def classic_iterator(urls,
reader=task_io.chain_reader,
input_stream=(func.map_input_stream, ),
notifier=func.notifier,
params=None,
ddfs=None):
"""
An iterator over records as seen by the classic map interface.
:type reader: :func:`disco.worker.task_io.input_stream`
:param reader: shortcut for the last input stream applied.
:type input_stream: sequence of :func:`disco.worker.task_io.input_stream`
:param input_stream: used to read from a custom file format.
:type notifier: :func:`disco.func.notifier`
:param notifier: called when the task opens a url.
"""
from disco.worker import Input
from disco.worker.classic.worker import Worker
worker = Worker(map_reader=reader, map_input_stream=input_stream)
settings = DiscoSettings(DISCO_MASTER=ddfs) if ddfs else DiscoSettings()
for input in util.inputlist(urls, settings=settings):
if isinstance(input, basestring):
dest = proxy_url(input, to_master=False)
elif isinstance(input, tuple):
dest = tuple([proxy_url(i, to_master=False) for i in input])
else:
dest = [proxy_url(i, to_master=False) for i in input]
notifier(dest)
for record in Input(dest, open=worker.opener('map', 'in', params)):
yield record
def classic_iterator(urls,
reader=task_io.chain_reader,
input_stream=(func.map_input_stream, ),
notifier=func.notifier,
params=None,
ddfs=None):
"""
An iterator over records as seen by the classic map interface.
:type reader: :func:`disco.worker.task_io.input_stream`
:param reader: shortcut for the last input stream applied.
:type input_stream: sequence of :func:`disco.worker.task_io.input_stream`
:param input_stream: used to read from a custom file format.
:type notifier: :func:`disco.func.notifier`
:param notifier: called when the task opens a url.
"""
from disco.worker import Input
from disco.worker.classic.worker import Worker
worker = Worker(map_reader=reader, map_input_stream=input_stream)
settings = DiscoSettings(DISCO_MASTER=ddfs) if ddfs else DiscoSettings()
for input in util.inputlist(urls, settings=settings):
if isinstance(input, basestring):
dest = proxy_url(input, to_master=False)
elif isinstance(input, tuple):
dest = tuple([proxy_url(i, to_master=False) for i in input])
else:
dest = [proxy_url(i, to_master=False) for i in input]
notifier(dest)
for record in Input(dest, open=worker.opener('map', 'in', params)):
yield record
def _download(self, url, data=None, token=None, method='GET'):
return json.loads(download(self._resolve(proxy_url(url,
proxy=self.proxy,
meth=method)),
data=data,
method=method,
token=self._token(url, token, method)))
def _upload(self, urls, source, token=None, to_master=True, **kwargs):
urls = [self._resolve(proxy_url(url,
proxy=self.proxy,
meth='PUT',
to_master=to_master))
for url in iterify(urls)]
return upload(urls, source, token=self._token(url, token, 'PUT'), **kwargs)
def curl(replicas):
for replica in replicas:
try:
return download(proxy_url(urlresolve(replica, master=program.ddfs.master),
to_master=False))
except Exception, e:
sys.stderr.write("%s\n" % e)
def curl(replicas):
for replica in replicas:
try:
return download(proxy_url(urlresolve(replica, master=program.ddfs.master),
to_master=False))
except Exception, e:
sys.stderr.write("%s\n" % e)
def result_iterator(results, notifier = None,\
proxy = None, reader = func.netstr_reader):
res = []
for dir_url in results:
if dir_url.startswith("dir://"):
res += util.parse_dir(dir_url, proxy)
else:
res.append(dir_url)
x, x, root = util.load_conf()
for url in res:
if url.startswith("file://"):
fname = url[7:]
fd = file(fname)
sze = os.stat(fname).st_size
elif url.startswith("disco://"):
host, fname = url[8:].split("/", 1)
url = util.proxy_url(proxy, fname, host)
if util.resultfs_enabled:
f = "%s/data/%s" % (root, fname)
fd = file(f)
sze = os.stat(f).st_size
else:
sze, fd = comm.open_remote(url)
else:
raise JobException("Invalid result url: %s" % url)
if notifier:
notifier(url)
for x in reader(fd, sze, fname):
yield x
def _download(self, url, data=None, token=None, method="GET", to_master=True):
