21 from collections import deque
30 _logger = logging.getLogger('arvados.collection')
32 def normalize_stream(s, stream):
34 sortedfiles = list(stream.keys())
41 if b[arvados.LOCATOR] not in blocks:
42 stream_tokens.append(b[arvados.LOCATOR])
43 blocks[b[arvados.LOCATOR]] = streamoffset
44 streamoffset += b[arvados.BLOCKSIZE]
46 if len(stream_tokens) == 1:
47 stream_tokens.append(config.EMPTY_BLOCK_LOCATOR)
51 fout = f.replace(' ', '\\040')
52 for segment in stream[f]:
53 segmentoffset = blocks[segment[arvados.LOCATOR]] + segment[arvados.OFFSET]
54 if current_span == None:
55 current_span = [segmentoffset, segmentoffset + segment[arvados.SEGMENTSIZE]]
57 if segmentoffset == current_span[1]:
58 current_span[1] += segment[arvados.SEGMENTSIZE]
60 stream_tokens.append("{0}:{1}:{2}".format(current_span[0], current_span[1] - current_span[0], fout))
61 current_span = [segmentoffset, segmentoffset + segment[arvados.SEGMENTSIZE]]
63 if current_span != None:
64 stream_tokens.append("{0}:{1}:{2}".format(current_span[0], current_span[1] - current_span[0], fout))
66 if len(stream[f]) == 0:
67 stream_tokens.append("0:0:{0}".format(fout))
71 def normalize(collection):
73 for s in collection.all_streams():
74 for f in s.all_files():
75 filestream = s.name() + "/" + f.name()
76 r = filestream.rindex("/")
77 streamname = filestream[:r]
78 filename = filestream[r+1:]
79 if streamname not in streams:
80 streams[streamname] = {}
81 if filename not in streams[streamname]:
82 streams[streamname][filename] = []
84 streams[streamname][filename].extend(s.locators_and_ranges(r[0], r[1]))
86 normalized_streams = []
87 sortedstreams = list(streams.keys())
89 for s in sortedstreams:
90 normalized_streams.append(normalize_stream(s, streams[s]))
91 return normalized_streams
94 class CollectionReader(object):
95 def __init__(self, manifest_locator_or_text):
96 if re.match(r'[a-f0-9]{32}(\+\d+)?(\+\S+)*$', manifest_locator_or_text):
97 self._manifest_locator = manifest_locator_or_text
98 self._manifest_text = None
99 elif re.match(r'(\S+)( [a-f0-9]{32}(\+\d+)(\+\S+)*)+( \d+:\d+:\S+)+\n', manifest_locator_or_text):
100 self._manifest_text = manifest_locator_or_text
101 self._manifest_locator = None
103 raise errors.ArgumentError(
104 "Argument to CollectionReader must be a manifest or a collection UUID")
114 if self._streams != None:
116 if not self._manifest_text:
118 c = arvados.api('v1').collections().get(
119 uuid=self._manifest_locator).execute()
120 self._manifest_text = c['manifest_text']
121 except Exception as e:
122 _logger.warning("API lookup failed for collection %s (%s: %s)",
123 self._manifest_locator, type(e), str(e))
124 self._manifest_text = Keep.get(self._manifest_locator)
126 for stream_line in self._manifest_text.split("\n"):
127 if stream_line != '':
128 stream_tokens = stream_line.split()
129 self._streams += [stream_tokens]
130 self._streams = normalize(self)
132 # now regenerate the manifest text based on the normalized stream
134 #print "normalizing", self._manifest_text
135 self._manifest_text = ''.join([StreamReader(stream).manifest_text() for stream in self._streams])
136 #print "result", self._manifest_text
139 def all_streams(self):
142 for s in self._streams:
143 resp.append(StreamReader(s))
147 for s in self.all_streams():
148 for f in s.all_files():
151 def manifest_text(self, strip=False):
154 m = ''.join([StreamReader(stream).manifest_text(strip=True) for stream in self._streams])
157 return self._manifest_text
159 class CollectionWriter(object):
160 KEEP_BLOCK_SIZE = 2**26
163 self._data_buffer = []
164 self._data_buffer_len = 0
165 self._current_stream_files = []
166 self._current_stream_length = 0
167 self._current_stream_locators = []
168 self._current_stream_name = '.'
169 self._current_file_name = None
170 self._current_file_pos = 0
171 self._finished_streams = []
172 self._close_file = None
173 self._queued_file = None
174 self._queued_dirents = deque()
175 self._queued_trees = deque()
183 def do_queued_work(self):
184 # The work queue consists of three pieces:
185 # * _queued_file: The file object we're currently writing to the
187 # * _queued_dirents: Entries under the current directory
188 # (_queued_trees[0]) that we want to write or recurse through.
189 # This may contain files from subdirectories if
190 # max_manifest_depth == 0 for this directory.
191 # * _queued_trees: Directories that should be written as separate
192 # streams to the Collection.
193 # This function handles the smallest piece of work currently queued
194 # (current file, then current directory, then next directory) until
195 # no work remains. The _work_THING methods each do a unit of work on
196 # THING. _queue_THING methods add a THING to the work queue.
