class BaiduFS(Operations):
'''Baidu netdisk filesystem'''
def __init__(self, username, password, *args, **kw):
self.disk = PCS(username, password,captcha_get)
self.buffer = {}
self.traversed_folder = {}
self.bufferLock = Lock()
self.upload_blocks = {} # 文件上传时用于记录块的md5,{PATH:{TMP:'',BLOCKS:''}
self.create_tmp = {} # {goutputstrem_path:file}
self.upload_fails = {} #
self.fd = 3
# 初始化百度服务器
print '设置pcs服务器'
pcs = self.disk.get_fastest_pcs_server()
self.disk.set_pcs_server(pcs)
print 'pcs api server:',pcs
'''
print '设置百度网盘服务器,时间比较长,请耐心等待'
pan = self.disk.get_fastest_mirror()
self.disk.set_pan_server(pan)
print 'baidupan server',pan
'''
self.uploadLock = Lock() # 上传文件时不刷新目录
self.readLock = Lock()
self.downloading_files = []
def unlink(self, path):
print '*'*10,'UNLINK CALLED',path
self.disk.delete([path])
def _add_file_to_buffer(self, path,file_info):
foo = File()
foo['st_ctime'] = file_info['local_ctime']
foo['st_mtime'] = file_info['local_mtime']
foo['st_mode'] = (stat.S_IFDIR | 0777) if file_info['isdir'] \
else (stat.S_IFREG | 0777)
foo['st_nlink'] = 2 if file_info['isdir'] else 1
foo['st_size'] = file_info['size']
self.buffer[path] = foo
def _del_file_from_buffer(self,path):
self.buffer.pop(path)
def getattr(self, path, fh=None):
#print 'getattr *',path
# 先看缓存中是否存在该文件
if not self.buffer.has_key(path):
print path,'未命中'
#print self.buffer
#print self.traversed_folder
jdata = json.loads(self.disk.meta([path]).content)
try:
if 'info' not in jdata:
raise FuseOSError(errno.ENOENT)
if jdata['errno'] != 0:
raise FuseOSError(errno.ENOENT)
file_info = jdata['info'][0]
self._add_file_to_buffer(path,file_info)
st = self.buffer[path].getDict()
return st
except:
raise FuseOSError(errno.ENOENT)
else:
#print path,'命中'
return self.buffer[path].getDict()
def readdir(self, path, offset):
self.uploadLock.acquire()
while True:
try:
foo = json.loads(self.disk.list_files(path).text)
break
except:
print 'error'
files = ['.', '..']
abs_files = [] # 该文件夹下文件的绝对路径
for file in foo['list']:
files.append(file['server_filename'])
abs_files.append(file['path'])
# 缓存文件夹下文件信息,批量查询meta info
# Update:解决meta接口一次不能查询超过100条记录
# 分成 ceil(file_num / 100.0) 组，利用商群
if not self.traversed_folder.has_key(path) or self.traversed_folder[path] == False:
print '正在对',path,'缓存中'
file_num = len(abs_files)
group = int(math.ceil(file_num / 100.0))
for i in range(group):
obj = [f for n,f in enumerate(abs_files) if n % group == i] #一组数据
while 1:
try:
ret = json.loads(self.disk.meta(obj).text)
break
except:
print 'error'
for file_info in ret['info']:
if not self.buffer.has_key(file_info['path']):
self._add_file_to_buffer(file_info['path'],file_info)
#print self.buffer
print '对',path,'的缓存完成'
self.traversed_folder[path] = True
for r in files:
yield r
self.uploadLock.release()
def _update_file_manual(self,path):
while 1:
try:
jdata = json.loads(self.disk.meta([path]).content)
break
except:
print 'error'
if 'info' not in jdata:
raise FuseOSError(errno.ENOENT)
if jdata['errno'] != 0:
raise FuseOSError(errno.ENOENT)
