def index(self, path, name, parentpath=None, parent=None, isdir=False):
''' index a folder and store in tree '''
if not parent:
parent = noder.dir_node(name, path, parent)
cnt = 0
for (root, dirs, files) in os.walk(path):
for f in files:
sub = os.path.join(root, f)
n = f
if len(n) > self.MAXLINE:
n = f[:self.MAXLINE] + '...'
Logger.progr('indexing: {:80}'.format(n))
self.noder.file_node(os.path.basename(f), sub, parent,
parentpath)
cnt += 1
for d in dirs:
base = os.path.basename(d)
sub = os.path.join(root, d)
dummy = self.noder.dir_node(base, sub, parent, parentpath)
_, cnt2 = self.index(sub,
base,
parent=dummy,
parentpath=parentpath)
cnt += cnt2
break
# clean line
Logger.progr('{:80}'.format(' '))
return parent, cnt
def _rec_size(self, node, store=True):
'''
recursively traverse tree and return size
@store: store the size in the node
'''
if self.verbose:
Logger.info('getting node size recursively')
if node.type == self.TYPE_FILE:
return node.size
size = 0
for i in node.children:
if node.type == self.TYPE_DIR:
sz = self._rec_size(i, store=store)
if store:
i.size = sz
size += sz
if node.type == self.TYPE_STORAGE:
sz = self._rec_size(i, store=store)
if store:
i.size = sz
size += sz
else:
continue
if store:
node.size = size
return size
def _restore_pickle(self):
'''restore the pickled tree'''
root = pickle.load(open(self.path, 'rb'))
if self.verbose:
m = 'Catalog imported from pickle \"{}\"'.format(self.path)
Logger.info(m)
return root
def _restore_json(self, string):
'''restore the tree from json'''
imp = JsonImporter()
root = imp.import_(string)
if self.verbose:
Logger.info('Catalog imported from json \"{}\"'.format(self.path))
return root
def file_node(self, name, path, parent, storagepath):
'''create a new node representing a file'''
if not os.path.exists(path):
Logger.err('File \"{}\" does not exist'.format(path))
return None
path = os.path.abspath(path)
try:
st = os.lstat(path)
except OSError as e:
Logger.err('OSError: {}'.format(e))
return None
md5 = None
if self.hash:
md5 = utils.md5sum(path)
relpath = os.path.join(os.path.basename(storagepath),
os.path.relpath(path, start=storagepath))
maccess = os.path.getmtime(path)
n = self._node(name, self.TYPE_FILE, relpath, parent,
size=st.st_size, md5=md5, maccess=maccess)
if self.arc:
ext = os.path.splitext(path)[1][1:]
if ext in self.decomp.get_format():
names = self.decomp.get_names(path)
self.list_to_tree(n, names)
return n
def _to_csv(self, node, with_header=False):
'''print the tree to csv'''
rend = anytree.RenderTree(node, childiter=self._sort_tree)
if with_header:
Logger.out(self.CSV_HEADER)
for _, _, node in rend:
self._node_to_csv(node)
def file_node(self, name, path, parent, storagepath):
'''create a new node representing a file'''
if not os.path.exists(path):
Logger.err('File \"{}\" does not exist'.format(path))
return None
path = os.path.abspath(path)
try:
st = os.lstat(path)
except OSError as e:
Logger.err('OSError: {}'.format(e))
return None
md5 = None
if self.hash:
md5 = self._get_hash(path)
relpath = os.sep.join([storagepath, name])
maccess = os.path.getmtime(path)
n = self._node(name,
self.TYPE_FILE,
relpath,
parent,
size=st.st_size,
md5=md5,
maccess=maccess)
if self.arc:
ext = os.path.splitext(path)[1][1:]
if ext.lower() in self.decomp.get_formats():
self._debug('{} is an archive'.format(path))
names = self.decomp.get_names(path)
self.list_to_tree(n, names)
else:
self._debug('{} is NOT an archive'.format(path))
return n
def _save_json(self, node):
'''export the catalog in json'''
exp = JsonExporter(indent=2, sort_keys=True)
with open(self.path, 'w') as f:
exp.write(node, f)
if self.verbose:
Logger.info('Catalog saved to json \"{}\"'.format(self.path))
return True
def get_node(self, top, path):
''' get the node at path '''
r = anytree.resolver.Resolver('name')
try:
return r.get(top, path)
except anytree.resolver.ChildResolverError:
Logger.err('No node at path \"{}\"'.format(path))
return None
def get_node(self, top, path, quiet=False):
'''get the node by internal tree path'''
r = anytree.resolver.Resolver('name')
try:
return r.get(top, path)
except anytree.resolver.ChildResolverError:
if not quiet:
Logger.err('No node at path \"{}\"'.format(path))
return None
def _log(self, string):
if self.debug:
return
if not string:
# clean
Logger.progr('{:80}'.format(' '))
return
if len(string) > self.MAXLINE:
string = string[:self.MAXLINE] + '...'