198 if self._queued_file:
200 elif self._queued_dirents:
202 elif self._queued_trees:
207 def _work_file(self):
209 buf = self._queued_file.read(self.KEEP_BLOCK_SIZE)
213 self.finish_current_file()
215 self._queued_file.close()
216 self._close_file = None
217 self._queued_file = None
219 def _work_dirents(self):
220 path, stream_name, max_manifest_depth = self._queued_trees[0]
221 if stream_name != self.current_stream_name():
222 self.start_new_stream(stream_name)
223 while self._queued_dirents:
224 dirent = self._queued_dirents.popleft()
225 target = os.path.join(path, dirent)
226 if os.path.isdir(target):
227 self._queue_tree(target,
228 os.path.join(stream_name, dirent),
229 max_manifest_depth - 1)
231 self._queue_file(target, dirent)
233 if not self._queued_dirents:
234 self._queued_trees.popleft()
236 def _work_trees(self):
237 path, stream_name, max_manifest_depth = self._queued_trees[0]
238 make_dirents = (util.listdir_recursive if (max_manifest_depth == 0)
240 d = make_dirents(path)
242 self._queue_dirents(stream_name, d)
244 self._queued_trees.popleft()
246 def _queue_file(self, source, filename=None):
247 assert (self._queued_file is None), "tried to queue more than one file"
248 if not hasattr(source, 'read'):
249 source = open(source, 'rb')
250 self._close_file = True
252 self._close_file = False
254 filename = os.path.basename(source.name)
255 self.start_new_file(filename)
256 self._queued_file = source
258 def _queue_dirents(self, stream_name, dirents):
259 assert (not self._queued_dirents), "tried to queue more than one tree"
260 self._queued_dirents = deque(sorted(dirents))
262 def _queue_tree(self, path, stream_name, max_manifest_depth):
263 self._queued_trees.append((path, stream_name, max_manifest_depth))
265 def write_file(self, source, filename=None):
266 self._queue_file(source, filename)
267 self.do_queued_work()
269 def write_directory_tree(self,
270 path, stream_name='.', max_manifest_depth=-1):
271 self._queue_tree(path, stream_name, max_manifest_depth)
272 self.do_queued_work()
274 def write(self, newdata):
275 if hasattr(newdata, '__iter__'):
279 self._data_buffer += [newdata]
280 self._data_buffer_len += len(newdata)
281 self._current_stream_length += len(newdata)
282 while self._data_buffer_len >= self.KEEP_BLOCK_SIZE:
285 def flush_data(self):
286 data_buffer = ''.join(self._data_buffer)
287 if data_buffer != '':
288 self._current_stream_locators += [Keep.put(data_buffer[0:self.KEEP_BLOCK_SIZE])]
289 self._data_buffer = [data_buffer[self.KEEP_BLOCK_SIZE:]]
290 self._data_buffer_len = len(self._data_buffer[0])
292 def start_new_file(self, newfilename=None):
293 self.finish_current_file()
294 self.set_current_file_name(newfilename)
296 def set_current_file_name(self, newfilename):
297 if re.search(r'[\t\n]', newfilename):
298 raise errors.AssertionError(
299 "Manifest filenames cannot contain whitespace: %s" %
301 self._current_file_name = newfilename
303 def current_file_name(self):
304 return self._current_file_name
306 def finish_current_file(self):
307 if self._current_file_name == None:
308 if self._current_file_pos == self._current_stream_length:
310 raise errors.AssertionError(
311 "Cannot finish an unnamed file " +
312 "(%d bytes at offset %d in '%s' stream)" %
313 (self._current_stream_length - self._current_file_pos,
314 self._current_file_pos,
315 self._current_stream_name))
316 self._current_stream_files += [[self._current_file_pos,
317 self._current_stream_length - self._current_file_pos,
318 self._current_file_name]]
319 self._current_file_pos = self._current_stream_length
321 def start_new_stream(self, newstreamname='.'):
322 self.finish_current_stream()
323 self.set_current_stream_name(newstreamname)
325 def set_current_stream_name(self, newstreamname):
326 if re.search(r'[\t\n]', newstreamname):
327 raise errors.AssertionError(
328 "Manifest stream names cannot contain whitespace")
329 self._current_stream_name = '.' if newstreamname=='' else newstreamname
331 def current_stream_name(self):
332 return self._current_stream_name
334 def finish_current_stream(self):
335 self.finish_current_file()
337 if len(self._current_stream_files) == 0:
339 elif self._current_stream_name == None:
340 raise errors.AssertionError(
341 "Cannot finish an unnamed stream (%d bytes in %d files)" %
342 (self._current_stream_length, len(self._current_stream_files)))
344 if len(self._current_stream_locators) == 0:
345 self._current_stream_locators += [config.EMPTY_BLOCK_LOCATOR]
346 self._finished_streams += [[self._current_stream_name,
347 self._current_stream_locators,
348 self._current_stream_files]]
349 self._current_stream_files = []
350 self._current_stream_length = 0
351 self._current_stream_locators = []
352 self._current_stream_name = None
353 self._current_file_pos = 0
354 self._current_file_name = None
357 # Store the manifest in Keep and return its locator.
