def good(self):
base_path = make_random_str()
for path in base_path, base_path + UNI_CHR:
hdfs.dump("foo\n", path)
self.assertTrue(hdfs.path.exists(path))
hdfs.rmr(path)
self.assertFalse(hdfs.path.exists(path))
def dump(self):
for test_path in self.hdfs_paths[0], self.local_paths[0]:
hdfs.dump(self.data, test_path)
with hdfs.open(test_path) as fi:
rdata = fi.read()
fi.fs.close()
self.assertEqual(rdata, self.data)
def __setup_remote_paths(self):
"""
Actually create the working directory and copy the module into it.
Note: the script has to be readable by Hadoop; though this may not
generally be a problem on HDFS, where the Hadoop user is usually
the superuser, things may be different if our working directory is
on a shared POSIX filesystem. Therefore, we make the directory
and the script accessible by all.
"""
self.logger.debug("remote_wd: %s", self.remote_wd)
self.logger.debug("remote_exe: %s", self.remote_exe)
self.logger.debug("remotes: %s", self.files_to_upload)
if self.args.module:
self.logger.debug(
'Generated pipes_code:\n\n %s', self._generate_pipes_code()
)
if not self.args.pretend:
hdfs.mkdir(self.remote_wd)
hdfs.chmod(self.remote_wd, "a+rx")
self.logger.debug("created and chmod-ed: %s", self.remote_wd)
pipes_code = self._generate_pipes_code()
hdfs.dump(pipes_code, self.remote_exe)
self.logger.debug("dumped pipes_code to: %s", self.remote_exe)
hdfs.chmod(self.remote_exe, "a+rx")
self.__warn_user_if_wd_maybe_unreadable(self.remote_wd)
for (l, h, _) in self.files_to_upload:
self.logger.debug("uploading: %s to %s", l, h)
hdfs.cp(l, h)
self.logger.debug("Created%sremote paths:" %
(' [simulation] ' if self.args.pretend else ' '))
def dump(self):
for test_path in self.hdfs_paths[0], self.local_paths[0]:
hdfs.dump(self.data, test_path, mode="wb")
with hdfs.open(test_path) as fi:
rdata = fi.read()
fi.fs.close()
self.assertEqual(rdata, self.data)
def good(self):
base_path = make_random_str()
for path in base_path, base_path + UNI_CHR:
hdfs.dump("foo\n", path)
self.assertTrue(hdfs.path.exists(path))
hdfs.rmr(path)
self.assertFalse(hdfs.path.exists(path))
def write_output(k, args, logger):
logger.debug("kinship matrix: shape=%r" % (k.shape, ))
logger.info("serializing output")
s = KinshipBuilder.serialize(k)
logger.debug("serialized matrix: %d bytes" % len(s))
logger.info("writing output to %r" % (args.output, ))
hdfs.dump(s, args.output, user=args.hdfs_user)
def _write(self, data):
"Internal Write API"
schema = self.schema
wmaid = self.wmaid(data)
year, month, _ = today()
hdir = '%s/%s/%s' % (self.hdir, year, month)
if not hdfs.path.isdir(hdir):
hdfs.mkdir(hdir)
fname = file_name(hdir, wmaid, self.compress)
# create Avro writer and binary encoder
writer = avro.io.DatumWriter(schema)
bytes_writer = io.BytesIO()
if self.compress:
# use gzip'ed writer with BytesIO file object
gzip_writer = gzip.GzipFile(fileobj=bytes_writer, mode='wb')
encoder = avro.io.BinaryEncoder(gzip_writer)
else:
# plain binary reader
encoder = avro.io.BinaryEncoder(bytes_writer)
# write records from given data stream to binary writer
writer.write(data, encoder)
# close gzip stream if necessary
if self.compress:
gzip_writer.flush()
gzip_writer.close()
# store raw data to hadoop via HDFS
hdfs.dump(bytes_writer.getvalue(), fname)
# close bytes stream
bytes_writer.close()
def __setup_remote_paths(self):
"""
Actually create the working directory and copy the module into it.
Note: the script has to be readable by Hadoop; though this may not
generally be a problem on HDFS, where the Hadoop user is usually
the superuser, things may be different if our working directory is
on a shared POSIX filesystem. Therefore, we make the directory
and the script accessible by all.