Logger.progr('indexing: {:80}'.format(string))
def find_name(self,
root,
key,
script=False,
directory=False,
startpath=None,
parentfromtree=False,
fmt='native'):
'''
find files based on their names
@script: output script
@directory: only search for directories
@startpath: node to start with
@parentfromtree: get path from parent instead of stored relpath
@fmt: output format
'''
self._debug('searching for \"{}\"'.format(key))
start = root
if startpath:
start = self.get_node(root, startpath)
self.term = key
found = anytree.findall(start, filter_=self._find_name)
paths = []
for f in found:
if f.type == self.TYPE_STORAGE:
# ignore storage nodes
continue
if directory and f.type != self.TYPE_DIR:
# ignore non directory
continue
# print the node
if fmt == 'native':
self._print_node(f,
withpath=True,
withdepth=True,
withstorage=True,
recalcparent=parentfromtree)
elif fmt == 'csv':
self._node_to_csv(f)
if parentfromtree:
paths.append(self._get_parents(f))
else:
paths.append(f.relpath)
if script:
tmp = ['${source}/' + x for x in paths]
cmd = 'op=file; source=/media/mnt; $op {}'.format(' '.join(tmp))
Logger.info(cmd)
return found
def rec_size(self, node):
'''recursively traverse tree and store dir size'''
if self.verbose:
Logger.info('getting directory size recursively')
if node.type == self.TYPE_FILE:
return node.size
size = 0
for i in node.children:
if node.type == self.TYPE_DIR:
size += self.rec_size(i)
if node.type == self.TYPE_STORAGE:
self.rec_size(i)
else:
continue
node.size = size
return size
def _node_to_csv(self, node, sep=','):
'''
print a node to csv
@node: the node to consider
'''
if not node:
return ''
if node.type == self.TYPE_TOP:
return ''
out = []
if node.type == self.TYPE_STORAGE:
# handle storage
out.append(node.name)
out.append(node.type)
out.append('') # full path
# size
sz = self._rec_size(node, store=False)
out.append(utils.human(sz))
out.append(utils.epoch_to_str(node.ts))
out.append('') # maccess
out.append('') # md5
else:
out.append(node.name)
out.append(node.type)
# node full path
parents = self._get_parents(node)
storage = self._get_storage(node)
fullpath = os.path.join(storage.name, parents)
out.append(fullpath)
out.append(utils.human(node.size))
out.append(utils.epoch_to_str(storage.ts))
out.append(utils.epoch_to_str(node.maccess))
# md5 if any
if node.md5:
out.append(node.md5)
else:
out.append('')
line = sep.join(['"' + o + '"' for o in out])
if len(line) > 0:
Logger.out(line)
def md5sum(path):
''' calculate md5 sum of a file '''
p = os.path.realpath(path)
if not os.path.exists(p):
Logger.err('\nunable to get md5sum on {}'.format(path))
return None
try:
with open(p, mode='rb') as f:
d = hashlib.md5()
while True:
buf = f.read(4096)
if not buf:
break
d.update(buf)
return d.hexdigest()
except PermissionError:
pass
return None
def find_name(self, root, key, script=False):
'''find files based on their names'''
if self.verbose:
Logger.info('searching for \"{}\"'.format(key))
self.term = key
found = anytree.findall(root, filter_=self._find_name)
paths = []
for f in found:
if f.type == self.TYPE_STORAGE:
# ignore storage nodes
continue
self._print_node(f, withpath=True, withdepth=True,
withstorage=True)
paths.append(f.relpath)
if script:
tmp = ['${source}/' + x for x in paths]
cmd = 'op=file; source=/media/mnt; $op {}'.format(' '.join(tmp))
Logger.info(cmd)
return found
def walk(self, root, path, rec=False):
'''walk the tree for ls based on names'''
if self.verbose:
Logger.info('walking path: \"{}\"'.format(path))
r = anytree.resolver.Resolver('name')
found = []
try:
found = r.glob(root, path)
if len(found) < 1:
return []
if rec:
self.print_tree(found[0].parent)
return found
found = sorted(found, key=self._sort, reverse=self.sortsize)
self._print_node(found[0].parent, withpath=False, withdepth=True)
for f in found:
self._print_node(f, withpath=False, pre='- ', withdepth=True)
except anytree.resolver.ChildResolverError:
pass
return found
def find_name(self,
root,
key,
script=False,
directory=False,
startpath=None,
parentfromtree=False):
'''find files based on their names'''
if self.verbose:
Logger.info('searching for \"{}\"'.format(key))
start = root
if startpath:
start = self.get_node(root, startpath)
self.term = key
found = anytree.findall(start, filter_=self._find_name)
paths = []
for f in found:
if f.type == self.TYPE_STORAGE:
# ignore storage nodes
continue
if directory and f.type != self.TYPE_DIR:
# ignore non directory
continue
self._print_node(f,
withpath=True,
withdepth=True,
withstorage=True,
recalcparent=parentfromtree)
if parentfromtree:
paths.append(self._get_parents(f))
else:
paths.append(f.relpath)
if script:
tmp = ['${source}/' + x for x in paths]
cmd = 'op=file; source=/media/mnt; $op {}'.format(' '.join(tmp))
Logger.info(cmd)
return found
def save(self, node):
'''save the catalog'''
if not self.path:
Logger.err('Path not defined')
return False
d = os.path.dirname(self.path)
if d and not os.path.exists(d):
os.makedirs(d)
elif os.path.exists(self.path) and not self.force:
if not utils.ask('Update catalog \"{}\"'.format(self.path)):
Logger.info('Catalog not saved')
return False
if d and not os.path.exists(d):
Logger.err('Cannot write to \"{}\"'.format(d))
return False
if self.metanode:
self.metanode.parent = node
if self.pickle:
return self._save_pickle(node)
return self._save_json(node)
def _save_pickle(self, node):
'''pickle the catalog'''
pickle.dump(node, open(self.path, 'wb'))
if self.verbose:
Logger.info('Catalog saved to pickle \"{}\"'.format(self.path))
return True
def to_dot(self, node, path='tree.dot'):
'''export to dot for graphing'''
anytree.exporter.DotExporter(node).to_dotfile(path)
Logger.info('dot file created under \"{}\"'.format(path))
return 'dot {} -T png -o /tmp/tree.png'.format(path)
def _print_node(self, node, pre='', withpath=False,
withdepth=False, withstorage=False):
'''print a node'''
if node.type == self.TYPE_TOP:
Logger.out('{}{}'.format(pre, node.name))
elif node.type == self.TYPE_FILE:
name = node.name
if withpath:
name = node.relpath
if withstorage:
storage = self._get_storage(node)
attr = ''
if node.md5:
attr = ', md5:{}'.format(node.md5)
compl = 'size:{}{}'.format(utils.human(node.size), attr)
if withstorage:
compl += ', storage:{}'.format(Logger.bold(storage.name))
Logger.file(pre, name, compl)
elif node.type == self.TYPE_DIR:
name = node.name
if withpath:
name = node.relpath
depth = ''
if withdepth:
depth = len(node.children)
if withstorage:
storage = self._get_storage(node)
attr = []
if node.size:
attr.append(['totsize', utils.human(node.size)])
if withstorage:
attr.append(['storage', Logger.bold(storage.name)])
Logger.dir(pre, name, depth=depth, attr=attr)
elif node.type == self.TYPE_STORAGE:
hf = utils.human(node.free)
ht = utils.human(node.total)
dt = ''
if self._has_attr(node, 'ts'):
dt = ', date:'
dt += utils.epoch_to_str(node.ts)
name = '{} (free:{}, total:{}{})'.format(node.name, hf, ht, dt)
Logger.storage(pre, name, node.attr)
elif node.type == self.TYPE_ARC:
if self.arc:
Logger.arc(pre, node.name, node.archive)
else:
Logger.err('Weird node encountered: {}'.format(node))
def _debug(self, string):
if not self.debug:
return
Logger.log(string)
def _debug(self, string):
'''print to debug'''
if not self.debug:
return
Logger.debug(string)
def _print_node(self,
node,
pre='',
withpath=False,
withdepth=False,
withstorage=False,
recalcparent=False):
'''
print a node
@node: the node to print
@pre: string to print before node
@withpath: print the node path
@withdepth: print the node depth info
@withstorage: print the node storage it belongs to
@recalcparent: get relpath from tree instead of relpath field
'''
if node.type == self.TYPE_TOP:
# top node
Logger.out('{}{}'.format(pre, node.name))
elif node.type == self.TYPE_FILE:
# node of type file
name = node.name
if withpath:
if recalcparent:
name = os.sep.join([self._get_parents(node.parent), name])
else:
name = node.relpath
name = name.lstrip(os.sep)
if withstorage:
storage = self._get_storage(node)
attr = ''
if node.md5:
attr = ', md5:{}'.format(node.md5)
compl = 'size:{}{}'.format(utils.human(node.size), attr)
if withstorage:
compl += ', storage:{}'.format(Logger.bold(storage.name))
Logger.file(pre, name, compl)
elif node.type == self.TYPE_DIR:
# node of type directory
name = node.name
if withpath:
if recalcparent:
name = os.sep.join([self._get_parents(node.parent), name])
else:
name = node.relpath
name = name.lstrip(os.sep)
depth = ''
if withdepth:
depth = len(node.children)
if withstorage:
storage = self._get_storage(node)
attr = []
if node.size:
attr.append(['totsize', utils.human(node.size)])
if withstorage:
attr.append(['storage', Logger.bold(storage.name)])
Logger.dir(pre, name, depth=depth, attr=attr)
elif node.type == self.TYPE_STORAGE:
# node of type storage
hf = utils.human(node.free)
ht = utils.human(node.total)
nbchildren = len(node.children)
# get the date
dt = ''
if self._has_attr(node, 'ts'):
dt = 'date:{}'.format(utils.epoch_to_str(node.ts))
ds = ''
# the children size
sz = self._rec_size(node, store=False)
sz = utils.human(sz)
ds = 'totsize:{}'.format(sz)
# format the output
name = '{}'.format(node.name)
args = [
'nbfiles:{}'.format(nbchildren), 'free:{}/{}'.format(hf, ht),
dt, ds
]
Logger.storage(pre, name, '({})'.format(','.join(args)), node.attr)
elif node.type == self.TYPE_ARC:
# archive node
if self.arc:
Logger.arc(pre, node.name, node.archive)
else:
Logger.err('bad node encountered: {}'.format(node))
def _log2file(self, string):
'''log to file'''
if not self.lpath:
return
line = '{}\n'.format(string)
Logger.flog(self.lpath, line, append=True)
def _debug(self, text):
if not self.debug:
return
Logger.debug(text)
