1 # Copyright (C) The Arvados Authors. All rights reserved.
3 # SPDX-License-Identifier: Apache-2.0
5 from __future__ import division
6 from future.utils import listitems, listvalues
7 from builtins import str
8 from builtins import object
11 import arvados.collection
33 from apiclient import errors as apiclient_errors
34 from arvados._version import __version__
35 from arvados.util import keep_locator_pattern
37 import arvados.commands._util as arv_cmd
41 upload_opts = argparse.ArgumentParser(add_help=False)
43 upload_opts.add_argument('--version', action='version',
44 version="%s %s" % (sys.argv[0], __version__),
45 help='Print version and exit.')
46 upload_opts.add_argument('paths', metavar='path', type=str, nargs='*',
48 Local file or directory. If path is a directory reference with a trailing
49 slash, then just upload the directory's contents; otherwise upload the
50 directory itself. Default: read from standard input.
53 _group = upload_opts.add_mutually_exclusive_group()
55 _group.add_argument('--max-manifest-depth', type=int, metavar='N',
56 default=-1, help=argparse.SUPPRESS)
58 _group.add_argument('--normalize', action='store_true',
60 Normalize the manifest by re-ordering files and streams after writing
64 _group.add_argument('--dry-run', action='store_true', default=False,
66 Don't actually upload files, but only check if any file should be
67 uploaded. Exit with code=2 when files are pending for upload.
70 _group = upload_opts.add_mutually_exclusive_group()
72 _group.add_argument('--as-stream', action='store_true', dest='stream',
77 _group.add_argument('--stream', action='store_true',
79 Store the file content and display the resulting manifest on
80 stdout. Do not write the manifest to Keep or save a Collection object
84 _group.add_argument('--as-manifest', action='store_true', dest='manifest',
86 Synonym for --manifest.
89 _group.add_argument('--in-manifest', action='store_true', dest='manifest',
91 Synonym for --manifest.
94 _group.add_argument('--manifest', action='store_true',
96 Store the file data and resulting manifest in Keep, save a Collection
97 object in Arvados, and display the manifest locator (Collection uuid)
98 on stdout. This is the default behavior.
101 _group.add_argument('--as-raw', action='store_true', dest='raw',
106 _group.add_argument('--raw', action='store_true',
108 Store the file content and display the data block locators on stdout,
109 separated by commas, with a trailing newline. Do not store a
113 upload_opts.add_argument('--update-collection', type=str, default=None,
114 dest='update_collection', metavar="UUID", help="""
115 Update an existing collection identified by the given Arvados collection
116 UUID. All new local files will be uploaded.
119 upload_opts.add_argument('--use-filename', type=str, default=None,
120 dest='filename', help="""
121 Synonym for --filename.
124 upload_opts.add_argument('--filename', type=str, default=None,
126 Use the given filename in the manifest, instead of the name of the
127 local file. This is useful when "-" or "/dev/stdin" is given as an
128 input file. It can be used only if there is exactly one path given and
129 it is not a directory. Implies --manifest.
132 upload_opts.add_argument('--portable-data-hash', action='store_true',
134 Print the portable data hash instead of the Arvados UUID for the collection
135 created by the upload.
138 upload_opts.add_argument('--replication', type=int, metavar='N', default=None,
140 Set the replication level for the new collection: how many different
141 physical storage devices (e.g., disks) should have a copy of each data
142 block. Default is to use the server-provided default (if any) or 2.
145 upload_opts.add_argument('--storage-classes', help="""
146 Specify comma separated list of storage classes to be used when saving data to Keep.
149 upload_opts.add_argument('--threads', type=int, metavar='N', default=None,
151 Set the number of upload threads to be used. Take into account that
152 using lots of threads will increase the RAM requirements. Default is
154 On high latency installations, using a greater number will improve
158 upload_opts.add_argument('--exclude', metavar='PATTERN', default=[],
159 action='append', help="""
160 Exclude files and directories whose names match the given glob pattern. When
161 using a path-like pattern like 'subdir/*.txt', all text files inside 'subdir'
162 directory, relative to the provided input dirs will be excluded.
163 When using a filename pattern like '*.txt', any text file will be excluded
164 no matter where it is placed.
165 For the special case of needing to exclude only files or dirs directly below
166 the given input directory, you can use a pattern like './exclude_this.gif'.
167 You can specify multiple patterns by using this argument more than once.
170 _group = upload_opts.add_mutually_exclusive_group()
171 _group.add_argument('--follow-links', action='store_true', default=True,
172 dest='follow_links', help="""
173 Follow file and directory symlinks (default).
175 _group.add_argument('--no-follow-links', action='store_false', dest='follow_links',
177 Do not follow file and directory symlinks.
181 run_opts = argparse.ArgumentParser(add_help=False)
183 run_opts.add_argument('--project-uuid', metavar='UUID', help="""
184 Store the collection in the specified project, instead of your Home
188 run_opts.add_argument('--name', help="""
189 Save the collection with the specified name.
192 _group = run_opts.add_mutually_exclusive_group()
193 _group.add_argument('--progress', action='store_true',
195 Display human-readable progress on stderr (bytes and, if possible,
196 percentage of total data size). This is the default behavior when
200 _group.add_argument('--no-progress', action='store_true',
202 Do not display human-readable progress on stderr, even if stderr is a
206 _group.add_argument('--batch-progress', action='store_true',
208 Display machine-readable progress on stderr (bytes and, if known,
212 run_opts.add_argument('--silent', action='store_true',
214 Do not print any debug messages to console. (Any error messages will
218 _group = run_opts.add_mutually_exclusive_group()
