def GetItemPath(self, item):
path = []
while item.IsOk():
path.append(self._tree.GetPyData(item))
item = self._tree.GetItemParent(item)
path.reverse()
return path
def extend_path(path, name):
"""Simpler version of the stdlib's :obj:`pkgutil.extend_path`.
It does not support ".pkg" files, and it does not require an
__init__.py (this is important: we want only one thing (pkgcore
itself) to install the __init__.py to avoid name clashes).
It also modifies the "path" list in place (and returns C{None})
instead of copying it and returning the modified copy.
"""
if not isinstance(path, list):
# This could happen e.g. when this is called from inside a
# frozen package. Return the path unchanged in that case.
return
# Reconstitute as relative path.
pname = os.path.join(*name.split('.'))
for entry in sys.path:
if not isinstance(entry, basestring) or not os.path.isdir(entry):
continue
subdir = os.path.join(entry, pname)
# XXX This may still add duplicate entries to path on
# case-insensitive filesystems
if subdir not in path:
path.append(subdir)
def default_lib_path(data_dir: str, pyversion: Tuple[int, int],
python_path: bool) -> List[str]:
"""Return default standard library search paths."""
# IDEA: Make this more portable.
path = [] # type: List[str]
auto = os.path.join(data_dir, 'stubs-auto')
if os.path.isdir(auto):
data_dir = auto
# We allow a module for e.g. version 3.5 to be in 3.4/. The assumption
# is that a module added with 3.4 will still be present in Python 3.5.
versions = ["%d.%d" % (pyversion[0], minor)
for minor in reversed(range(pyversion[1] + 1))]
# E.g. for Python 3.5, try 2and3/, then 3/, then 3.5/, then 3.4/, 3.3/, ...
for v in ['2and3', str(pyversion[0])] + versions:
for lib_type in ['stdlib', 'builtins', 'third_party']:
stubdir = os.path.join(data_dir, 'typeshed', lib_type, v)
if os.path.isdir(stubdir):
path.append(stubdir)
# Add fallback path that can be used if we have a broken installation.
if sys.platform != 'win32':
path.append('/usr/local/lib/mypy')
# Contents of Python's sys.path go last, to prefer the stubs
# TODO: To more closely model what Python actually does, builtins should
# go first, then sys.path, then anything in stdlib and third_party.
if python_path:
path.extend(sys.path)
return path
def _extract_system_path(self, script):
"""
Extracts and normalizes additional paths for code execution.
For now, there's a default path of data/course/code; this may be removed
at some point.
script : ?? (TODO)
"""
DEFAULT_PATH = ["code"]
# Separate paths by :, like the system path.
raw_path = script.get("system_path", "").split(":") + DEFAULT_PATH
# find additional comma-separated modules search path
path = []
for dir in raw_path:
if not dir:
continue
# path is an absolute path or a path relative to the data dir
dir = os.path.join(self.capa_system.filestore.root_path, dir)
# Check that we are within the filestore tree.
reldir = os.path.relpath(dir, self.capa_system.filestore.root_path)
if ".." in reldir:
log.warning("Ignoring Python directory outside of course: %r", dir)
continue
abs_dir = os.path.normpath(dir)
path.append(abs_dir)
return path
def test_func_dir(tmpdir):
# Test the creation of the memory cache directory for the function.
memory = Memory(location=tmpdir.strpath, verbose=0)
path = __name__.split('.')