"""
self.logger.debug("remote_wd: %s", self.remote_wd)
self.logger.debug("remote_exe: %s", self.remote_exe)
self.logger.debug("remotes: %s", self.files_to_upload)
if self.args.module:
self.logger.debug('Generated pipes_code:\n\n %s',
self._generate_pipes_code())
if not self.args.pretend:
hdfs.mkdir(self.remote_wd)
hdfs.chmod(self.remote_wd, "a+rx")
self.logger.debug("created and chmod-ed: %s", self.remote_wd)
pipes_code = self._generate_pipes_code()
hdfs.dump(pipes_code, self.remote_exe)
self.logger.debug("dumped pipes_code to: %s", self.remote_exe)
hdfs.chmod(self.remote_exe, "a+rx")
self.__warn_user_if_wd_maybe_unreadable(self.remote_wd)
for (l, h, _) in self.files_to_upload:
self.logger.debug("uploading: %s to %s", l, h)
hdfs.cp(l, h)
self.logger.debug("Created%sremote paths:" %
(' [simulation] ' if self.args.pretend else ' '))
def _write(self, data):
"Internal Write API"
schema = self.schema
wmaid = self.wmaid(data)
year, month, _ = today()
hdir = '%s/%s/%s' % (self.hdir, year, month)
if not hdfs.path.isdir(hdir):
hdfs.mkdir(hdir)
fname = file_name(hdir, wmaid, self.compress)
# create Avro writer and binary encoder
writer = avro.io.DatumWriter(schema)
bytes_writer = io.BytesIO()
if self.compress:
# use gzip'ed writer with BytesIO file object
gzip_writer = gzip.GzipFile(fileobj=bytes_writer, mode='wb')
encoder = avro.io.BinaryEncoder(gzip_writer)
else:
# plain binary reader
encoder = avro.io.BinaryEncoder(bytes_writer)
# write records from given data stream to binary writer
writer.write(data, encoder)
# close gzip stream if necessary
if self.compress:
gzip_writer.flush()
gzip_writer.close()
# store raw data to hadoop via HDFS
hdfs.dump(bytes_writer.getvalue(), fname)
# close bytes stream
bytes_writer.close()
def write(writeFlag):
if (writeFlag == True):
# instantiate hadoop
hdfs.hdfs()
targetPath = config.targetPath;
targetDirectory = config.targetDirectory;
sourceFile = config.sourceFile;
print("Target Path: " + targetPath);
print("Target Directory: " + targetDirectory);
print("Source Path: " + sourceFile);
dumpFile = open(sourceFile, "r");
fullText = dumpFile.read();
dumpFile.close();
# write to hadoop
#hdfs.mkdir(targetDirectory)
hdfs.dump(fullText, targetPath)
#hdfs.cp(sourceFile, targetPath)
#print (hdfs.ls("test4"))
#files = hdfs.ls("test4")
# read from hadoop
#hdfs.get("test4/hello.txt", "/tmp/hello.txt")
#with open("/tmp/hello.txt") as f:
# print f.read()
#print(hdfs.ls("test", "hduser1"))
#text = hdfs.load("test/hello.txt")
#print text
def get(self):
src = self.hdfs_paths[0]
dest = hdfs.path.split(self.local_paths[0])[-1]
hdfs.dump(self.data, src)
hdfs.get(src, dest)
with open(dest) as fi:
rdata = fi.read()
self.assertEqual(rdata, self.data)
def renames(self):
test_path = self.hdfs_paths[0]
hdfs.dump(self.data, test_path)
new_d = hdfs.path.join(self.hdfs_wd, "new_dir")
new_path = hdfs.path.join(new_d, "new_p")
hdfs.renames(test_path, new_path)
self.assertFalse(hdfs.path.exists(test_path))
self.assertTrue(hdfs.path.exists(new_path))
def renames(self):
test_path = self.hdfs_paths[0]
hdfs.dump(self.data, test_path, mode="wb")
new_d = hdfs.path.join(self.hdfs_wd, "new_dir")
new_path = hdfs.path.join(new_d, "new_p")
hdfs.renames(test_path, new_path)
self.assertFalse(hdfs.path.exists(test_path))
self.assertTrue(hdfs.path.exists(new_path))
def get(self):
src = self.hdfs_paths[0]
dest = hdfs.path.split(self.local_paths[0])[-1]
hdfs.dump(self.data, src, mode="wb")
hdfs.get(src, dest, mode="wb")
with open(dest, 'rb') as fi:
rdata = fi.read()
self.assertEqual(rdata, self.data)
def rename(self):
test_path = self.hdfs_paths[0]
new_path = "%s.new" % test_path
hdfs.dump(self.data, test_path, mode="wb")
hdfs.rename(test_path, new_path)
self.assertFalse(hdfs.path.exists(test_path))
self.assertTrue(hdfs.path.exists(new_path))
self.assertRaises(RuntimeError, hdfs.rename, test_path,
self.local_paths[0])
def set_exe(self, pipes_code):
"""
Dump launcher code to the distributed file system.