219 _group.add_argument('--resume', action='store_true', default=True,
221 Continue interrupted uploads from cached state (default).
223 _group.add_argument('--no-resume', action='store_false', dest='resume',
225 Do not continue interrupted uploads from cached state.
228 _group = run_opts.add_mutually_exclusive_group()
229 _group.add_argument('--cache', action='store_true', dest='use_cache', default=True,
231 Save upload state in a cache file for resuming (default).
233 _group.add_argument('--no-cache', action='store_false', dest='use_cache',
235 Do not save upload state in a cache file for resuming.
238 _group = upload_opts.add_mutually_exclusive_group()
239 _group.add_argument('--trash-at', metavar='YYYY-MM-DDTHH:MM', default=None,
241 Set the trash date of the resulting collection to an absolute date in the future.
242 The accepted format is defined by the ISO 8601 standard. Examples: 20090103, 2009-01-03, 20090103T181505, 2009-01-03T18:15:05.\n
243 Timezone information can be added. If not, the provided date/time is assumed as being in the local system's timezone.
245 _group.add_argument('--trash-after', type=int, metavar='DAYS', default=None,
247 Set the trash date of the resulting collection to an amount of days from the
248 date/time that the upload process finishes.
251 arg_parser = argparse.ArgumentParser(
252 description='Copy data from the local filesystem to Keep.',
253 parents=[upload_opts, run_opts, arv_cmd.retry_opt])
255 def parse_arguments(arguments):
256 args = arg_parser.parse_args(arguments)
258 if len(args.paths) == 0:
261 args.paths = ["-" if x == "/dev/stdin" else x for x in args.paths]
263 if len(args.paths) != 1 or os.path.isdir(args.paths[0]):
266 --filename argument cannot be used when storing a directory or
270 # Turn on --progress by default if stderr is a tty.
271 if (not (args.batch_progress or args.no_progress or args.silent)
272 and os.isatty(sys.stderr.fileno())):
275 # Turn off --resume (default) if --no-cache is used.
276 if not args.use_cache:
279 if args.paths == ['-']:
280 if args.update_collection:
282 --update-collection cannot be used when reading from stdin.
285 args.use_cache = False
286 if not args.filename:
287 args.filename = 'stdin'
289 # Remove possible duplicated patterns
290 if len(args.exclude) > 0:
291 args.exclude = list(set(args.exclude))
296 class PathDoesNotExistError(Exception):
300 class CollectionUpdateError(Exception):
304 class ResumeCacheConflict(Exception):
308 class ResumeCacheInvalidError(Exception):
311 class ArvPutArgumentConflict(Exception):
315 class ArvPutUploadIsPending(Exception):
319 class ArvPutUploadNotPending(Exception):
323 class FileUploadList(list):
324 def __init__(self, dry_run=False):
326 self.dry_run = dry_run
328 def append(self, other):
330 raise ArvPutUploadIsPending()
331 super(FileUploadList, self).append(other)
334 # Appends the X-Request-Id to the log message when log level is ERROR or DEBUG
335 class ArvPutLogFormatter(logging.Formatter):
336 std_fmtr = logging.Formatter(arvados.log_format, arvados.log_date_format)
338 request_id_informed = False
340 def __init__(self, request_id):
341 self.err_fmtr = logging.Formatter(
342 arvados.log_format+' (X-Request-Id: {})'.format(request_id),
343 arvados.log_date_format)
345 def format(self, record):
346 if (not self.request_id_informed) and (record.levelno in (logging.DEBUG, logging.ERROR)):
347 self.request_id_informed = True
348 return self.err_fmtr.format(record)
349 return self.std_fmtr.format(record)
352 class ResumeCache(object):
353 CACHE_DIR = '.cache/arvados/arv-put'
355 def __init__(self, file_spec):
356 self.cache_file = open(file_spec, 'a+')
357 self._lock_file(self.cache_file)
358 self.filename = self.cache_file.name
361 def make_path(cls, args):
363 md5.update(arvados.config.get('ARVADOS_API_HOST', '!nohost').encode())
364 realpaths = sorted(os.path.realpath(path) for path in args.paths)
365 md5.update(b'\0'.join([p.encode() for p in realpaths]))
366 if any(os.path.isdir(path) for path in realpaths):
369 md5.update(args.filename.encode())
371 arv_cmd.make_home_conf_dir(cls.CACHE_DIR, 0o700, 'raise'),
374 def _lock_file(self, fileobj):
376 fcntl.flock(fileobj, fcntl.LOCK_EX | fcntl.LOCK_NB)
378 raise ResumeCacheConflict(u"{} locked".format(fileobj.name))
381 self.cache_file.seek(0)
382 return json.load(self.cache_file)
384 def check_cache(self, api_client=None, num_retries=0):
389 if "_finished_streams" in state and len(state["_finished_streams"]) > 0:
390 locator = state["_finished_streams"][0][1][0]
391 elif "_current_stream_locators" in state and len(state["_current_stream_locators"]) > 0:
392 locator = state["_current_stream_locators"][0]
393 if locator is not None:
394 kc = arvados.keep.KeepClient(api_client=api_client)
395 kc.head(locator, num_retries=num_retries)
396 except Exception as e:
401 def save(self, data):
403 new_cache_fd, new_cache_name = tempfile.mkstemp(
404 dir=os.path.dirname(self.filename))
405 self._lock_file(new_cache_fd)
406 new_cache = os.fdopen(new_cache_fd, 'r+')
407 json.dump(data, new_cache)
408 os.rename(new_cache_name, self.filename)
409 except (IOError, OSError, ResumeCacheConflict):
411 os.unlink(new_cache_name)
412 except NameError: # mkstemp failed.