file_info = jdata['info'][0]
self._add_file_to_buffer(path,file_info)
def rename(self, old, new):
#logging.debug('* rename',old,os.path.basename(new))
print '*'*10,'RENAME CALLED',old,os.path.basename(new),type(old),type(new)
while True:
try:
ret = self.disk.rename([(old,os.path.basename(new))]).content
jdata = json.loads(ret)
break
except:
print 'error'
if jdata['errno'] != 0:
# 文件名已存在,删除原文件
print self.disk.delete([new]).content
print self.disk.rename([(old,os.path.basename(new))])
self._update_file_manual(new)
self.buffer.pop(old)
def open(self, path, flags):
self.readLock.acquire()
print '*'*10,'OPEN CALLED',path,flags
#print '[****]',path
"""
Permission denied
accmode = os.O_RDONLY | os.O_WRONLY | os.O_RDWR
if (flags & accmode) != os.O_RDONLY:
raise FuseOSError(errno.EACCES)
"""
self.fd += 1
self.readLock.release()
return self.fd
def create(self, path, mode,fh=None):
# 创建文件
# 中文路径有问题
print '*'*10,'CREATE CALLED',path,mode,type(path)
#if 'outputstream' not in path:
tmp_file = tempfile.TemporaryFile('r+w+b')
foo = self.disk.upload(os.path.dirname(path),tmp_file,os.path.basename(path)).content
ret = json.loads(foo)
print ret
print 'create-not-outputstream',ret
if ret['path'] != path:
# 文件已存在
print '文件已存在'
raise FuseOSError(errno.EEXIST)
'''
else:
print 'create:',path
foo = File()
foo['st_ctime'] = int(time.time())
foo['st_mtime'] = int(time.time())
foo['st_mode'] = (stat.S_IFREG | 0777)
foo['st_nlink'] = 1
foo['st_size'] = 0
self.buffer[path] = foo
'''
'''
dict(st_mode=(stat.S_IFREG | mode), st_nlink=1,
st_size=0, st_ctime=time.time(), st_mtime=time.time(),
st_atime=time.time())
'''
self.fd += 1
return 0
def write(self, path, data, offset, fp):
# 上传文件时会调用
# 4kb ( 4096 bytes ) 每块，data中是块中的数据
# 最后一块的判断：len(data) < 4096
# 文件大小 = 最后一块的offset + len(data)
# 4kb传太慢了，合计成2M传一次
#print '*'*10,path,offset, len(data)
def _block_size(stream):
stream.seek(0,2)
return stream.tell()
_BLOCK_SIZE = 16 * 2 ** 20
# 第一块的任务
if offset == 0:
#self.uploadLock.acquire()
#self.readLock.acquire()
# 初始化块md5列表
self.upload_blocks[path] = {'tmp':None,
'blocks':[]}
# 创建缓冲区临时文件
tmp_file = tempfile.TemporaryFile('r+w+b')
self.upload_blocks[path]['tmp'] = tmp_file
# 向临时文件写入数据，检查是否>= _BLOCK_SIZE 是则上传该块并将临时文件清空
try:
tmp = self.upload_blocks[path]['tmp']
except KeyError:
return 0
tmp.write(data)
if _block_size(tmp) > _BLOCK_SIZE:
print path,'发生上传'
tmp.seek(0)
try:
foo = self.disk.upload_tmpfile(tmp,callback=ProgressBar()).content
foofoo = json.loads(foo)
block_md5 = foofoo['md5']
except:
print foo
# 在 upload_blocks 中插入本块的 md5
self.upload_blocks[path]['blocks'].append(block_md5)
# 创建缓冲区临时文件
self.upload_blocks[path]['tmp'].close()
tmp_file = tempfile.TemporaryFile('r+w+b')
self.upload_blocks[path]['tmp'] = tmp_file
print '创建临时文件',tmp_file.name
# 最后一块的任务
if len(data) < 4096:
# 检查是否有重名，有重名则删除它
while True:
try:
foo = self.disk.meta([path]).content
foofoo = json.loads(foo)
break
except:
print 'error'
if foofoo['errno'] == 0:
logging.debug('Deleted the file which has same name.')