path.append('f')
path = tmpdir.join('joblib', *path).strpath
g = memory.cache(f)
# Test that the function directory is created on demand
func_id = _build_func_identifier(f)
location = os.path.join(g.store_backend.location, func_id)
assert location == path
assert os.path.exists(path)
assert memory.location == os.path.dirname(g.store_backend.location)
with warns(DeprecationWarning) as w:
assert memory.cachedir == g.store_backend.location
assert len(w) == 1
assert "The 'cachedir' attribute has been deprecated" in str(w[-1].message)
# Test that the code is stored.
# For the following test to be robust to previous execution, we clear
# the in-memory store
_FUNCTION_HASHES.clear()
assert not g._check_previous_func_code()
assert os.path.exists(os.path.join(path, 'func_code.py'))
assert g._check_previous_func_code()
# Test the robustness to failure of loading previous results.
func_id, args_id = g._get_output_identifiers(1)
output_dir = os.path.join(g.store_backend.location, func_id, args_id)
a = g(1)
assert os.path.exists(output_dir)
os.remove(os.path.join(output_dir, 'output.pkl'))
assert a == g(1)
def extend_path(path, name):
if not isinstance(path, list):
return path
pname = os.path.join(*name.split('.'))
sname = os.extsep.join(name.split('.'))
sname_pkg = sname + os.extsep + 'pkg'
init_py = '__init__' + os.extsep + 'py'
path = path[:]
for dir in sys.path:
if not isinstance(dir, basestring) or not os.path.isdir(dir):
continue
subdir = os.path.join(dir, pname)
initfile = os.path.join(subdir, init_py)
if subdir not in path and os.path.isfile(initfile):
path.append(subdir)
pkgfile = os.path.join(dir, sname_pkg)
if os.path.isfile(pkgfile):
try:
f = open(pkgfile)
except IOError as msg:
sys.stderr.write("Can't open %s: %s\n" % (pkgfile, msg))
else:
for line in f:
line = line.rstrip('\n')
if not line or line.startswith('#'):
continue
path.append(line)
f.close()
return path
def _read_xml(stream):
document = XmlNode()
current_node = document
path = []
while not stream.atEnd():
stream.readNext()
if stream.isStartElement():
node = XmlNode()
attrs = stream.attributes()
for i in xrange(attrs.count()):
attr = attrs.at(i)
node.attribs[_node_name(attr.name())] = unicode(attr.value())
current_node.append_child(_node_name(stream.name()), node)
path.append(current_node)
current_node = node
elif stream.isEndElement():
current_node = path.pop()
elif stream.isCharacters():
current_node.text += unicode(stream.text())
return document
def main():
output_path = "/Users/ramapriyasridharan/Documents/clients_file.txt"
path = []
with open(output_path) as f:
for exptini_path_raw in f:
exptini_path = exptini_path_raw.strip()
path.append(exptini_path)
print path
for i in range(6,11):
print i
for j in range(0,len(path)):
print j
if i < 10:
p = "/%s/0%d"%(path[j],i)
else:
p = "/%s/%d"%(path[j],i)
print p
for root, _, files in os.walk(p):
for f in files:
if f.endswith('.tgz'):
print 'going to extract %s'%f
f1 = os.path.join(p,f)
print f1
tar = tarfile.open(f1)
tar.extractall(p)
tar.close()
def random_walk(self, path_length, alpha=0, rand=random.Random(), start=None):
""" Returns a truncated random walk.
path_length: Length of the random walk.
alpha: probability of restarts.
start: the start node of the random walk.
"""
G = self
if start:
path = [start]
else:
# Sampling is uniform w.r.t V, and not w.r.t E
path = [rand.choice(G.keys())]
while len(path) < path_length:
cur = path[-1]
if len(G[cur]) > 0:
if rand.random() >= alpha:
path.append(rand.choice(G[cur]))
else:
path.append(path[0])
else:
break
return path
def findBestPathFromWordToWord( self, word, toWord ):
"""
Do a breadth first search, which will find the shortest
path between the nodes we are interested in
This will only find a path if the words are
linked by parent relationship
you won't be able to find your cousins
"""
queue = [word]
previous = { word: word }
while queue:
node = queue.pop(0)
if node == toWord: # If we've found our target word.