415 self.cache_file.close()
416 self.cache_file = new_cache
419 self.cache_file.close()
423 os.unlink(self.filename)
424 except OSError as error:
425 if error.errno != errno.ENOENT: # That's what we wanted anyway.
431 self.__init__(self.filename)
434 class ArvPutUploadJob(object):
435 CACHE_DIR = '.cache/arvados/arv-put'
437 'manifest' : None, # Last saved manifest checkpoint
438 'files' : {} # Previous run file list: {path : {size, mtime}}
441 def __init__(self, paths, resume=True, use_cache=True, reporter=None,
442 name=None, owner_uuid=None, api_client=None,
443 ensure_unique_name=False, num_retries=None,
444 put_threads=None, replication_desired=None, filename=None,
445 update_time=60.0, update_collection=None, storage_classes=None,
446 logger=logging.getLogger('arvados.arv_put'), dry_run=False,
447 follow_links=True, exclude_paths=[], exclude_names=None,
451 self.use_cache = use_cache
453 self.reporter = reporter
454 # This will set to 0 before start counting, if no special files are going
456 self.bytes_expected = None
457 self.bytes_written = 0
458 self.bytes_skipped = 0
460 self.owner_uuid = owner_uuid
461 self.ensure_unique_name = ensure_unique_name
462 self.num_retries = num_retries
463 self.replication_desired = replication_desired
464 self.put_threads = put_threads
465 self.filename = filename
466 self.storage_classes = storage_classes
467 self._api_client = api_client
468 self._state_lock = threading.Lock()
469 self._state = None # Previous run state (file list & manifest)
470 self._current_files = [] # Current run file list
471 self._cache_file = None
472 self._collection_lock = threading.Lock()
473 self._remote_collection = None # Collection being updated (if asked)
474 self._local_collection = None # Collection from previous run manifest
475 self._file_paths = set() # Files to be updated in remote collection
476 self._stop_checkpointer = threading.Event()
477 self._checkpointer = threading.Thread(target=self._update_task)
478 self._checkpointer.daemon = True
479 self._update_task_time = update_time # How many seconds wait between update runs
480 self._files_to_upload = FileUploadList(dry_run=dry_run)
481 self._upload_started = False
483 self.dry_run = dry_run
484 self._checkpoint_before_quit = True
485 self.follow_links = follow_links
486 self.exclude_paths = exclude_paths
487 self.exclude_names = exclude_names
488 self._trash_at = trash_at
490 if self._trash_at is not None:
491 if type(self._trash_at) not in [datetime.datetime, datetime.timedelta]:
492 raise TypeError('trash_at should be None, timezone-naive datetime or timedelta')
493 if type(self._trash_at) == datetime.datetime and self._trash_at.tzinfo is not None:
494 raise TypeError('provided trash_at datetime should be timezone-naive')
496 if not self.use_cache and self.resume:
497 raise ArvPutArgumentConflict('resume cannot be True when use_cache is False')
499 # Check for obvious dry-run responses
500 if self.dry_run and (not self.use_cache or not self.resume):
501 raise ArvPutUploadIsPending()
503 # Load cached data if any and if needed
504 self._setup_state(update_collection)
506 # Build the upload file list, excluding requested files and counting the
507 # bytes expected to be uploaded.
508 self._build_upload_list()
510 def _build_upload_list(self):
512 Scan the requested paths to count file sizes, excluding requested files
513 and dirs and building the upload file list.
515 # If there aren't special files to be read, reset total bytes count to zero
517 if not any([p for p in self.paths
518 if not (os.path.isfile(p) or os.path.isdir(p))]):
519 self.bytes_expected = 0
521 for path in self.paths:
522 # Test for stdin first, in case some file named '-' exist
525 raise ArvPutUploadIsPending()
526 self._write_stdin(self.filename or 'stdin')
527 elif not os.path.exists(path):
528 raise PathDoesNotExistError(u"file or directory '{}' does not exist.".format(path))
529 elif os.path.isdir(path):
530 # Use absolute paths on cache index so CWD doesn't interfere
531 # with the caching logic.
533 path = os.path.abspath(path)
534 if orig_path[-1:] == os.sep:
535 # When passing a directory reference with a trailing slash,
536 # its contents should be uploaded directly to the
540 # When passing a directory reference with no trailing slash,
541 # upload the directory to the collection's root.