"""
if not self.output:
raise RuntimeError("no output directory, can't create launcher")
parent = hdfs.path.dirname(hdfs.path.abspath(self.output.rstrip("/")))
self.exe = hdfs.path.join(parent, utils.make_random_str())
hdfs.dump(pipes_code, self.exe)
def set_exe(self, pipes_code):
"""
Dump launcher code to the distributed file system.
"""
if not self.output:
raise RuntimeError("no output directory, can't create launcher")
parent = hdfs.path.dirname(hdfs.path.abspath(self.output.rstrip("/")))
self.exe = hdfs.path.join(parent, utils.make_random_str())
hdfs.dump(pipes_code, self.exe)
def __ls(self, ls_func, path_transform):
for wd, paths in izip(
(self.local_wd, self.hdfs_wd), (self.local_paths, self.hdfs_paths)
):
for p in paths:
hdfs.dump(self.data, p)
self.assertEqual(path_transform(ls_func(p)[0]), p)
dir_list = [path_transform(p) for p in ls_func(wd)]
self.assertEqual(set(dir_list), set(paths))
def rename(self):
test_path = self.hdfs_paths[0]
new_path = "%s.new" % test_path
hdfs.dump(self.data, test_path)
hdfs.rename(test_path, new_path)
self.assertFalse(hdfs.path.exists(test_path))
self.assertTrue(hdfs.path.exists(new_path))
self.assertRaises(
RuntimeError, hdfs.rename, test_path, self.local_paths[0]
)
def get_hosts(self):
if hdfs.default_is_local():
# only run on HDFS
return
hdfs.dump(self.data, self.hdfs_paths[0], mode="wb")
fs = hdfs.hdfs("default", 0)
hs = fs.get_hosts(self.hdfs_paths[0], 0, 10)
self.assertTrue(len(hs) > 0)
self.assertRaises(ValueError, fs.get_hosts, self.hdfs_paths[0], -10,
10)
self.assertRaises(ValueError, fs.get_hosts, self.hdfs_paths[0], 0, -10)
def runTest(self):
path = make_random_str() + UNI_CHR
hdfs.dump("foo\n", path)
st = hdfs.path.stat(path)
atime, mtime = [getattr(st, 'st_%stime' % _) for _ in 'am']
new_atime, new_mtime = atime + 100, mtime + 200
hdfs.path.utime(path, (new_atime, new_mtime))
st = hdfs.path.stat(path)
self.assertEqual(st.st_atime, new_atime)
self.assertEqual(st.st_mtime, new_mtime)
hdfs.rmr(path)
def __make_tree(self, wd):
d1 = "%s/d1" % wd
t1 = FSTree(d1)
d2 = "%s/d2" % d1
t2 = t1.add(d2)
hdfs.mkdir(d2)
for t, d, bn in ((t1, d1, "f1"), (t2, d2, "f2")):
f = "%s/%s" % (d, bn)
hdfs.dump(self.data, f)
t.add(f, 0)
return t1
def __make_tree(self, wd):
d1 = "%s/d1" % wd
t1 = FSTree(d1)
d2 = "%s/d2" % d1
t2 = t1.add(d2)
hdfs.mkdir(d2)
for t, d, bn in ((t1, d1, "f1"), (t2, d2, "f2")):
f = "%s/%s" % (d, bn)
hdfs.dump(self.data, f)
t.add(f, 0)
return t1
def run_phase_two(args, mappers, input_, logger):
launcher_text = generate_launcher("bl.core.gt.mr.kinship.phase_two")
launcher_name, mr_out_dir = random_str(args), random_str(args)
logger.debug("launcher_name: %r" % (launcher_name, ))
logger.debug("mr_out_dir: %r" % (mr_out_dir, ))
hdfs.dump(launcher_text, launcher_name)
mr_conf = MR_CONF.copy()
mr_conf["mapred.map.tasks"] = mappers
mr_conf["mapred.job.name"] = "kinship_phase_two"
hadut.run_pipes(launcher_name, input_, mr_out_dir, properties=mr_conf)
return mr_out_dir
def runTest(self):
