class DFS(object):
def __init__(self, hostname='speedy', port=8020):
self.hfs = Hfilesystem(hostname, port)
self.urlhead = 'hdfs://' + hostname + ':' + str(port)
self.urlheadlen = len(self.urlhead)
def mkdir(self, dirname):
if not self.exists(dirname):
self.hfs.mkdir(dirname)
def rmdir(self, dirname):
if self.exists(dirname):
self.hfs(dirname)
def rm_rf(self, d):
self.rmdir(d)
def rename(self, srcpath, destpath):
if os.path.exists(destpath):
self.rm_rf(destpath)
self.hfs.rename(srcpath, destpath)
def is_done(self, dirname):
return self.exists(os.path.join(dirname, DONE_TAG))
def get_subdirs(self, dirname, checkdone = False):
if not self.exists(dirname):
return []
return [sdir[sdir.rindex['/']+1:] for sdir in self.hfs.listdir(dirname) if self.hfs.isDir(os.path.join(dirname, sdir))
if not checkdone or self.is_done(os.path.join(dirname, sdir))]
def get_abs_subdirs(self, dirname, checkdone = False):
if not self.exists(dirname):
return []
return [sdir[self.urlheadlen:] for sdir in self.hfs.listdir(dirname) if self.hfs.isDir(os.path.join(dirname, sdir))
if not checkdone or self.is_done(os.path.join(dirname, sdir))]
def get_unfinished_subdirs(self, dirname, jobname = '', checkdone = False):
return [sdir for sdir in self.get_subdirs(dirname, checkdone) if
not self.exists(os.path.join(dirname, sdir, jobname + FINISHED_TAG))]
def get_buffered_subdirs(self, dirname, jobname = '', checkdone = False):
return [sdir for sdir in self.get_subdirs(dirname, checkdone = False) if
not self.exists(os.path.join(dirname, sdir, jobname + FINISHED_TAG))
and not self.exists(os.path.join(dirname, sdir, jobname + STARTED_TAG))]
def get_unfinished_abs_subdirs(self, dirname, jobname = '', checkdone = False):
return [sdir for sdir in self.get_abs_subdirs(dirname, checkdone) if
not self.exists(os.path.join(dirname, sdir, jobname + FINISHED_TAG))]
def get_buffered_abs_subdirs(self, dirname, jobname = '', checkdone = False):
return [os.path.join(dirname, sdir) for sdir in self.get_abs_subdirs(dirname, checkdone = False) if
not self.exists(os.path.join(dirname, sdir, jobname + FINISHED_TAG))
and not self.exists(os.path.join(dirname, sdir, jobname + STARTED_TAG))]
def get_subfiles(self, dirname):
if not self.exists(dirname):
return []
return [sfile for sfile in self.hfs.listdir(dirname) if self.hfs.isFile(os.path.join(dirname, sfile)) ]
def get_abs_subfiles(self, dirname):
return [os.path.join(dirname, sfile) for sfile in self.get_subfiles(dirname) ]
def get_subdir_num(self, dirname):
return len(self.get_subdirs(dirname))
def get_unfinished_subdir_num(self, dirname, jobname = '', checkdone = False):
return len(self.get_unfinished_subdirs(dirname, jobname, checkdone = False))
def get_buffered_subdir_num(self, dirname, jobname = '', checkdone = False):
return len(self.get_buffered_subdirs(dirname, jobname, checkdone = False))
def get_subfile_num(self, dirname):
return len(self.get_subfiles(dirname))
# recursive
def get_dir_size(self, dirname):
size = 0L
if self.exists(dirname):
for node in self.hfs.listdir(dirname):
if self.hfs.isFile(node):
size += self.hfs.stat(node).mSize
else:
size += self.get_dir_size(node)
return size
# recursive
def get_unfinished_dir_size(self, dirname, jobname = ''):
size = 0L
if self.exists(dirname) and not self.exists(os.path.join(dirname, jobname + FINISHED_TAG)):
for node in self.hfs.listdir(dirname):
if self.hfs.isFile(node):
size += self.hfs.stat(node).mSize
else:
size += self.get_dir_size(node)
return size
def get_buffered_dir_size(self, dirname, jobname = ''):
size = 0L
if self.exists(dirname) and not self.exists(os.path.join(dirname, jobname + FINISHED_TAG)):
for node in self.hfs.listdir(dirname):
if self.hfs.isFile(node):
size += self.hfs.stat(node).mSize
else:
size += self.get_dir_size(node)
return size
def exists(self, pathname):
return self.hfs.exists(pathname)