358 return Keep.put(self.manifest_text())
360 def stripped_manifest(self):
362 Return the manifest for the current collection with all permission
363 hints removed from the locators in the manifest.
365 raw = self.manifest_text()
367 for line in raw.split("\n"):
368 fields = line.split()
370 locators = [ re.sub(r'\+A[a-z0-9@_-]+', '', x)
371 for x in fields[1:-1] ]
372 clean += fields[0] + ' ' + ' '.join(locators) + ' ' + fields[-1] + "\n"
375 def manifest_text(self):
376 self.finish_current_stream()
379 for stream in self._finished_streams:
380 if not re.search(r'^\.(/.*)?$', stream[0]):
382 manifest += stream[0].replace(' ', '\\040')
383 manifest += ' ' + ' '.join(stream[1])
384 manifest += ' ' + ' '.join("%d:%d:%s" % (sfile[0], sfile[1], sfile[2].replace(' ', '\\040')) for sfile in stream[2])
387 if len(manifest) > 0:
388 return CollectionReader(manifest).manifest_text()
392 def data_locators(self):
394 for name, locators, files in self._finished_streams:
399 class ResumableCollectionWriter(CollectionWriter):
400 STATE_PROPS = ['_current_stream_files', '_current_stream_length',
401 '_current_stream_locators', '_current_stream_name',
402 '_current_file_name', '_current_file_pos', '_close_file',
403 '_data_buffer', '_dependencies', '_finished_streams',
404 '_queued_dirents', '_queued_trees']
407 self._dependencies = {}
408 super(ResumableCollectionWriter, self).__init__()
411 def from_state(cls, state, *init_args, **init_kwargs):
412 # Try to build a new writer from scratch with the given state.
413 # If the state is not suitable to resume (because files have changed,
414 # been deleted, aren't predictable, etc.), raise a
415 # StaleWriterStateError. Otherwise, return the initialized writer.
416 # The caller is responsible for calling writer.do_queued_work()
417 # appropriately after it's returned.
418 writer = cls(*init_args, **init_kwargs)
419 for attr_name in cls.STATE_PROPS:
420 attr_value = state[attr_name]
421 attr_class = getattr(writer, attr_name).__class__
422 # Coerce the value into the same type as the initial value, if
424 if attr_class not in (type(None), attr_value.__class__):
425 attr_value = attr_class(attr_value)
426 setattr(writer, attr_name, attr_value)
427 # Check dependencies before we try to resume anything.
428 if any(KeepLocator(ls).permission_expired()
429 for ls in writer._current_stream_locators):
430 raise errors.StaleWriterStateError(
431 "locators include expired permission hint")
432 writer.check_dependencies()
433 if state['_current_file'] is not None:
434 path, pos = state['_current_file']
436 writer._queued_file = open(path, 'rb')
437 writer._queued_file.seek(pos)
438 except IOError as error:
439 raise errors.StaleWriterStateError(
440 "failed to reopen active file {}: {}".format(path, error))
443 def check_dependencies(self):
444 for path, orig_stat in self._dependencies.items():
445 if not S_ISREG(orig_stat[ST_MODE]):
446 raise errors.StaleWriterStateError("{} not file".format(path))
448 now_stat = tuple(os.stat(path))
449 except OSError as error:
450 raise errors.StaleWriterStateError(
451 "failed to stat {}: {}".format(path, error))
452 if ((not S_ISREG(now_stat[ST_MODE])) or
453 (orig_stat[ST_MTIME] != now_stat[ST_MTIME]) or
454 (orig_stat[ST_SIZE] != now_stat[ST_SIZE])):
455 raise errors.StaleWriterStateError("{} changed".format(path))
457 def dump_state(self, copy_func=lambda x: x):
458 state = {attr: copy_func(getattr(self, attr))
459 for attr in self.STATE_PROPS}
460 if self._queued_file is None:
461 state['_current_file'] = None
463 state['_current_file'] = (os.path.realpath(self._queued_file.name),
464 self._queued_file.tell())
467 def _queue_file(self, source, filename=None):
469 src_path = os.path.realpath(source)
471 raise errors.AssertionError("{} not a file path".format(source))
473 path_stat = os.stat(src_path)
474 except OSError as stat_error:
476 super(ResumableCollectionWriter, self)._queue_file(source, filename)
477 fd_stat = os.fstat(self._queued_file.fileno())
478 if not S_ISREG(fd_stat.st_mode):
479 # We won't be able to resume from this cache anyway, so don't
480 # worry about further checks.
481 self._dependencies[source] = tuple(fd_stat)
482 elif path_stat is None:
483 raise errors.AssertionError(
484 "could not stat {}: {}".format(source, stat_error))
485 elif path_stat.st_ino != fd_stat.st_ino:
486 raise errors.AssertionError(
487 "{} changed between open and stat calls".format(source))
489 self._dependencies[src_path] = tuple(fd_stat)
491 def write(self, data):
492 if self._queued_file is None:
493 raise errors.AssertionError(
494 "resumable writer can't accept unsourced data")
495 return super(ResumableCollectionWriter, self).write(data)