542 prefixdir = os.path.dirname(path)
544 for root, dirs, files in os.walk(path,
545 followlinks=self.follow_links):
546 root_relpath = os.path.relpath(root, path)
547 if root_relpath == '.':
549 # Exclude files/dirs by full path matching pattern
550 if self.exclude_paths:
551 dirs[:] = [d for d in dirs
552 if not any(pathname_match(
553 os.path.join(root_relpath, d), pat)
554 for pat in self.exclude_paths)]
555 files = [f for f in files
556 if not any(pathname_match(
557 os.path.join(root_relpath, f), pat)
558 for pat in self.exclude_paths)]
559 # Exclude files/dirs by name matching pattern
560 if self.exclude_names is not None:
561 dirs[:] = [d for d in dirs
562 if not self.exclude_names.match(d)]
563 files = [f for f in files
564 if not self.exclude_names.match(f)]
565 # Make os.walk()'s dir traversing order deterministic
569 filepath = os.path.join(root, f)
570 # Add its size to the total bytes count (if applicable)
571 if self.follow_links or (not os.path.islink(filepath)):
572 if self.bytes_expected is not None:
573 self.bytes_expected += os.path.getsize(filepath)
574 self._check_file(filepath,
575 os.path.join(root[len(prefixdir):], f))
577 filepath = os.path.abspath(path)
578 # Add its size to the total bytes count (if applicable)
579 if self.follow_links or (not os.path.islink(filepath)):
580 if self.bytes_expected is not None:
581 self.bytes_expected += os.path.getsize(filepath)
582 self._check_file(filepath,
583 self.filename or os.path.basename(path))
584 # If dry-mode is on, and got up to this point, then we should notify that
585 # there aren't any file to upload.
587 raise ArvPutUploadNotPending()
588 # Remove local_collection's files that don't exist locally anymore, so the
589 # bytes_written count is correct.
590 for f in self.collection_file_paths(self._local_collection,
592 if f != 'stdin' and f != self.filename and not f in self._file_paths:
593 self._local_collection.remove(f)
595 def start(self, save_collection):
597 Start supporting thread & file uploading
599 self._checkpointer.start()
601 # Update bytes_written from current local collection and
602 # report initial progress.
605 self._upload_started = True # Used by the update thread to start checkpointing
607 except (SystemExit, Exception) as e:
608 self._checkpoint_before_quit = False
609 # Log stack trace only when Ctrl-C isn't pressed (SIGINT)
610 # Note: We're expecting SystemExit instead of
611 # KeyboardInterrupt because we have a custom signal
612 # handler in place that raises SystemExit with the catched
614 if isinstance(e, PathDoesNotExistError):
615 # We aren't interested in the traceback for this case
617 elif not isinstance(e, SystemExit) or e.code != -2:
618 self.logger.warning("Abnormal termination:\n{}".format(
619 traceback.format_exc()))
623 # Stop the thread before doing anything else
624 self._stop_checkpointer.set()
625 self._checkpointer.join()
626 if self._checkpoint_before_quit:
627 # Commit all pending blocks & one last _update()
628 self._local_collection.manifest_text()
629 self._update(final=True)
631 self.save_collection()
633 self._cache_file.close()
635 def _collection_trash_at(self):
637 Returns the trash date that the collection should use at save time.
638 Takes into account absolute/relative trash_at values requested
641 if type(self._trash_at) == datetime.timedelta:
642 # Get an absolute datetime for trash_at
643 return datetime.datetime.utcnow() + self._trash_at
644 return self._trash_at
646 def save_collection(self):
648 # Check if files should be updated on the remote collection.
649 for fp in self._file_paths:
650 remote_file = self._remote_collection.find(fp)
652 # File don't exist on remote collection, copy it.
653 self._remote_collection.copy(fp, fp, self._local_collection)
654 elif remote_file != self._local_collection.find(fp):
655 # A different file exist on remote collection, overwrite it.
656 self._remote_collection.copy(fp, fp, self._local_collection, overwrite=True)
658 # The file already exist on remote collection, skip it.
660 self._remote_collection.save(storage_classes=self.storage_classes,
661 num_retries=self.num_retries,
662 trash_at=self._collection_trash_at())
664 if self.storage_classes is None:
665 self.storage_classes = ['default']
666 self._local_collection.save_new(
667 name=self.name, owner_uuid=self.owner_uuid,
668 storage_classes=self.storage_classes,
669 ensure_unique_name=self.ensure_unique_name,
670 num_retries=self.num_retries,
671 trash_at=self._collection_trash_at())
673 def destroy_cache(self):
676 os.unlink(self._cache_filename)
677 except OSError as error:
678 # That's what we wanted anyway.
679 if error.errno != errno.ENOENT:
681 self._cache_file.close()
683 def _collection_size(self, collection):
685 Recursively get the total size of the collection
688 for item in listvalues(collection):
689 if isinstance(item, arvados.collection.Collection) or isinstance(item, arvados.collection.Subcollection):
690 size += self._collection_size(item)
695 def _update_task(self):
697 Periodically called support task. File uploading is
698 asynchronous so we poll status from the collection.
700 while not self._stop_checkpointer.wait(1 if not self._upload_started else self._update_task_time):
703 def _update(self, final=False):