path = make_random_str() + UNI_CHR
hdfs.dump("foo\n", path)
st = hdfs.path.stat(path)
atime, mtime = [getattr(st, 'st_%stime' % _) for _ in 'am']
new_atime, new_mtime = atime + 100, mtime + 200
hdfs.path.utime(path, (new_atime, new_mtime))
st = hdfs.path.stat(path)
self.assertEqual(st.st_atime, new_atime)
self.assertEqual(st.st_mtime, new_mtime)
hdfs.rmr(path)
def get_hosts(self):
if hdfs.default_is_local():
# only run on HDFS
return
hdfs.dump(self.data, self.hdfs_paths[0])
fs = hdfs.hdfs("default", 0)
hs = fs.get_hosts(self.hdfs_paths[0], 0, 10)
self.assertTrue(len(hs) > 0)
self.assertRaises(
ValueError, fs.get_hosts, self.hdfs_paths[0], -10, 10
)
self.assertRaises(ValueError, fs.get_hosts, self.hdfs_paths[0], 0, -10)
def create(self, path, mode, fi=None):
'''Open a nonexistent file. This will just create a new file and generate a
new filehandle for you.'''
mode = mode & 1
maxfh = resource.getrlimit(resource.RLIMIT_NOFILE)[0]
if len(self.filehandles) == maxfh:
raise FuseOSError(errno.EMFILE)
while self.fhmax in self.filehandles.keys():
self.fhmax = (self.fhmax+1)%maxfh
hdfs.dump('',path)
self.filehandles[self.fhmax] = self.hdfs.open_file(path, mode)
return self.fhmax
def __cp_file(self, wd):
fn = "%s/fn" % wd
hdfs.dump(self.data, fn, mode="wb")
dest_dir = "%s/dest_dir" % wd
hdfs.mkdir(dest_dir)
fn_copy_on_wd = "%s/fn_copy" % wd
hdfs.cp(fn, fn_copy_on_wd, mode="wb")
self.assertEqual(hdfs.load(fn_copy_on_wd), self.data)
self.assertRaises(IOError, hdfs.cp, fn, fn_copy_on_wd)
fn_copy_on_dest_dir = "%s/fn" % dest_dir
hdfs.cp(fn, dest_dir, mode="wb")
self.assertEqual(hdfs.load(fn_copy_on_dest_dir), self.data)
self.assertRaises(IOError, hdfs.cp, fn, dest_dir)
def __cp_file(self, wd): fn = "%s/fn" % wd hdfs.dump(self.data, fn) dest_dir = "%s/dest_dir" % wd hdfs.mkdir(dest_dir) fn_copy_on_wd = "%s/fn_copy" % wd hdfs.cp(fn, fn_copy_on_wd) self.assertEqual(hdfs.load(fn_copy_on_wd), self.data) self.assertRaises(IOError, hdfs.cp, fn, fn_copy_on_wd) fn_copy_on_dest_dir = "%s/fn" % dest_dir hdfs.cp(fn, dest_dir) self.assertEqual(hdfs.load(fn_copy_on_dest_dir), self.data) self.assertRaises(IOError, hdfs.cp, fn, dest_dir)
def test_kind(self):
for path in self.path, self.u_path:
self.assertTrue(hdfs.path.kind(path) is None)
try:
hdfs.dump("foo\n", path)
self.assertEqual('file', hdfs.path.kind(path))
hdfs.rmr(path)
hdfs.mkdir(path)
self.assertEqual('directory', hdfs.path.kind(path))
finally:
try:
hdfs.rmr(path)
except IOError:
pass
def test_kind(self):
path = utils.make_random_str()
self.assertTrue(hdfs.path.kind(path) is None)
try:
hdfs.dump("foo\n", path)
self.assertEqual('file', hdfs.path.kind(path))
hdfs.rmr(path)
hdfs.mkdir(path)
self.assertEqual('directory', hdfs.path.kind(path))
finally:
try:
hdfs.rmr(path)
except IOError:
pass
def test_isdir(self):
path = utils.make_random_str()
self.assertFalse(hdfs.path.isdir(path))
try:
hdfs.dump("foo\n", path)
self.assertFalse(hdfs.path.isdir(path))
hdfs.rmr(path)
hdfs.mkdir(path)