# Work out how we got here.
path = [node]
while previous[node] != node:
node = previous[node]
path.append( node )
return path
# Continue on up the tree.
for parent in self.getParentRelationships(node):
# unless we've been to this parent before.
if parent not in previous and parent not in queue:
previous[parent] = node
queue.append( parent )
# We didn't find a path to our target word
return None
def mapping_to_alignment(mapping, sam_file, region_fetcher):
''' Convert a mapping represented by a pysam.AlignedRead into an alignment. '''
# The read indices in AlignedRead's aligned_pairs are relative to qstart.
# Convert them to be absolute indices in seq.
path = []
mismatches = set()
deletions = set()
rname = sam_file.getrname(mapping.tid)
for read_i, ref_i in mapping.aligned_pairs:
if read_i != None:
read_i = read_i + mapping.qstart
if ref_i == None:
ref_i = sw.GAP
else:
read_base = mapping.seq[read_i]
ref_base = region_fetcher(rname, ref_i, ref_i + 1).upper()
if read_base != ref_base:
mismatches.add((read_i, ref_i))
path.append((read_i, ref_i))
else:
deletions.add(ref_i)
alignment = {'path': path,
'XO': mapping.opt('XO'),
'XM': mapping.opt('XM'),
'is_reverse': mapping.is_reverse,
'mismatches': mismatches,
'deletions': deletions,
'rname': rname,
'query': mapping.seq,
}
return alignment
def url_to_filename(self, url):
""" Map url to a unique file name/path. Register the directories in the path. """
filename = re.sub(r'^https?://', '', url)
filename = os.path.normpath(filename)
path = []
path_parts = filename.split("/")
for idx, part in enumerate(path_parts):
tries = 0
unique_part = part
new_path = "/".join(path + [part])
while new_path in self.path_types and (self.path_types[new_path] == "file" or len(path_parts)-1 == idx):
tries += 1
unique_part = "%s.%d" % (part, tries)
new_path = "/".join(path + [unique_part])
if len(path_parts)-1 == idx:
self.path_types[new_path] = "file"
else:
self.path_types[new_path] = "dir"
path.append(unique_part)
return "/".join(path)
def test_func_dir():
# Test the creation of the memory cache directory for the function.
memory = Memory(cachedir=env["dir"], verbose=0)
memory.clear()
path = __name__.split(".")
path.append("f")
path = os.path.join(env["dir"], "joblib", *path)
g = memory.cache(f)
# Test that the function directory is created on demand
yield nose.tools.assert_equal, g._get_func_dir(), path
yield nose.tools.assert_true, os.path.exists(path)
# Test that the code is stored.
# For the following test to be robust to previous execution, we clear
# the in-memory store
_FUNCTION_HASHES.clear()
yield nose.tools.assert_false, g._check_previous_func_code()
yield nose.tools.assert_true, os.path.exists(os.path.join(path, "func_code.py"))
yield nose.tools.assert_true, g._check_previous_func_code()
# Test the robustness to failure of loading previous results.
dir, _ = g.get_output_dir(1)
a = g(1)
yield nose.tools.assert_true, os.path.exists(dir)
os.remove(os.path.join(dir, "output.pkl"))
yield nose.tools.assert_equal, a, g(1)
def _gen(self, node, path, prefix):
global args
tail, heads = self._name_node(path), []
edges = node.sufs.keys()
sort = args.sort == "all" or (args.sort == "root" and not path)
if sort:
edges.sort()
for i, edge in enumerate(edges):
child = node.sufs[edge]
path.append(i)
prefix.append(edge)
self._gen_node(child, path, "".join(prefix))
head = self._name_node(path)
if heads and sort:
self._gen_edge(heads[-1], head, style="invisible", dir="none")
heads.append(head)
edge_label = [cgi.escape(edge)]
edge_attrs = {'label': edge_label, 'tooltip': edge_label}
if len(path) == 1:
root = self._name_node([i + len(node.sufs)])
self._print(root, self._stringify_attrs(style="invisible"))
self._gen_edge(root + ":e", head + ":w", **edge_attrs)
else:
self._gen_edge(tail, head, **edge_attrs)
self._gen(child, path, prefix)
path.pop()
prefix.pop()
if sort and heads:
self._print("{ rank=same", " ".join(head for head in heads), "}")
def default_lib_path(data_dir: str, target: int, pyversion: int) -> List[str]:
"""Return default standard library search paths."""