self.disk.delete([path])
# 看看是否需要上传
if _block_size(tmp) != 0:
# 此时临时文件有数据，需要上传
print path,'发生上传,块末尾,文件大小',_block_size(tmp)
tmp.seek(0)
while True:
try:
foo = self.disk.upload_tmpfile(tmp,callback=ProgressBar()).content
foofoo = json.loads(foo)
break
except:
print 'exception, retry.'
block_md5 = foofoo['md5']
# 在 upload_blocks 中插入本块的 md5
self.upload_blocks[path]['blocks'].append(block_md5)
# 调用 upload_superfile 以合并块文件
print '合并文件',path,type(path)
self.disk.upload_superfile(path,self.upload_blocks[path]['blocks'])
# 删除upload_blocks中数据
self.upload_blocks.pop(path)
# 更新本地文件列表缓存
self._update_file_manual(path)
#self.readLock.release()
#self.uploadLock.release()
return len(data)
def mkdir(self, path, mode):
logger.debug("mkdir is:" + path)
self.disk.mkdir(path)
def rmdir(self, path):
logger.debug("rmdir is:" + path)
self.disk.delete([path])
def read(self, path, size, offset, fh):
#print '*'*10,'READ CALLED',path,size,offset
#logger.debug("read is: " + path)
paras = {'Range': 'bytes=%s-%s' % (offset, offset + size - 1)}
while True:
try:
foo = self.disk.download(path, headers=paras).content
return foo
except:
pass
access = None
statfs = None
class BaiduFS(Operations):
'''Baidu netdisk filesystem'''
def __init__(self, username, password, *args, **kw):
self.disk = PCS(username, password)
self.buffer = {}
self.traversed_folder = {}
self.bufferLock = Lock()
self.upload_blocks = {} # 文件上传时用于记录块的md5,{PATH:{TMP:'',BLOCKS:''}
self.create_tmp = {} # {goutputstrem_path:file}
self.upload_fails = {} #
self.fd = 3
# 初始化百度服务器
print '设置pcs服务器'
pcs = self.disk.get_fastest_pcs_server()
self.disk.set_pcs_server(pcs)
print 'pcs api server:', pcs
'''
print '设置百度网盘服务器,时间比较长,请耐心等待'
pan = self.disk.get_fastest_mirror()
self.disk.set_pan_server(pan)
print 'baidupan server',pan
'''
self.uploadLock = Lock() # 上传文件时不刷新目录
self.readLock = Lock()
self.downloading_files = []
def unlink(self, path):
print '*' * 10, 'UNLINK CALLED', path
self.disk.delete([path])
def _add_file_to_buffer(self, path, file_info):
foo = File()
foo['st_ctime'] = file_info['local_ctime']
foo['st_mtime'] = file_info['local_mtime']
foo['st_mode'] = (stat.S_IFDIR | 0777) if file_info['isdir'] \
else (stat.S_IFREG | 0777)
foo['st_nlink'] = 2 if file_info['isdir'] else 1
foo['st_size'] = file_info['size']
self.buffer[path] = foo
def _del_file_from_buffer(self, path):
self.buffer.pop(path)
def getattr(self, path, fh=None):
#print 'getattr *',path
# 先看缓存中是否存在该文件
if not self.buffer.has_key(path):
print path, '未命中'
#print self.buffer
#print self.traversed_folder
jdata = json.loads(self.disk.meta([path]).content)
try:
if 'info' not in jdata:
raise FuseOSError(errno.ENOENT)
if jdata['errno'] != 0:
raise FuseOSError(errno.ENOENT)
file_info = jdata['info'][0]
self._add_file_to_buffer(path, file_info)
st = self.buffer[path].getDict()
return st
except:
raise FuseOSError(errno.ENOENT)
else:
#print path,'命中'
return self.buffer[path].getDict()
def readdir(self, path, offset):
self.uploadLock.acquire()
while True:
try:
foo = json.loads(self.disk.list_files(path).text)
break
except:
print 'error'
files = ['.', '..']