self.assertTrue(hdfs.path.isdir(path))
finally:
try:
hdfs.rmr(path)
except IOError:
pass
def test_isdir(self):
for path in self.path, self.u_path:
self.assertFalse(hdfs.path.isdir(path))
try:
hdfs.dump("foo\n", path)
self.assertFalse(hdfs.path.isdir(path))
hdfs.rmr(path)
hdfs.mkdir(path)
self.assertTrue(hdfs.path.isdir(path))
finally:
try:
hdfs.rmr(path)
except IOError:
pass
def lowercase():
f = request.files['file']
f.save(
os.path.join(app.config['UPLOAD_FOLDER'], secure_filename(f.filename)))
with open(f.filename, 'r') as fopen:
hdfs.dump(fopen.read(), '/user/input_lowercase/text')
os.system(
"pydoop script --num-reducers 0 -t '' lowercase.py /user/input_lowercase /user/output_lowercase"
)
list_files = hdfs.hdfs().list_directory('/user/output_lowercase')
return json.dumps([
hdfs.load(file['name'], mode='rt') for file in list_files
if 'SUCCESS' not in file['name']
])
def chown(self):
new_user = '******'
test_path = self.hdfs_paths[0]
hdfs.dump(self.data, test_path)
hdfs.chown(test_path, user=new_user)
path_info = hdfs.lsl(test_path)[0]
self.assertEqual(path_info['owner'], new_user)
prev_owner = path_info['owner']
prev_grp = path_info['group']
# owner and group should remain unchanged
hdfs.chown(test_path, user='', group='')
path_info = hdfs.lsl(test_path)[0]
self.assertEqual(path_info['owner'], prev_owner)
self.assertEqual(path_info['group'], prev_grp)
def test_isdir(self):
path = utils.make_random_str()
self.assertFalse(hdfs.path.isdir(path))
try:
hdfs.dump("foo\n", path)
self.assertFalse(hdfs.path.isdir(path))
hdfs.rmr(path)
hdfs.mkdir(path)
self.assertTrue(hdfs.path.isdir(path))
finally:
try:
hdfs.rmr(path)
except IOError:
pass
def test_kind(self):
path = utils.make_random_str()
self.assertTrue(hdfs.path.kind(path) is None)
try:
hdfs.dump("foo\n", path)
self.assertEqual('file', hdfs.path.kind(path))
hdfs.rmr(path)
hdfs.mkdir(path)
self.assertEqual('directory', hdfs.path.kind(path))
finally:
try:
hdfs.rmr(path)
except IOError:
pass
def wordcount():
f = request.files['file']
f.save(
os.path.join(app.config['UPLOAD_FOLDER'], secure_filename(f.filename)))
with open(f.filename, 'r') as fopen:
hdfs.dump(fopen.read(), '/user/input_wordcount/text')
os.system(
'pydoop script -c combiner wordcount.py /user/input_wordcount /user/output_wordcount'
)
list_files = hdfs.hdfs().list_directory('/user/output_wordcount')
return json.dumps([
hdfs.load(file['name'], mode='rt') for file in list_files
if 'SUCCESS' not in file['name']
])
def chown(self):
new_user = '******'
test_path = self.hdfs_paths[0]
hdfs.dump(self.data, test_path, mode="wb")
hdfs.chown(test_path, user=new_user)
path_info = hdfs.lsl(test_path)[0]
self.assertEqual(path_info['owner'], new_user)
prev_owner = path_info['owner']
prev_grp = path_info['group']
# owner and group should remain unchanged
hdfs.chown(test_path, user='', group='')
path_info = hdfs.lsl(test_path)[0]
self.assertEqual(path_info['owner'], prev_owner)
self.assertEqual(path_info['group'], prev_grp)