# IDEA: Make this more portable.
path = List[str]()
# Add MYPYPATH environment variable to library path, if defined.
path_env = os.getenv('MYPYPATH')
if path_env is not None:
path[:0] = path_env.split(os.pathsep)
if target in [ICODE, C]:
# Add C back end library directory.
path.append(os.path.join(data_dir, 'lib'))
else:
# Add library stubs directory. By convention, they are stored in the
# stubs/x.y directory of the mypy installation.
version_dir = '3.2'
if pyversion < 3:
version_dir = '2.7'
path.append(os.path.join(data_dir, 'stubs', version_dir))
path.append(os.path.join(data_dir, 'stubs-auto', version_dir))
#Add py3.3 and 3.4 stubs
if sys.version_info.major == 3:
versions = ['3.' + str(x) for x in range(3, sys.version_info.minor + 1)]
for v in versions:
path.append(os.path.join(data_dir, 'stubs', v))
# Add fallback path that can be used if we have a broken installation.
if sys.platform != 'win32':
path.append('/usr/local/lib/mypy')
return path
def evaluateTopicPaths(inputFl, pathFl, outputFl, topicId, metric, cutoff):
profiles = ranking.readInputFl.readInputFl(inputFl, topicId)
# path files can be quite large, so it opens and processes
# the path file line-by-line.
with open(pathFl) as pthFl:
metricFun = ranking.calcUtilities.getMetricFunction(metric)
with tempfile.NamedTemporaryFile("r+", delete=False) as tmpF:
for line in pthFl:
# 15017299068 1 5 d4 : 0 d4:0 d1:0 d3:1 d2:0
(usell, ll) = utils.utils.stripComments(line)
if usell:
lline = re.sub("\s:\s", ":", ll)
lst = lline.split(None)
instId = lst[0]
tTopicId = lst[1]
if topicId == tTopicId:
profileId = lst[2]
path = []
for dc in lst[3 : len(lst)]:
(d, _, c) = dc.partition(":")
path.append(d)
if profiles.getDocRelMap(profileId) == None:
raise Exception("Profile id not found.")
docsToRels = dict(profiles.getDocRelMap(profileId))
util = metricFun(path, docsToRels, cutoff)
tmpF.write(instId + " " + topicId + " " + profileId + " " + str(util) + "\n")
tmpF.flush()
tmpF.close()
shutil.move(tmpF.name, outputFl)
if os.path.exists(tmpF.name):
os.remove(tmpF.name)
return outputFl
def GetCertificatePath(self, cert):
"""
Return the certification path for cert.
"""
path = []
repo = self.GetCertificateRepository()
c = cert
checked = {}
while 1:
subj = str(c.GetSubject())
path.append(c)
if c.GetSubject().as_der() == c.GetIssuer().as_der():
break
# If we come back to a place we've been before, we're in a cycle
# and won't get anywhere. Bail.
if subj in checked:
return ""
checked[subj] = 1
issuers = repo.FindCertificatesWithSubject(str(c.GetIssuer()))
validIssuers = filter(lambda x: not x.IsExpired(), issuers)
# Find an issuer to return. If none is valid, pick one to return.
if len(validIssuers) == 0:
if len(issuers) > 0:
path.append(issuers[0])
break
c = issuers[0]
return path
def searchPath(goal, path, graph):
if goal == path[-1]:
return
for key in graph[path[-1]]:
if key not in path:
path.append(key)
searchPath(goal, path, graph)
def test_func_dir(tmpdir):
# Test the creation of the memory cache directory for the function.
memory = Memory(cachedir=tmpdir.strpath, verbose=0)
memory.clear()
path = __name__.split('.')