abs_files = [] # 该文件夹下文件的绝对路径
for file in foo['list']:
files.append(file['server_filename'])
abs_files.append(file['path'])
# 缓存文件夹下文件信息,批量查询meta info
# Update:解决meta接口一次不能查询超过100条记录
# 分成 ceil(file_num / 100.0) 组，利用商群
if not self.traversed_folder.has_key(
path) or self.traversed_folder[path] == False:
print '正在对', path, '缓存中'
file_num = len(abs_files)
group = int(math.ceil(file_num / 100.0))
for i in range(group):
obj = [f for n, f in enumerate(abs_files)
if n % group == i] #一组数据
while 1:
try:
ret = json.loads(self.disk.meta(obj).text)
break
except:
print 'error'
for file_info in ret['info']:
if not self.buffer.has_key(file_info['path']):
self._add_file_to_buffer(file_info['path'], file_info)
#print self.buffer
print '对', path, '的缓存完成'
self.traversed_folder[path] = True
for r in files:
yield r
self.uploadLock.release()
def _update_file_manual(self, path):
while 1:
try:
jdata = json.loads(self.disk.meta([path]).content)
break
except:
print 'error'
if 'info' not in jdata:
raise FuseOSError(errno.ENOENT)
if jdata['errno'] != 0:
raise FuseOSError(errno.ENOENT)
file_info = jdata['info'][0]
self._add_file_to_buffer(path, file_info)
def rename(self, old, new):
#logging.debug('* rename',old,os.path.basename(new))
print '*' * 10, 'RENAME CALLED', old, os.path.basename(new), type(
old), type(new)
while True:
try:
ret = self.disk.rename([(old, os.path.basename(new))]).content
jdata = json.loads(ret)
break
except:
print 'error'
if jdata['errno'] != 0:
# 文件名已存在,删除原文件
print self.disk.delete([new]).content
print self.disk.rename([(old, os.path.basename(new))])
self._update_file_manual(new)
self.buffer.pop(old)
def open(self, path, flags):
self.readLock.acquire()
print '*' * 10, 'OPEN CALLED', path, flags
#print '[****]',path
"""
Permission denied
accmode = os.O_RDONLY | os.O_WRONLY | os.O_RDWR
if (flags & accmode) != os.O_RDONLY:
raise FuseOSError(errno.EACCES)
"""
self.fd += 1
self.readLock.release()
return self.fd
def create(self, path, mode, fh=None):
# 创建文件
# 中文路径有问题
print '*' * 10, 'CREATE CALLED', path, mode, type(path)
#if 'outputstream' not in path:
tmp_file = tempfile.TemporaryFile('r+w+b')
foo = self.disk.upload(os.path.dirname(path), tmp_file,
os.path.basename(path)).content
ret = json.loads(foo)
print ret
print 'create-not-outputstream', ret
if ret['path'] != path:
# 文件已存在
print '文件已存在'
raise FuseOSError(errno.EEXIST)
'''
else:
print 'create:',path
foo = File()
foo['st_ctime'] = int(time.time())
foo['st_mtime'] = int(time.time())
foo['st_mode'] = (stat.S_IFREG | 0777)
foo['st_nlink'] = 1
foo['st_size'] = 0
self.buffer[path] = foo
'''
'''
dict(st_mode=(stat.S_IFREG | mode), st_nlink=1,
