def file_status(self, path='/'):
"""Return a file status object that represents the path.
Args:
path: The path we want information from
Returns:
FileStatus JSON object:
{
"FileStatus":
{
"accessTime" : 0,
"blockSize" : 0,
"group" : "supergroup",
"length" : 0, //in bytes, zero for directories
"modificationTime": 1320173277227,
"owner" : "webuser",
"pathSuffix" : "",
"permission" : "777",
"replication" : 0,
"type" : "DIRECTORY" //enum {FILE, DIRECTORY}
}
}
"""
url = '%s%s?user.name=%s&op=%s' % (self.prefix, path, self.user,
'GETFILESTATUS')
response = request.push(self.host, self.port, 'GET', url)
js = json.loads(response)
return js
def list(self, path):
"""Get list of directory's content in WebHDFS JSON format.
Args:
path: Given path.
Returns:
The client receives a response with a FileStatuses JSON object:
"""
url = '%s%s?user.name=%s&op=%s' % (self.prefix, path, self.user,
'LISTSTATUS')
response = request.push(self.host, self.port, "GET", url)
js = json.loads(response)
return js['FileStatuses']['FileStatus']
def rename(self, src_path, dst_path):
"""Rename file or directory.
Args:
src_path: Current path
dst_path: New path
Returns:
True if operation is successful and False if not.
"""
url = '%s%s?user.name=%s&op=%s&destination=%s' % (
self.prefix, src_path, self.user, 'RENAME', dst_path)
response = request.push(self.host, self.port, 'PUT', url)
js = json.loads(response)
return js['boolean']
def delete(self, path, recursive="false"):
"""
Args:
path: The path to file or directory we want delete.
Returns:
curl -i -X DELETE "http://<host>:<port>/webhdfs/v1/<path>?op=DELETE
[&recursive=<true|false>]"
"""
url = '%s%s?user.name=%s&op=DELETE&recursive=%s' % (
self.prefix, path, self.user, recursive)
response = request.push(self.host, self.port, 'DELETE', url)
js = json.loads(response)
return js['boolean']
def mkdir(self, path, permission=775):
"""Create directory.
Has the semantics of Unix 'mkdir -p'.
Args:
path: Path to the directory you want to create.
permission: The permission of a file/directory. <Octal>
Returns:
True if successful or False if not.
"""
url = '%s%s?user.name=%s&op=%s&permission=%s' % (
self.prefix, path, self.user, 'MKDIRS', permission)
response = request.push(self.host, self.port, 'PUT', url)
js = json.loads(response)
return js['boolean']
def set_rf(self, path, rf):
"""Set replication factor for file.
Args:
path: The path we want change RF for.
rf: Replication factor <SHORT>
Returns:
True if changing of replication factor is successful,
False if not.
"""
url = '%s%s?user.name=%s&op=SETREPLICATION&replication=%s' % (
self.prefix, path, self.user, rf)
response = request.push(self.host, self.port, 'PUT', url)
js = json.loads(response)
return js['boolean']
def file_checksum(self, path):
"""Get the checksum of a file.
Args:
path: Given path.
Returns:
The checksum object contains:
self.algorithm: The name of the checksum algorithm.
self.bytes: The byte sequence of the checksum in hexadecimal
self.length: The length of the bytes (not the length of the string).
"""
url = '%s%s?user.name=%s&op=%s' % (self.prefix, path, self.user,
'GETFILECHECKSUM')
response = request.push(self.host, self.port, 'GET', url)
js = json.loads(response)
result = structures.Checksum()
result.algorithm = js['FileChecksum']['algorithm']
result.bytes = js['FileChecksum']['bytes']
result.length = js['FileChecksum']['length']
return result