byts = download(
self._resolve(proxy_url(url, proxy=self.proxy, meth=method, to_master=to_master)),
data=data,
method=method,
token=self._token(url, token, method),
)
return json.loads(bytes_to_str(byts))
def _download(self, url, data=None, token=None, method='GET', to_master=True):
return json.loads(download(self._resolve(proxy_url(url,
proxy=self.proxy,
meth=method,
to_master=to_master)),
data=data,
method=method,
token=self._token(url, token, method)))
def request(self, url, data=None, offset=0):
try:
return download(proxy_url('%s%s' % (self.master, url), proxy=self.proxy),
data=data,
offset=offset)
except CommError, e:
if e.code == None:
e.msg += " (is disco master running at %s?)" % self.master
raise
def request(self, url, data=None, offset=0):
try:
return download(proxy_url('%s%s' % (self.master, url),
proxy=self.proxy),
data=data,
offset=offset)
except CommError, e:
if e.code == None:
e.msg += " (is disco master running at %s?)" % self.master
raise
def curl(replicas):
for replica in replicas:
try:
return download(proxy_url(urlresolve(replica, master=program.ddfs.master),
to_master=False))
except Exception as e:
sys.stderr.write("{0}\n".format(e))
if not ignore_missing:
raise Exception("Failed downloading all replicas: {0}".format(replicas))
return ''
def curl(replicas):
for replica in replicas:
try:
return download(proxy_url(urlresolve(replica, master=program.ddfs.master),
to_master=False))
except Exception as e:
sys.stderr.write("{0}\n".format(e))
if not ignore_missing:
raise Exception("Failed downloading all replicas: {0}".format(replicas))
return ''
def request(self, url, data=None, offset=0):
try:
byts = download(proxy_url('{0}{1}'.format(self.master, url), proxy=self.proxy),
data=data,
offset=offset)
return byts.decode('utf-8')
except CommError as e:
if e.code == None:
e.msg += " (is disco master running at {0}?)".format(self.master)
raise
def pull(self, tag, blobfilter=lambda x: True, token=None):
for repl in self.urls(tag, token=token):
if blobfilter(self.blob_name(repl[0])):
random.shuffle(repl)
for url in repl:
try:
yield open_remote(proxy_url(url,proxy=self.proxy))
break
except CommError, error:
continue
else:
raise error
def result_iterator(urls,
reader=task_io.chain_reader,
input_stream=(func.map_input_stream, ),
notifier=func.notifier,
params=None,
ddfs=None):
"""
An iterator over records stored in either disco or ddfs.
:type reader: :func:`disco.worker.task_io.input_stream`
:param reader: shortcut for the last input stream applied.
:type input_stream: sequence of :func:`disco.worker.task_io.input_stream`
:param input_stream: used to read from a custom file format.
:type notifier: :func:`disco.func.notifier`
:param notifier: called when the task opens a url.
"""
from disco.worker import Input
from disco.worker.task_io import StreamCombiner
settings = DiscoSettings(DISCO_MASTER=ddfs) if ddfs else DiscoSettings()
for input in util.inputlist(urls, settings=settings):
if isinstance(input, basestring):
dest = proxy_url(input, to_master=False)
elif isinstance(input, tuple):
dest = tuple([proxy_url(i, to_master=False) for i in input])
else:
dest = [proxy_url(i, to_master=False) for i in input]
notifier(dest)
def open(url):
streams = [s for s in input_stream]
if reader:
streams += [reader]
return StreamCombiner(url, streams, params)
for record in Input(dest, open=open):
yield record
def result_iterator(urls,
reader=task_io.chain_reader,
input_stream=(func.map_input_stream, ),
notifier=func.notifier,
params=None,
ddfs=None):
"""
An iterator over records stored in either disco or ddfs.
:type reader: :func:`disco.worker.task_io.input_stream`
:param reader: shortcut for the last input stream applied.
:type input_stream: sequence of :func:`disco.worker.task_io.input_stream`
:param input_stream: used to read from a custom file format.
:type notifier: :func:`disco.func.notifier`
:param notifier: called when the task opens a url.