705 Update cached manifest text and report progress.
707 if self._upload_started:
708 with self._collection_lock:
709 self.bytes_written = self._collection_size(self._local_collection)
712 manifest = self._local_collection.manifest_text()
714 # Get the manifest text without comitting pending blocks
715 manifest = self._local_collection.manifest_text(strip=False,
719 with self._state_lock:
720 self._state['manifest'] = manifest
724 except Exception as e:
725 self.logger.error("Unexpected error trying to save cache file: {}".format(e))
726 # Keep remote collection's trash_at attribute synced when using relative expire dates
727 if self._remote_collection is not None and type(self._trash_at) == datetime.timedelta:
729 self._api_client.collections().update(
730 uuid=self._remote_collection.manifest_locator(),
731 body={'trash_at': self._collection_trash_at().strftime("%Y-%m-%dT%H:%M:%S.%fZ")}
732 ).execute(num_retries=self.num_retries)
733 except Exception as e:
734 self.logger.error("Unexpected error trying to update remote collection's expire date: {}".format(e))
736 self.bytes_written = self.bytes_skipped
737 # Call the reporter, if any
738 self.report_progress()
740 def report_progress(self):
741 if self.reporter is not None:
742 self.reporter(self.bytes_written, self.bytes_expected)
744 def _write_stdin(self, filename):
745 output = self._local_collection.open(filename, 'wb')
746 self._write(sys.stdin, output)
749 def _check_file(self, source, filename):
751 Check if this file needs to be uploaded
753 # Ignore symlinks when requested
754 if (not self.follow_links) and os.path.islink(source):
757 should_upload = False
758 new_file_in_cache = False
759 # Record file path for updating the remote collection before exiting
760 self._file_paths.add(filename)
762 with self._state_lock:
763 # If no previous cached data on this file, store it for an eventual
765 if source not in self._state['files']:
766 self._state['files'][source] = {
767 'mtime': os.path.getmtime(source),
768 'size' : os.path.getsize(source)
770 new_file_in_cache = True
771 cached_file_data = self._state['files'][source]
773 # Check if file was already uploaded (at least partially)
774 file_in_local_collection = self._local_collection.find(filename)
776 # If not resuming, upload the full file.
779 # New file detected from last run, upload it.
780 elif new_file_in_cache:
782 # Local file didn't change from last run.
783 elif cached_file_data['mtime'] == os.path.getmtime(source) and cached_file_data['size'] == os.path.getsize(source):
784 if not file_in_local_collection:
785 # File not uploaded yet, upload it completely
787 elif file_in_local_collection.permission_expired():
788 # Permission token expired, re-upload file. This will change whenever
789 # we have a API for refreshing tokens.
790 self.logger.warning(u"Uploaded file '{}' access token expired, will re-upload it from scratch".format(filename))
792 self._local_collection.remove(filename)
793 elif cached_file_data['size'] == file_in_local_collection.size():
794 # File already there, skip it.
795 self.bytes_skipped += cached_file_data['size']
796 elif cached_file_data['size'] > file_in_local_collection.size():
797 # File partially uploaded, resume!
798 resume_offset = file_in_local_collection.size()
799 self.bytes_skipped += resume_offset
802 # Inconsistent cache, re-upload the file
804 self._local_collection.remove(filename)
805 self.logger.warning(u"Uploaded version of file '{}' is bigger than local version, will re-upload it from scratch.".format(source))
806 # Local file differs from cached data, re-upload it.
808 if file_in_local_collection:
809 self._local_collection.remove(filename)
814 self._files_to_upload.append((source, resume_offset, filename))
815 except ArvPutUploadIsPending:
816 # This could happen when running on dry-mode, close cache file to
817 # avoid locking issues.
818 self._cache_file.close()
821 def _upload_files(self):
822 for source, resume_offset, filename in self._files_to_upload:
823 with open(source, 'rb') as source_fd:
824 with self._state_lock:
825 self._state['files'][source]['mtime'] = os.path.getmtime(source)
826 self._state['files'][source]['size'] = os.path.getsize(source)
827 if resume_offset > 0:
828 # Start upload where we left off
829 output = self._local_collection.open(filename, 'ab')
830 source_fd.seek(resume_offset)
833 output = self._local_collection.open(filename, 'wb')
834 self._write(source_fd, output)
835 output.close(flush=False)
837 def _write(self, source_fd, output):
839 data = source_fd.read(arvados.config.KEEP_BLOCK_SIZE)
844 def _my_collection(self):
845 return self._remote_collection if self.update else self._local_collection
847 def _get_cache_filepath(self):
848 # Set up cache file name from input paths.
850 md5.update(arvados.config.get('ARVADOS_API_HOST', '!nohost').encode())
851 realpaths = sorted(os.path.realpath(path) for path in self.paths)
852 md5.update(b'\0'.join([p.encode() for p in realpaths]))
854 md5.update(self.filename.encode())
855 cache_filename = md5.hexdigest()
856 cache_filepath = os.path.join(
857 arv_cmd.make_home_conf_dir(self.CACHE_DIR, 0o700, 'raise'),
859 return cache_filepath
861 def _setup_state(self, update_collection):
863 Create a new cache file or load a previously existing one.
865 # Load an already existing collection for update
866 if update_collection and re.match(arvados.util.collection_uuid_pattern,
869 self._remote_collection = arvados.collection.Collection(
870 update_collection, api_client=self._api_client)
871 except arvados.errors.ApiError as error:
872 raise CollectionUpdateError("Cannot read collection {} ({})".format(update_collection, error))
875 elif update_collection:
876 # Collection locator provided, but unknown format
877 raise CollectionUpdateError("Collection locator unknown: '{}'".format(update_collection))
880 cache_filepath = self._get_cache_filepath()
881 if self.resume and os.path.exists(cache_filepath):
882 self.logger.info(u"Resuming upload from cache file {}".format(cache_filepath))
883 self._cache_file = open(cache_filepath, 'a+')
885 # --no-resume means start with a empty cache file.
886 self.logger.info(u"Creating new cache file at {}".format(cache_filepath))
887 self._cache_file = open(cache_filepath, 'w+')
888 self._cache_filename = self._cache_file.name
889 self._lock_file(self._cache_file)
890 self._cache_file.seek(0)
892 with self._state_lock:
895 self._state = json.load(self._cache_file)
896 if not set(['manifest', 'files']).issubset(set(self._state.keys())):
897 # Cache at least partially incomplete, set up new cache
898 self._state = copy.deepcopy(self.EMPTY_STATE)
900 # Cache file empty, set up new cache
901 self._state = copy.deepcopy(self.EMPTY_STATE)
903 self.logger.info("No cache usage requested for this run.")