st_size=0, st_ctime=time.time(), st_mtime=time.time(),
st_atime=time.time())
'''
self.fd += 1
return 0
def write(self, path, data, offset, fp):
# 上传文件时会调用
# 4kb ( 4096 bytes ) 每块，data中是块中的数据
# 最后一块的判断：len(data) < 4096
# 文件大小 = 最后一块的offset + len(data)
# 4kb传太慢了，合计成2M传一次
#print '*'*10,path,offset, len(data)
def _block_size(stream):
stream.seek(0, 2)
return stream.tell()
_BLOCK_SIZE = 16 * 2**20
# 第一块的任务
if offset == 0:
#self.uploadLock.acquire()
#self.readLock.acquire()
# 初始化块md5列表
self.upload_blocks[path] = {'tmp': None, 'blocks': []}
# 创建缓冲区临时文件
tmp_file = tempfile.TemporaryFile('r+w+b')
self.upload_blocks[path]['tmp'] = tmp_file
# 向临时文件写入数据，检查是否>= _BLOCK_SIZE 是则上传该块并将临时文件清空
try:
tmp = self.upload_blocks[path]['tmp']
except KeyError:
return 0
tmp.write(data)
if _block_size(tmp) > _BLOCK_SIZE:
print path, '发生上传'
tmp.seek(0)
try:
foo = self.disk.upload_tmpfile(tmp,
callback=ProgressBar()).content
foofoo = json.loads(foo)
block_md5 = foofoo['md5']
except:
print foo
# 在 upload_blocks 中插入本块的 md5
self.upload_blocks[path]['blocks'].append(block_md5)
# 创建缓冲区临时文件
self.upload_blocks[path]['tmp'].close()
tmp_file = tempfile.TemporaryFile('r+w+b')
self.upload_blocks[path]['tmp'] = tmp_file
print '创建临时文件', tmp_file.name
# 最后一块的任务
if len(data) < 4096:
# 检查是否有重名，有重名则删除它
while True:
try:
foo = self.disk.meta([path]).content
foofoo = json.loads(foo)
break
except:
print 'error'
if foofoo['errno'] == 0:
logging.debug('Deleted the file which has same name.')
self.disk.delete([path])
# 看看是否需要上传
if _block_size(tmp) != 0:
# 此时临时文件有数据，需要上传
print path, '发生上传,块末尾,文件大小', _block_size(tmp)
tmp.seek(0)
while True:
try:
foo = self.disk.upload_tmpfile(
tmp, callback=ProgressBar()).content
foofoo = json.loads(foo)
break
except:
print 'exception, retry.'
block_md5 = foofoo['md5']
# 在 upload_blocks 中插入本块的 md5
self.upload_blocks[path]['blocks'].append(block_md5)
# 调用 upload_superfile 以合并块文件
print '合并文件', path, type(path)
self.disk.upload_superfile(path,
self.upload_blocks[path]['blocks'])
# 删除upload_blocks中数据
self.upload_blocks.pop(path)
# 更新本地文件列表缓存
self._update_file_manual(path)
#self.readLock.release()
#self.uploadLock.release()
return len(data)
def mkdir(self, path, mode):
logger.debug("mkdir is:" + path)
self.disk.mkdir(path)
def rmdir(self, path):
logger.debug("rmdir is:" + path)
self.disk.delete([path])
def read(self, path, size, offset, fh):
#print '*'*10,'READ CALLED',path,size,offset
#logger.debug("read is: " + path)
paras = {'Range': 'bytes=%s-%s' % (offset, offset + size - 1)}
while True:
try:
foo = self.disk.download(path, headers=paras).content
return foo
except:
pass
access = None
statfs = None
class BaiduPCS(object):
"""
Baidu disk uploader.
