress_update(f.file_size)
number_of_skipped_files += 1
continue
else:
self.exception = e
break
with db_task_write_lock:
annotation.set_content_visibility(
channel_id,
file_checksums_to_annotate,
node_ids=node_ids,
exclude_node_ids=exclude_node_ids,
public=public,
)
resources_after_transfer = (ContentNode.objects.filter(
channel_id=channel_id,
available=True).exclude(kind=content_kinds.TOPIC).values(
"content_id").distinct().count())
if job:
job.extra_metadata[
"transferred_file_size"] = transferred_file_size
job.extra_metadata["transferred_resources"] = (
resources_after_transfer - resources_before_transfer)
job.save_meta()
if number_of_skipped_files > 0:
logger.warning(
"{} files are skipped, because errors occurred during the import."
.format(number_of_skipped_files))
overall_progress_update(dummy_bytes_for_annotation)
if self.exception:
raise self.exception
if self.is_cancelled():
self.cancel()
def make_job_updates(flag):
job = get_current_job()
for i in range(3):
job.update_progress(i, 2)
set_flag(flag)
def _transfer( # noqa: max-complexity=16
self,
method,
channel_id,
path=None,
drive_id=None,
node_ids=None,
exclude_node_ids=None,
baseurl=None,
peer_id=None,
renderable_only=True,
):
try:
nodes_for_transfer = get_nodes_to_transfer(
channel_id,
node_ids,
exclude_node_ids,
False,
renderable_only=renderable_only,
drive_id=drive_id,
peer_id=peer_id,
)
except LocationError:
if drive_id:
raise CommandError(
"The external drive with given drive id {} does not exist."
.format(drive_id))
if peer_id:
raise CommandError(
"The network location with the id {} does not exist".
format(peer_id))
total_resource_count = (nodes_for_transfer.exclude(
kind=content_kinds.TOPIC).values("content_id").distinct().count())
(files_to_download,
total_bytes_to_transfer) = calculate_files_to_transfer(
nodes_for_transfer, False)
job = get_current_job()
if job:
job.extra_metadata["file_size"] = total_bytes_to_transfer
job.extra_metadata["total_resources"] = total_resource_count
job.save_meta()
number_of_skipped_files = 0
transferred_file_size = 0
file_checksums_to_annotate = []
resources_before_transfer = (ContentNode.objects.filter(
channel_id=channel_id,
available=True).exclude(kind=content_kinds.TOPIC).values(
"content_id").distinct().count())
with self.start_progress(
total=total_bytes_to_transfer) as overall_progress_update:
exception = None # Exception that is not caught by the retry logic
if method == DOWNLOAD_METHOD:
session = requests.Session()
for f in files_to_download:
if self.is_cancelled():
break
filename = f.get_filename()
try:
dest = paths.get_content_storage_file_path(filename)
except InvalidStorageFilenameError:
# If the destination file name is malformed, just stop now.
overall_progress_update(f.file_size)
continue
# if the file already exists, add its size to our overall progress, and skip
if os.path.isfile(dest) and os.path.getsize(
dest) == f.file_size:
overall_progress_update(f.file_size)
file_checksums_to_annotate.append(f.id)
transferred_file_size += f.file_size
continue
# determine where we're downloading/copying from, and create appropriate transfer object
if method == DOWNLOAD_METHOD:
url = paths.get_content_storage_remote_url(filename,
baseurl=baseurl)
filetransfer = transfer.FileDownload(url,
dest,
session=session)
elif method == COPY_METHOD:
try:
srcpath = paths.get_content_storage_file_path(
filename, datafolder=path)
except InvalidStorageFilenameError:
# If the source file name is malformed, just stop now.
overall_progress_update(f.file_size)
continue
filetransfer = transfer.FileCopy(srcpath, dest)
finished = False
try:
while not finished:
finished, status = self._start_file_transfer(
f, filetransfer, overall_progress_update)
if self.is_cancelled():
break
if status == FILE_TRANSFERRED:
file_checksums_to_annotate.append(f.id)
transferred_file_size += f.file_size
elif status == FILE_SKIPPED:
number_of_skipped_files += 1
except Exception as e:
exception = e
break
with db_task_write_lock:
annotation.set_content_visibility(
channel_id,
file_checksums_to_annotate,
node_ids=node_ids,
exclude_node_ids=exclude_node_ids,
)
resources_after_transfer = (ContentNode.objects.filter(
channel_id=channel_id,
available=True).exclude(kind=content_kinds.TOPIC).values(
"content_id").distinct().count())
if job:
job.extra_metadata[
"transferred_file_size"] = transferred_file_size
job.extra_metadata["transferred_resources"] = (
resources_after_transfer - resources_before_transfer)
job.save_meta()
if number_of_skipped_files > 0:
logger.warning(
"{} files are skipped, because errors occurred during the import."
.format(number_of_skipped_files))
if exception:
raise exception
if self.is_cancelled():
self.cancel()
def diff_stats(channel_id, method, drive_id=None, baseurl=None):
"""
Download the channel database to an upgraded path.
Annotate the local file availability of the upgraded channel db.
Calculate diff stats comparing default db and annotated channel db.
"""
# upgraded content database path
source_path = paths.get_upgrade_content_database_file_path(channel_id)
# annotated db to be used for calculating diff stats
destination_path = paths.get_annotated_content_database_file_path(channel_id)
try:
if method == "network":
call_command(
"importchannel", "network", channel_id, baseurl=baseurl, no_upgrade=True
)
elif method == "disk":
drive = get_mounted_drive_by_id(drive_id)
call_command(
"importchannel", "disk", channel_id, drive.datafolder, no_upgrade=True
)
# create all fields/tables at the annotated destination db, based on the current schema version
bridge = Bridge(
sqlite_file_path=destination_path, schema_version=CURRENT_SCHEMA_VERSION
)
bridge.Base.metadata.create_all(bridge.engine)
# initialize import manager based on annotated destination path, pulling from source db path
import_manager = channel_import.initialize_import_manager(
channel_id,
cancel_check=False,
source=source_path,
destination=destination_path,
)
# import channel data from source db path
import_manager.import_channel_data()
import_manager.end()
# annotate file availability on destination db
annotation.set_local_file_availability_from_disk(destination=destination_path)
# get the diff count between whats on the default db and the annotated db
new_resources_count = count_new_resources_available_for_import(
destination_path, channel_id
)
# get the count for leaf nodes which are in the default db, but not in the annotated db
resources_to_be_deleted_count = count_removed_resources(
destination_path, channel_id
)
# get the ids of leaf nodes which are now incomplete due to missing local files
updated_resources_ids = automatically_updated_resource_ids(
destination_path, channel_id
)
# remove the annotated database
try:
os.remove(destination_path)
except OSError as e:
logger.info(
"Tried to remove {}, but exception {} occurred.".format(
destination_path, e
)
)
# annotate job metadata with diff stats
job = get_current_job()
if job:
job.extra_metadata["new_resources_count"] = new_resources_count
job.extra_metadata[
"deleted_resources_count"
] = resources_to_be_deleted_count
job.extra_metadata["updated_node_ids"] = updated_resources_ids
job.save_meta()
except UserCancelledError:
# remove the annotated database
try:
os.remove(destination_path)
except OSError:
pass
raise