904 # No cache file, set empty state
905 self._state = copy.deepcopy(self.EMPTY_STATE)
906 if not self._cached_manifest_valid():
907 raise ResumeCacheInvalidError()
908 # Load the previous manifest so we can check if files were modified remotely.
909 self._local_collection = arvados.collection.Collection(
910 self._state['manifest'],
911 replication_desired=self.replication_desired,
912 put_threads=self.put_threads,
913 api_client=self._api_client)
915 def _cached_manifest_valid(self):
917 Validate the oldest non-expired block signature to check if cached manifest
918 is usable: checking if the cached manifest was not created with a different
921 if self._state.get('manifest', None) is None:
922 # No cached manifest yet, all good.
924 now = datetime.datetime.utcnow()
928 for m in keep_locator_pattern.finditer(self._state['manifest']):
931 exp = datetime.datetime.utcfromtimestamp(int(loc.split('@')[1], 16))
933 # Locator without signature
936 if exp > now and (oldest_exp is None or exp < oldest_exp):
940 # No block signatures found => no invalid block signatures.
942 if oldest_loc is None:
943 # Locator signatures found, but all have expired.
944 # Reset the cache and move on.
945 self.logger.info('Cache expired, starting from scratch.')
946 self._state['manifest'] = ''
948 kc = arvados.KeepClient(api_client=self._api_client,
949 num_retries=self.num_retries)
952 except arvados.errors.KeepRequestError:
953 # Something is wrong, cached manifest is not valid.
957 def collection_file_paths(self, col, path_prefix='.'):
958 """Return a list of file paths by recursively go through the entire collection `col`"""
960 for name, item in listitems(col):
961 if isinstance(item, arvados.arvfile.ArvadosFile):
962 file_paths.append(os.path.join(path_prefix, name))
963 elif isinstance(item, arvados.collection.Subcollection):
964 new_prefix = os.path.join(path_prefix, name)
965 file_paths += self.collection_file_paths(item, path_prefix=new_prefix)
968 def _lock_file(self, fileobj):
970 fcntl.flock(fileobj, fcntl.LOCK_EX | fcntl.LOCK_NB)
972 raise ResumeCacheConflict(u"{} locked".format(fileobj.name))
974 def _save_state(self):
976 Atomically save current state into cache.
978 with self._state_lock:
979 # We're not using copy.deepcopy() here because it's a lot slower
980 # than json.dumps(), and we're already needing JSON format to be
982 state = json.dumps(self._state)
984 new_cache = tempfile.NamedTemporaryFile(
986 dir=os.path.dirname(self._cache_filename), delete=False)
987 self._lock_file(new_cache)
988 new_cache.write(state)
991 os.rename(new_cache.name, self._cache_filename)
992 except (IOError, OSError, ResumeCacheConflict) as error:
993 self.logger.error("There was a problem while saving the cache file: {}".format(error))
995 os.unlink(new_cache_name)
996 except NameError: # mkstemp failed.
999 self._cache_file.close()
1000 self._cache_file = new_cache
1002 def collection_name(self):
1003 return self._my_collection().api_response()['name'] if self._my_collection().api_response() else None
1005 def collection_trash_at(self):
1006 return self._my_collection().get_trash_at()
1008 def manifest_locator(self):
1009 return self._my_collection().manifest_locator()
1011 def portable_data_hash(self):
1012 pdh = self._my_collection().portable_data_hash()
1013 m = self._my_collection().stripped_manifest().encode()
1014 local_pdh = '{}+{}'.format(hashlib.md5(m).hexdigest(), len(m))
1015 if pdh != local_pdh:
1016 self.logger.warning("\n".join([
1017 "arv-put: API server provided PDH differs from local manifest.",
1018 " This should not happen; showing API server version."]))
1021 def manifest_text(self, stream_name=".", strip=False, normalize=False):
1022 return self._my_collection().manifest_text(stream_name, strip, normalize)
1024 def _datablocks_on_item(self, item):
1026 Return a list of datablock locators, recursively navigating
1027 through subcollections
1029 if isinstance(item, arvados.arvfile.ArvadosFile):
1030 if item.size() == 0:
1031 # Empty file locator
1032 return ["d41d8cd98f00b204e9800998ecf8427e+0"]
1035 for segment in item.segments():
1036 loc = segment.locator
1037 locators.append(loc)
1039 elif isinstance(item, arvados.collection.Collection):
1040 l = [self._datablocks_on_item(x) for x in listvalues(item)]
1041 # Fast list flattener method taken from:
1042 # http://stackoverflow.com/questions/952914/making-a-flat-list-out-of-list-of-lists-in-python
1043 return [loc for sublist in l for loc in sublist]
1047 def data_locators(self):
1048 with self._collection_lock:
1049 # Make sure all datablocks are flushed before getting the locators
1050 self._my_collection().manifest_text()
1051 datablocks = self._datablocks_on_item(self._my_collection())
1054 _machine_format = "{} {}: {{}} written {{}} total\n".format(sys.argv[0],
1057 # Simulate glob.glob() matching behavior without the need to scan the filesystem
1058 # Note: fnmatch() doesn't work correctly when used with pathnames. For example the
1059 # pattern 'tests/*.py' will match 'tests/run_test.py' and also 'tests/subdir/run_test.py',
1060 # so instead we're using it on every path component.