def test_kind(self):
for path in self.path, self.u_path:
self.assertTrue(hdfs.path.kind(path) is None)
try:
hdfs.dump("foo\n", path)
self.assertEqual('file', hdfs.path.kind(path))
hdfs.rmr(path)
hdfs.mkdir(path)
self.assertEqual('directory', hdfs.path.kind(path))
finally:
try:
hdfs.rmr(path)
except IOError:
pass
def test_isdir(self):
for path in self.path, self.u_path:
self.assertFalse(hdfs.path.isdir(path))
try:
hdfs.dump("foo\n", path)
self.assertFalse(hdfs.path.isdir(path))
hdfs.rmr(path)
hdfs.mkdir(path)
self.assertTrue(hdfs.path.isdir(path))
finally:
try:
hdfs.rmr(path)
except IOError:
pass
def __ls(self, ls_func, path_transform):
for wd, paths in izip(
(self.local_wd, self.hdfs_wd), (self.local_paths, self.hdfs_paths)
):
for p in paths:
hdfs.dump(self.data, p)
test_dir = "%s/%s" % (wd, "test_dir")
test_path = "%s/%s" % (test_dir, "test_path")
hdfs.dump(self.data, test_path)
paths.append(test_dir)
for recursive in False, True:
if recursive:
paths.append(test_path)
dir_list = [path_transform(p) for p in ls_func(wd, recursive=recursive)]
self.assertEqual(sorted(dir_list), sorted(paths))
def __ls(self, ls_func, path_transform):
for wd, paths in czip((self.local_wd, self.hdfs_wd),
(self.local_paths, self.hdfs_paths)):
for p in paths:
hdfs.dump(self.data, p, mode="wb")
test_dir = "%s/%s" % (wd, "test_dir")
test_path = "%s/%s" % (test_dir, "test_path")
hdfs.dump(self.data, test_path, mode="wb")
paths.append(test_dir)
for recursive in False, True:
if recursive:
paths.append(test_path)
dir_list = [
path_transform(p) for p in ls_func(wd, recursive=recursive)
]
self.assertEqual(sorted(dir_list), sorted(paths))
def __make_tree(self, wd, root="d1", create=True):
"""
d1
|-- d2
| `-- f2
`-- f1
"""
d1 = "%s/%s" % (wd, root)
t1 = FSTree(d1)
d2 = "%s/d2" % d1
t2 = t1.add(d2)
if create:
hdfs.mkdir(d2)
for t, d, bn in ((t1, d1, "f1"), (t2, d2, "f2")):
f = "%s/%s" % (d, bn)
if create:
hdfs.dump(self.data, f, mode="wb")
t.add(f, 0)
return t1
def __make_tree(self, wd, root="d1", create=True):
"""
d1
|-- d2
| `-- f2
`-- f1
"""
d1 = "%s/%s" % (wd, root)
t1 = FSTree(d1)
d2 = "%s/d2" % d1
t2 = t1.add(d2)
if create:
hdfs.mkdir(d2)
for t, d, bn in ((t1, d1, "f1"), (t2, d2, "f2")):
f = "%s/%s" % (d, bn)
if create:
hdfs.dump(self.data, f, mode="wb")
t.add(f, 0)
return t1
def dump(data, hdfs_path):
"""
Dumps data to a file
Args:
:data: data to write to hdfs_path
:hdfs_path: You can specify either a full hdfs pathname or a relative one (relative to your Project's path in HDFS).
"""
hdfs_path = _expand_path(hdfs_path, exists=False)
return hdfs.dump(data, hdfs_path)
def __setup_remote_paths(self):
"""
Actually create the working directory and copy the module into it.
Note: the script has to be readable by Hadoop; though this may not
generally be a problem on HDFS, where the Hadoop user is usually
the superuser, things may be different if our working directory is
on a shared POSIX filesystem. Therefore, we make the directory
and the script accessible by all.