path.append('f')
path = tmpdir.join('joblib', *path).strpath
g = memory.cache(f)
# Test that the function directory is created on demand
assert g._get_func_dir() == path
assert os.path.exists(path)
# Test that the code is stored.
# For the following test to be robust to previous execution, we clear
# the in-memory store
_FUNCTION_HASHES.clear()
assert not g._check_previous_func_code()
assert os.path.exists(os.path.join(path, 'func_code.py'))
assert g._check_previous_func_code()
# Test the robustness to failure of loading previous results.
dir, _ = g.get_output_dir(1)
a = g(1)
assert os.path.exists(dir)
os.remove(os.path.join(dir, 'output.pkl'))
assert a == g(1)
def chain_wires(wires):
assert wires.num_vertices > 0
visited = np.zeros(wires.num_vertices, dtype=bool)
path = [0]
visited[0] = True
while not np.all(visited):
front = path[0]
front_neighbors = wires.get_vertex_neighbors(int(front))
for v in front_neighbors:
if visited[v]:
continue
path.insert(0, v)
visited[v] = True
break
end = path[-1]
end_neighbors = wires.get_vertex_neighbors(int(end))
for v in end_neighbors:
if visited[v]:
continue
visited[v] = True
path.append(v)
break
first_neighbors = wires.get_vertex_neighbors(int(path[0])).squeeze()
if len(path) > 2 and path[-1] in first_neighbors:
# Close the loop.
path.append(path[0])
path = wires.vertices[path]
return path
def test_func_dir():
# Test the creation of the memory cache directory for the function.
memory = Memory(cachedir=env['dir'], verbose=0)
path = __name__.split('.')
path.append('f')
path = os.path.join(env['dir'], 'joblib', *path)
g = memory.cache(f)
# Test that the function directory is created on demand
yield nose.tools.assert_equal, g._get_func_dir(), path
yield nose.tools.assert_true, os.path.exists(path)
# Test that the code is stored.
yield nose.tools.assert_false, \
g._check_previous_func_code()
yield nose.tools.assert_true, \
os.path.exists(os.path.join(path, 'func_code.py'))
yield nose.tools.assert_true, \
g._check_previous_func_code()
# Test the robustness to failure of loading previous results.
dir, _ = g.get_output_dir(1)
a = g(1)
yield nose.tools.assert_true, os.path.exists(dir)
os.remove(os.path.join(dir, 'output.pkl'))
yield nose.tools.assert_equal, a, g(1)
def _read_structure_lazy(infile=None, include_hosts=True):
'''Determine and return the organizational structure from a given
host file.