1061 def pathname_match(pathname, pattern):
1062 name = pathname.split(os.sep)
1063 # Fix patterns like 'some/subdir/' or 'some//subdir'
1064 pat = [x for x in pattern.split(os.sep) if x != '' and x != '.']
1065 if len(name) != len(pat):
1067 for i in range(len(name)):
1068 if not fnmatch.fnmatch(name[i], pat[i]):
1072 def machine_progress(bytes_written, bytes_expected):
1073 return _machine_format.format(
1074 bytes_written, -1 if (bytes_expected is None) else bytes_expected)
1076 def human_progress(bytes_written, bytes_expected):
1078 return "\r{}M / {}M {:.1%} ".format(
1079 bytes_written >> 20, bytes_expected >> 20,
1080 float(bytes_written) / bytes_expected)
1082 return "\r{} ".format(bytes_written)
1084 def progress_writer(progress_func, outfile=sys.stderr):
1085 def write_progress(bytes_written, bytes_expected):
1086 outfile.write(progress_func(bytes_written, bytes_expected))
1087 return write_progress
1089 def desired_project_uuid(api_client, project_uuid, num_retries):
1090 if not project_uuid:
1091 query = api_client.users().current()
1092 elif arvados.util.user_uuid_pattern.match(project_uuid):
1093 query = api_client.users().get(uuid=project_uuid)
1094 elif arvados.util.group_uuid_pattern.match(project_uuid):
1095 query = api_client.groups().get(uuid=project_uuid)
1097 raise ValueError("Not a valid project UUID: {}".format(project_uuid))
1098 return query.execute(num_retries=num_retries)['uuid']
1100 def main(arguments=None, stdout=sys.stdout, stderr=sys.stderr,
1101 install_sig_handlers=True):
1104 args = parse_arguments(arguments)
1105 logger = logging.getLogger('arvados.arv_put')
1107 logger.setLevel(logging.WARNING)
1109 logger.setLevel(logging.INFO)
1112 request_id = arvados.util.new_request_id()
1114 formatter = ArvPutLogFormatter(request_id)
1115 logging.getLogger('arvados').handlers[0].setFormatter(formatter)
1117 if api_client is None:
1118 api_client = arvados.api('v1', request_id=request_id)
1120 if install_sig_handlers:
1121 arv_cmd.install_signal_handlers()
1123 # Trash arguments validation
1125 if args.trash_at is not None:
1126 # ciso8601 considers YYYYMM as invalid but YYYY-MM as valid, so here we
1127 # make sure the user provides a complete YYYY-MM-DD date.
1128 if not re.match(r'^\d{4}(?P<dash>-?)\d{2}?(?P=dash)\d{2}', args.trash_at):
1129 logger.error("--trash-at argument format invalid, use --help to see examples.")
1131 # Check if no time information was provided. In that case, assume end-of-day.
1132 if re.match(r'^\d{4}(?P<dash>-?)\d{2}?(?P=dash)\d{2}$', args.trash_at):
1133 args.trash_at += 'T23:59:59'
1135 trash_at = ciso8601.parse_datetime(args.trash_at)
1137 logger.error("--trash-at argument format invalid, use --help to see examples.")
1140 if trash_at.tzinfo is not None:
1141 # Timezone aware datetime provided.
1142 utcoffset = -trash_at.utcoffset()
1144 # Timezone naive datetime provided. Assume is local.
1146 utcoffset = datetime.timedelta(seconds=time.altzone)
1148 utcoffset = datetime.timedelta(seconds=time.timezone)
1149 # Convert to UTC timezone naive datetime.
1150 trash_at = trash_at.replace(tzinfo=None) + utcoffset
1152 if trash_at <= datetime.datetime.utcnow():
1153 logger.error("--trash-at argument must be set in the future")
1155 if args.trash_after is not None:
1156 if args.trash_after < 1:
1157 logger.error("--trash-after argument must be >= 1")
1159 trash_at = datetime.timedelta(seconds=(args.trash_after * 24 * 60 * 60))
1161 # Determine the name to use
1163 if args.stream or args.raw:
1164 logger.error("Cannot use --name with --stream or --raw")
1166 elif args.update_collection:
1167 logger.error("Cannot use --name with --update-collection")
1169 collection_name = args.name
1171 collection_name = "Saved at {} by {}@{}".format(
1172 datetime.datetime.utcnow().strftime("%Y-%m-%d %H:%M:%S UTC"),
1173 pwd.getpwuid(os.getuid()).pw_name,
1174 socket.gethostname())
1176 if args.project_uuid and (args.stream or args.raw):
1177 logger.error("Cannot use --project-uuid with --stream or --raw")
1180 # Determine the parent project
1182 project_uuid = desired_project_uuid(api_client, args.project_uuid,
1184 except (apiclient_errors.Error, ValueError) as error:
1189 reporter = progress_writer(human_progress)
1190 elif args.batch_progress:
1191 reporter = progress_writer(machine_progress)
1195 # Split storage-classes argument
1196 storage_classes = None
1197 if args.storage_classes:
1198 storage_classes = args.storage_classes.strip().split(',')
1199 if len(storage_classes) > 1:
1200 logger.error("Multiple storage classes are not supported currently.")