"""
def __init__(self, filepath, username, password):
"""
Login
"""
(self.filepath, self.filename,
self.dirname, self.filesize) = (filepath, os.path.basename(filepath),
os.path.dirname(filepath),
os.path.getsize(filepath))
self.path = self.dirname + '\\' + self.filename.split('.')[0]
self.pcs = PCS(username, password) #Login
def create_upload(self, num):
self.uplog['md5'][num] = (json.loads(
self.pcs.upload_tmpfile(self.block(
(num - 1) * self.chinksize)).content)['md5'])
self.count += 1
with open(self.dirname + '\\' + self.filename.split('.')[0] + '.json',
'w') as self.new_uplog:
json.dump(self.uplog, self.new_uplog)
print('[' + str(self.count) + '/' + str(self.fid) +
' Uploaded BlockID: ' + str(num) + ' md5: ' +
self.uplog['md5'][num] + ']')
def read_uplog(self):
if os.path.exists(self.dirname + '\\' + self.filename.split('.')[0] +
'.json'):
with open(
self.dirname + '\\' + self.filename.split('.')[0] +
'.json', 'r') as self.uplog_file:
self.uplog = json.load(self.uplog_file)
tmp_dict = {}
for i in sorted(self.uplog['md5'].keys()):
tmp_dict[int(i)] = self.uplog['md5'][i]
self.uplog['md5'] = tmp_dict
else:
self.uplog_file = open(
self.dirname + '\\' + self.filename.split('.')[0] + '.json',
'w')
self.uplog = {'block': 0, 'size': 0, 'md5': {}}
def block(self, location=None):
if location == None:
return math.ceil(os.path.getsize(self.filepath) / self.chinksize)
file = open(self.filepath, 'rb')
file.seek(location, 0)
return io.BytesIO(file.read(self.chinksize))
def upload(self):
"""
Biadu upload module
"""
self.read_uplog()
if int(self.uplog['size']) == 0:
self.chinksize = 1024 * 1024 * 24
self.uplog['size'] = self.chinksize
else:
self.chinksize = self.uplog['size']
self.thread_num = 25
if int(self.uplog['block']) == 0:
self.fid = self.block()
self.count = len(self.uplog['md5'])
with open(self.dirname + '\\' + self.filename.split('.')[0] + '.json',
'w') as self.new_uplog:
json.dump(self.uplog, self.new_uplog)
print('start uploading...')
self.threads = []
for i in range(self.fid):
if not i + 1 in self.uplog['md5']:
while len(threading.enumerate()) - 1 >= self.thread_num:
time.sleep(1)
self.t = threading.Thread(target=self.create_upload,
kwargs={'num': i + 1})
self.t.setDaemon(True)
self.t.start()
self.threads.append(self.t)
for self.thread in self.threads:
self.thread.join()
def superfile(self):
self.pcs.upload_superfile('/' + self.filename,
[(self.uplog['md5'][k])
for k in sorted(self.uplog['md5'].keys())])
def CheckUpload(self):
"""
Check upload status.
Retry if file not uploaded.
"""
if not self.fid == len(self.uplog['md5']):
return 0
return 1
def quota_remaining(self):
self.quota_info = json.loads(self.pcs.quota().content.decode(
"utf-8", "ignore"))
self.remaining = self.quota_info['total'] - self.quota_info['used']
return self.remaining
import sys
import os, json, sys, tempfile
from baidupcsapi import PCS
pcs = PCS('username', 'password')
chinksize = 1024 * 1024 * 16
fid = 1
md5list = []
tmpdir = tempfile.mkdtemp('bdpcs')
with open(sys.argv[1], 'rb') as infile:
while 1:
data = infile.read(chinksize)
if len(data) == 0: break
smallfile = os.path.join(tmpdir, 'tmp%d' % fid)
with open(smallfile, 'wb') as f:
f.write(data)
print('chunk%d size %d' % (fid, len(data)))
fid += 1
print('start uploading...')
ret = pcs.upload_tmpfile(open(smallfile, 'rb'))
md5list.append(json.loads(ret.content)['md5'])
print('md5: %s' % (md5list[-1]))
os.remove(smallfile)
os.rmdir(tmpdir)
remote_path = sys.argv[2] if not sys.argv[2].endswith(
'/') else sys.argv[2] + os.path.basename(sys.argv[1])
ret = pcs.upload_superfile(remote_path, md5list)
print ret.content