"""
from disco.worker import Input
from disco.worker.task_io import StreamCombiner
settings = DiscoSettings(DISCO_MASTER=ddfs) if ddfs else DiscoSettings()
for input in util.inputlist(urls, settings=settings):
if isinstance(input, basestring):
dest = proxy_url(input, to_master=False)
elif isinstance(input, tuple):
dest = tuple([proxy_url(i, to_master=False) for i in input])
else:
dest = [proxy_url(i, to_master=False) for i in input]
notifier(dest)
def open(url):
streams = [s for s in input_stream]
if reader:
streams += [reader]
return StreamCombiner(url, streams, params)
for record in Input(dest, open=open):
yield record
def request(self, url, data=None, offset=0, as_bytes=False):
try:
byts = download(proxy_url('{0}{1}'.format(self.master, url),
proxy=self.proxy),
data=data,
offset=offset)
if as_bytes:
return byts
return byts.decode('utf-8')
except CommError as e:
if e.code == None:
e.msg += " (is disco master running at {0}?)".format(
self.master)
raise
def pull(self, tag, blobfilter=lambda x: True, token=None):
for repl in self.urls(tag, token=token):
if blobfilter(self.blob_name(repl[0])):
random.shuffle(repl)
for url in repl:
url = self._resolve(
proxy_url(url,
meth='GET',
proxy=self.proxy,
to_master=False))
try:
yield open_remote(url)
break
except CommError, error:
continue
else:
raise error
def pull(self, tag, blobfilter=lambda x: True, token=None):
"""
Iterate over the blobs in a ``tag`` after optionally applying
a ``blobfilter`` over the blob names.
"""
comm_error = None
for repl in self.urls(tag, token=token):
if blobfilter(self.blob_name(repl[0])):
random.shuffle(repl)
for url in repl:
url = self._resolve(proxy_url(url, meth="GET", proxy=self.proxy, to_master=False))
try:
yield open_remote(url)
break
except CommError as error:
comm_error = error
continue
else:
raise comm_error
def pull(self, tag, blobfilter=lambda x: True, token=None):
comm_error = None
for repl in self.urls(tag, token=token):
if blobfilter(self.blob_name(repl[0])):
random.shuffle(repl)
for url in repl:
url = self._resolve(
proxy_url(url,
meth='GET',
proxy=self.proxy,
to_master=False)
)
try:
yield open_remote(url)
break
except CommError as error:
comm_error = error
continue
else:
raise comm_error
def xcat(program, *urls):
"""Usage: [urls ...]
Concatenate the extracted results stored in url[s] and print to stdout.
If any of the url[s] are tags,
the blobs reachable from the tags will be printed after any non-tag url[s].
"""
from itertools import chain
from disco.core import classic_iterator
from disco.util import iterify, reify, urlresolve, proxy_url
tags, urls = program.separate_tags(*program.input(*urls))
stream = reify(program.options.stream)
reader = program.options.reader
reader = reify('disco.func.chain_reader' if reader is None else reader)
bloburls = [[proxy_url(urlresolve(u), to_master=False) for u in repset]
for repset in chain(urls, program.blobs(*tags))]
for record in classic_iterator(bloburls,
input_stream=stream,
reader=reader):
print '\t'.join('%s' % (e, ) for e in iterify(record)).rstrip()
def xcat(program, *urls):
"""Usage: [urls ...]
Concatenate the extracted results stored in url[s] and print to stdout.
If any of the url[s] are tags,
the blobs reachable from the tags will be printed after any non-tag url[s].
"""
from itertools import chain
from disco.core import classic_iterator
from disco.util import iterify, reify, urlresolve, proxy_url
tags, urls = program.separate_tags(*program.input(*urls))
stream = reify(program.options.stream)
reader = program.options.reader
reader = reify('disco.func.chain_reader' if reader is None else reader)
bloburls = [[proxy_url(urlresolve(u), to_master=False) for u in repset]
for repset in chain(urls, program.blobs(*tags))]
for record in classic_iterator(bloburls,
input_stream=stream,
reader=reader):
print '\t'.join('%s' % (e,) for e in iterify(record)).rstrip()
def pull(self, tag, blobfilter=lambda x: True, token=None):
"""
Iterate over the blobs in a ``tag`` after optionally applying
a ``blobfilter`` over the blob names.
"""
comm_error = None
for repl in self.urls(tag, token=token):
if blobfilter(self.blob_name(repl[0])):
random.shuffle(repl)
for url in repl:
url = self._resolve(
proxy_url(url,
meth='GET',
proxy=self.proxy,
to_master=False))
try:
yield open_remote(url)
break
except CommError as error:
comm_error = error
continue
else:
raise comm_error