1204 # Setup exclude regex from all the --exclude arguments provided
1207 exclude_names = None
1208 if len(args.exclude) > 0:
1209 # We're supporting 2 kinds of exclusion patterns:
1210 # 1) --exclude '*.jpg' (file/dir name patterns, will only match
1211 # the name, wherever the file is on the tree)
1212 # 2.1) --exclude 'foo/bar' (file/dir path patterns, will match the
1213 # entire path, and should be relative to
1214 # any input dir argument)
1215 # 2.2) --exclude './*.jpg' (Special case for excluding files/dirs
1216 # placed directly underneath the input dir)
1217 for p in args.exclude:
1218 # Only relative paths patterns allowed
1219 if p.startswith(os.sep):
1220 logger.error("Cannot use absolute paths with --exclude")
1222 if os.path.dirname(p):
1223 # We don't support of path patterns with '..'
1224 p_parts = p.split(os.sep)
1227 "Cannot use path patterns that include or '..'")
1229 # Path search pattern
1230 exclude_paths.append(p)
1232 # Name-only search pattern
1233 name_patterns.append(p)
1234 # For name only matching, we can combine all patterns into a single
1235 # regexp, for better performance.
1236 exclude_names = re.compile('|'.join(
1237 [fnmatch.translate(p) for p in name_patterns]
1238 )) if len(name_patterns) > 0 else None
1239 # Show the user the patterns to be used, just in case they weren't
1240 # specified inside quotes and got changed by the shell expansion.
1241 logger.info("Exclude patterns: {}".format(args.exclude))
1243 # If this is used by a human, and there's at least one directory to be
1244 # uploaded, the expected bytes calculation can take a moment.
1245 if args.progress and any([os.path.isdir(f) for f in args.paths]):
1246 logger.info("Calculating upload size, this could take some time...")
1248 writer = ArvPutUploadJob(paths = args.paths,
1249 resume = args.resume,
1250 use_cache = args.use_cache,
1251 filename = args.filename,
1252 reporter = reporter,
1253 api_client = api_client,
1254 num_retries = args.retries,
1255 replication_desired = args.replication,
1256 put_threads = args.threads,
1257 name = collection_name,
1258 owner_uuid = project_uuid,
1259 ensure_unique_name = True,
1260 update_collection = args.update_collection,
1261 storage_classes=storage_classes,
1263 dry_run=args.dry_run,
1264 follow_links=args.follow_links,
1265 exclude_paths=exclude_paths,
1266 exclude_names=exclude_names,
1268 except ResumeCacheConflict:
1269 logger.error("\n".join([
1270 "arv-put: Another process is already uploading this data.",
1271 " Use --no-cache if this is really what you want."]))
1273 except ResumeCacheInvalidError:
1274 logger.error("\n".join([
1275 "arv-put: Resume cache contains invalid signature: it may have expired",
1276 " or been created with another Arvados user's credentials.",
1277 " Switch user or use one of the following options to restart upload:",
1278 " --no-resume to start a new resume cache.",
1279 " --no-cache to disable resume cache."]))
1281 except (CollectionUpdateError, PathDoesNotExistError) as error:
1282 logger.error("\n".join([
1283 "arv-put: %s" % str(error)]))
1285 except ArvPutUploadIsPending:
1286 # Dry run check successful, return proper exit code.
1288 except ArvPutUploadNotPending:
1289 # No files pending for upload
1292 if not args.dry_run and not args.update_collection and args.resume and writer.bytes_written > 0:
1293 logger.warning("\n".join([
1294 "arv-put: Resuming previous upload from last checkpoint.",
1295 " Use the --no-resume option to start over."]))
1297 if not args.dry_run:
1298 writer.report_progress()
1301 writer.start(save_collection=not(args.stream or args.raw))
1302 except arvados.errors.ApiError as error:
1303 logger.error("\n".join([
1304 "arv-put: %s" % str(error)]))
1307 if args.progress: # Print newline to split stderr from stdout for humans.
1312 output = writer.manifest_text(normalize=True)
1314 output = writer.manifest_text()
1316 output = ','.join(writer.data_locators())
1319 expiration_notice = ""
1320 if writer.collection_trash_at() is not None:
1321 # Get the local timezone-naive version, and log it with timezone information.
1323 local_trash_at = writer.collection_trash_at().replace(tzinfo=None) - datetime.timedelta(seconds=time.altzone)
1325 local_trash_at = writer.collection_trash_at().replace(tzinfo=None) - datetime.timedelta(seconds=time.timezone)
1326 expiration_notice = ". It will expire on {} {}.".format(
1327 local_trash_at.strftime("%Y-%m-%d %H:%M:%S"), time.strftime("%z"))
1328 if args.update_collection:
1329 logger.info(u"Collection updated: '{}'{}".format(
1330 writer.collection_name(), expiration_notice))
1332 logger.info(u"Collection saved as '{}'{}".format(
1333 writer.collection_name(), expiration_notice))
1334 if args.portable_data_hash:
1335 output = writer.portable_data_hash()
1337 output = writer.manifest_locator()
1338 except apiclient_errors.Error as error:
1340 "arv-put: Error creating Collection on project: {}.".format(
1344 # Print the locator (uuid) of the new collection.
1346 status = status or 1
1347 elif not args.silent:
1348 stdout.write(output)
1349 if not output.endswith('\n'):
1352 if install_sig_handlers:
1353 arv_cmd.restore_signal_handlers()
1362 if __name__ == '__main__':