'''
path, hosts, structure = [], [], []
lines = _read_lines_lazy(infile)
for line in lines:
if not _iscomment(line):
if include_hosts:
hosts.append(line)
continue
match = _match_open.search(line)
if match:
path.append(match.group(1))
continue
match = _match_close.search(line)
if match:
if include_hosts:
yield (list(path), list(hosts))
hosts = []
else:
yield list(path)
path.pop()
return
def _find_path(self, src, tgt):
visited = set()
to_visit = [src]
came_from = {}
while len(to_visit) > 0:
elem = to_visit[0]
del to_visit[0]
if elem in visited:
continue
# if path found, redo path and return
if elem is tgt and src in visited: # elem in visited for the case where src = tgt
path = []
prev_elem = elem
while prev_elem is not src:
path.append(prev_elem)
prev_elem = came_from[prev_elem]
path.append(src)
path.reverse()
return path
visited.add(elem)
for outelem in elem.outset:
came_from[outelem] = elem
to_visit.append(outelem)
return [] # if no path found, return an empty path
def read_mda(self, attribute):
lines = attribute.split('\n')
mda = {}
current_dict = mda
path = []
for line in lines:
if not line:
continue
if line == 'END':
break
key, val = line.split('=')
key = key.strip()
val = val.strip()
try:
val = eval(val)
except NameError:
pass
if key in ['GROUP', 'OBJECT']:
new_dict = {}
path.append(val)
current_dict[val] = new_dict
current_dict = new_dict
elif key in ['END_GROUP', 'END_OBJECT']:
if val != path[-1]:
raise SyntaxError
path = path[:-1]
current_dict = mda
for item in path:
current_dict = current_dict[item]
elif key in ['CLASS', 'NUM_VAL']:
pass
else:
current_dict[key] = val
return mda
def build_model_results_path(self, model):
opt = self.process[model]
K_part = "K-{0}".format(self.process["options"]["K"])
path = [opt["init"].upper(), model, K_part]
for key in ["burnIn", "maxIter", "sampleLag", "alpha", "beta"]:
path.append("{0}-{1}".format(self.argmap(key), opt[key]))
return "_".join(path)
def get_node_path(self):
node = self
path = []
while node:
if not node._key:
break
path.append(node._key)
node = node._parent
return '.'.join(path)
def dp(dist_mat):
"""
Find minimum-cost path through matrix `dist_mat` using dynamic programming.
The cost of a path is defined as the sum of the matrix entries on that
path. See the following for details of the algorithm:
- http://en.wikipedia.org/wiki/Dynamic_time_warping
- http://www.ee.columbia.edu/ln/labrosa/matlab/dtw/dp.m
The notation in the first reference was followed, while Dan Ellis's code
(second reference) was used to check for correctness. Return a list path
of indices and the cost matrix.
"""
N, M = dist_mat.shape
# Initialize the cost matrix
cost_mat = np.zeros((N + 1, M + 1))
for i in range(1, N + 1):
cost_mat[i, 0] = np.inf
for i in range(1, M + 1):
cost_mat[0, i] = np.inf
# Fill the cost matrix while keeping traceback information
traceback_mat = np.zeros((N, M))
for i in range(N):
for j in range(M):
penalty = [
cost_mat[i, j], # match (0)
cost_mat[i, j + 1], # insertion (1)
cost_mat[i + 1, j]] # deletion (2)
i_penalty = np.argmin(penalty)
cost_mat[i + 1, j + 1] = dist_mat[i, j] + penalty[i_penalty]
traceback_mat[i, j] = i_penalty
# Traceback from bottom right
i = N - 1
j = M - 1
path = [(i, j)]
while i > 0 or j > 0:
tb_type = traceback_mat[i, j]
if tb_type == 0:
# Match
i = i - 1
j = j - 1
elif tb_type == 1:
# Insertion
i = i - 1
elif tb_type == 2:
# Deletion
j = j - 1
path.append((i, j))
# Strip infinity edges from cost_mat before returning
cost_mat = cost_mat[1:, 1:]
return (path[::-1], cost_mat)
def load_path_file(filename):
path = []
with open(filename, "r") as fin:
for line in fin:
fields = line.split()
if len(fields) == 0 or line[0] == "#":
continue
path.append([float(x) for x in fields])
return np.array(path, dtype=float)
def gen_refs(o, path=None):
path = path or ['']
path.append(o.xpathname())
yield (o.id, '/'.join(path))
if hasattr(o, 'items'):
for child in o.items:
for entry in gen_refs(child, list(path)):
yield entry
def path(self):
'''Nodes from root to self, every tree node has a unique path.
Path can represente a tree node identically, so we use
a path to represent a file in tree file system.