"""
pipes_code = self.__generate_pipes_code()
hdfs.mkdir(self.remote_wd)
hdfs.chmod(self.remote_wd, "a+rx")
hdfs.dump(pipes_code, self.remote_exe)
hdfs.chmod(self.remote_exe, "a+rx")
hdfs.put(self.args.module, self.remote_module)
hdfs.chmod(self.remote_module, "a+r")
self.__warn_user_if_wd_maybe_unreadable(self.remote_wd)
self.logger.debug("Created remote paths:")
self.logger.debug(self.remote_wd)
self.logger.debug(self.remote_exe)
self.logger.debug(self.remote_module)
def __setup_remote_paths(self):
"""
Actually create the working directory and copy the module into it.
Note: the script has to be readable by Hadoop; though this may not
generally be a problem on HDFS, where the Hadoop user is usually
the superuser, things may be different if our working directory is
on a shared POSIX filesystem. Therefore, we make the directory
and the script accessible by all.
"""
pipes_code = self.__generate_pipes_code()
hdfs.mkdir(self.remote_wd)
hdfs.chmod(self.remote_wd, "a+rx")
hdfs.dump(pipes_code, self.remote_exe)
hdfs.chmod(self.remote_exe, "a+rx")
hdfs.put(self.args.module, self.remote_module)
hdfs.chmod(self.remote_module, "a+r")
self.__warn_user_if_wd_maybe_unreadable(self.remote_wd)
self.logger.debug("Created remote paths:")
self.logger.debug(self.remote_wd)
self.logger.debug(self.remote_exe)
self.logger.debug(self.remote_module)
def dump(data, hdfs_path):
"""
Dumps data to a file
Args:
:data: data to write to hdfs_path
:hdfs_path: You can specify either a full hdfs pathname or a relative one (relative to your Project's path in HDFS).
"""
#split = hdfs_path.split('/')
#filename = split[len(split) - 1]
#directory = "/".join(split[0:len(split)-1])
hdfs_path = _expand_path(hdfs_path, exists=False)
return hdfs.dump(data, hdfs_path)
def samefile_rel(self):
p = make_random_str() + UNI_CHR
hdfs.dump("foo\n", p)
self.assertTrue(hdfs.path.samefile(p, hdfs.path.abspath(p)))
hdfs.rmr(p)
def setUp(self):
self.path = make_random_str() + UNI_CHR
hdfs.dump("foo\n", self.path)
def samefile_rel(self):
p = make_random_str() + UNI_CHR
hdfs.dump("foo\n", p)
self.assertTrue(hdfs.path.samefile(p, hdfs.path.abspath(p)))
hdfs.rmr(p)
def good(self):
path = utils.make_random_str()
hdfs.dump("foo\n", path)
self.assertTrue(hdfs.path.exists(path))
hdfs.rmr(path)
self.assertFalse(hdfs.path.exists(path))
def load(self):
for test_path in self.hdfs_paths[0], self.local_paths[0]:
hdfs.dump(self.data, test_path)
rdata = hdfs.load(test_path)
self.assertEqual(rdata, self.data)
def dump(self, data, fname):
"Dump given data directly to HDFS"
hdfs.dump(data, fname)
import os
import itertools
import swiftclient
import pydoop.hdfs as hdfs
container = 'w251-enron'
prefix = 'clean_v2'
hdfs_prefix = '/enron'
authurl = os.environ['SWIFT_AUTH_URL']
user = os.environ['SWIFT_USER']
key = os.environ['SWIFT_KEY']
conn = swiftclient.client.Connection(
authurl=authurl, user=user, key=key)
header, objects = conn.get_container(container, prefix=prefix, full_listing=True)
hdfs.mkdir(hdfs_prefix)
total = len(objects)
count = 1
for obj in objects:
name = obj['name']
print 'Downloading %s (%d of %d)' % (name, count, total)
header, contents = conn.get_object(container, name)
filename = name.replace('/', '_')
hdfs.dump(contents, '%s/%s' % (hdfs_prefix, filename))
count += 1
def setUp(self):
self.path = make_random_str() + UNI_CHR
hdfs.dump("foo\n", self.path)
def load(self):
for test_path in self.hdfs_paths[0], self.local_paths[0]:
hdfs.dump(self.data, test_path, mode="wb")
rdata = hdfs.load(test_path)
self.assertEqual(rdata, self.data)