'''
if self.isRoot(): return [self]
path = self.parent.path()
path.append(self)
return path
def make_process(*args, **kwargs):
"""Choose whether to use python built in process or jasper."""
process_cls = process.Process
if config.SPAWN_USING == "jasper":
process_cls = jasper_process.Process
# Add the current working directory and /data/multiversion to the PATH.
env_vars = kwargs.get("env_vars", {}).copy()
path = [
os.getcwd(),
config.DEFAULT_MULTIVERSION_DIR,
]
# If installDir is provided, add it early to the path
if config.INSTALL_DIR is not None:
path.append(config.INSTALL_DIR)
env_vars["INSTALL_DIR"] = config.INSTALL_DIR
path.append(env_vars.get("PATH", os.environ.get("PATH", "")))
env_vars["PATH"] = os.pathsep.join(path)
kwargs["env_vars"] = env_vars
return process_cls(*args, **kwargs)
def find_shortest_path(self, src, dest):
# does not need to be recalculated for all
src = self.Cities.inverse[src]
dest = self.Cities.inverse[dest]
a, pr = shortest_path(self.DistanceMap,
directed=False,
return_predecessors=True)
path = [dest]
k = dest
while pr[src, k] != -9999:
path.append(pr[src, k])
k = pr[src, k]
if path[-1] == src:
path = path[::-1]
res = [(self.Cities[path[0]], self.Cities[path[1]], a[path[0],
path[1]])]
for i in range(len(path) - 2):
res.append((self.Cities[path[i + 1]], self.Cities[path[i + 2]],
res[-1][2] + a[path[i + 1], path[i + 2]]))
return res
else:
return np.zeros(1, int)
def move_up_tree(graph, hpo, move_up, min_dist):
c = hpo
d = 0
path = [hpo]
try:
while graph.nodes[c]['name'] != 'All':
c = graph.nodes[c]['is_a'][0]
d += 1
path.append(c)
except KeyError:
print('Key Error ' + current + ' dist: ' + str(dist))
return None
if min_dist == None or d > min_dist:
if len(path) - move_up > min_dist:
return path[move_up]
else:
# return the item at min_dist
print('Cannot move ' + str(move_up) +
' spaces up the tree. Returning node at minimum distance')
return path[-(min_dist + 1)]
else:
print('Path shorter than min distance. Returning starting point')
return hpo
def on_click(event, x, y, flags, param):
if event != cv2.EVENT_LBUTTONUP:
return
# kp = clostest_key_points(key_points, (x,y), 1)[0]
res_frame = resize_frame(frame)
kp_grid = key_point_grid(orb, res_frame)
print("len(kp_grid)", len(kp_grid))
pos = (y, x)
grid_offset_x = ((frame.shape[0] - 32) % stride) / 2.0 + 16
grid_offset_y = ((frame.shape[1] - 32) % stride) / 2.0 + 16
g_pos = (int(math.floor((pos[0] - grid_offset_x) / stride)),
int(math.floor((pos[1] - grid_offset_y) / stride)))
print("g_pos", g_pos)
path = []
for i in range(playback_random_walk_length):
g_pos, pos = next_pos_play(kp_grid, res_frame.shape, g_pos)
print("g_pos, pos", g_pos, pos)
if g_pos is None:
break
path.append(pos)
path = list(set(path))
windows = np.array([extract_window(res_frame, p) for p in path])
preprocess_input(windows)
features = model.predict(windows)
features = features.reshape((windows.shape[0], 512))
print("windows.shape, feats.shape", windows.shape, features.shape)
show_patches(windows, features, path, frame.shape, memory_graph)
def A_star(start, goal, MyWorld):
# frontier of points that havent been searched
frontier = PriorityQueue()
# put a point in frontier
frontier.put(start, 0)
# dictionary of squares that other squares came from
came_from = {}
came_from[start] = None
cost_so_far = {}
cost_so_far[start] = 0
while not frontier.empty():
current = frontier.get()
print("Searching {0}".format(current))
# If the place we are at is the goal end the search
if current == goal:
break
# Search for each point that is next to current
for next in MyWorld.neighbors(current):
new_cost = cost_so_far[current] + distance(
current, next) + angle_cost(came_from[current], current, next)
if new_cost < cost_so_far.get(next, math.inf):
cost_so_far[next] = new_cost
priority = new_cost + distance(next, goal)
frontier.put(next, priority)
came_from[next] = current
# Find the path from the start to end
current = goal
path = []
while current != start:
print("Current: {0}".format(current))
path.append(current)
current = came_from[current]
path.append(start)
path.reverse()
return path
def random_walk(self,
path_length,
alpha=0,
rand=random.Random(),
cumulated_cache=None,
nodes=None,
q=None,
p=None):
""" Returns a truncated random walk.
path_length: Length of the random walk.
alpha: probability of restarts.
start: the start node of the random walk.
"""
G = self
paths = []
for node in nodes:
if node:
path = [node]
else:
# Sampling is uniform w.r.t V, and not w.r.t E
path = [rand.choice(G.keys())]
while len(path) < path_length:
cur = path[-1]
if len(G[cur]) > 0:
if cur == path[0] or rand.random() >= alpha:
#path.append(rand.choice(G[cur].keys()))
path.append(
weighted_choice(G[cur], cumulated_cache[cur]))
#path.append(choice(G[cur].keys(), p=weights[cur]))
else:
path.append(path[0])
else:
break
paths.append(path)
q.put(paths)
def default_lib_path(data_dir: str, pyversion: Tuple[int, int],
python_path: bool) -> List[str]:
"""Return default standard library search paths."""
# IDEA: Make this more portable.
path = [] # type: List[str]
# Add MYPYPATH environment variable to library path, if defined.
path.extend(mypy_path())
auto = os.path.join(data_dir, 'stubs-auto')
if os.path.isdir(auto):
data_dir = auto
# We allow a module for e.g. version 3.5 to be in 3.4/. The assumption
# is that a module added with 3.4 will still be present in Python 3.5.
versions = [
"%d.%d" % (pyversion[0], minor)
for minor in reversed(range(pyversion[1] + 1))
]
# E.g. for Python 3.5, try 2and3/, then 3/, then 3.5/, then 3.4/, 3.3/, ...
for v in ['2and3', str(pyversion[0])] + versions:
for lib_type in ['stdlib', 'builtins', 'third_party']:
stubdir = os.path.join(data_dir, 'typeshed', lib_type, v)
if os.path.isdir(stubdir):
path.append(stubdir)
# Add fallback path that can be used if we have a broken installation.
if sys.platform != 'win32':
path.append('/usr/local/lib/mypy')
# Contents of Python's sys.path go last, to prefer the stubs
# TODO: To more closely model what Python actually does, builtins should
# go first, then sys.path, then anything in stdlib and third_party.
if python_path:
path.extend(sys.path)
return path
def viterbi(self, sentence):
pmv = np.zeros(shape=(len(self.states), len(sentence)))
bp = np.zeros(shape=(len(self.states), len(sentence)))
for i in range(len(self.states)):
pmv[i][0] = self.start_probas[self.states[
i]] * self.emission_probas[self.states[i]][sentence[0]]
bp[i][0] = 0
for i in range(1, len(sentence)):
for j in range(len(self.states)):
li = []
li1 = []
for k in range(len(self.states)):
li.append(
pmv[k][i - 1] *
self.emission_probas[self.states[j]][sentence[i]] *
self.transition_probas[self.states[k]][self.states[j]])
li1.append(
pmv[k][i - 1] *
self.transition_probas[self.states[k]][self.states[j]])
pmv[j][i] = max(li)
bp[j][i] = li1.index(max(li1))
li2 = []
for k in range(len(self.states) - 1):
li2.append(pmv[k][len(sentence) - 1])
self.proba = max(li2)
bptr = li2.index(max(li2))
li3 = [bptr]
for i in range(len(sentence) - 1, 0, -1):
a = bp[int(bptr)][i]
li3.append(int(a))
bptr = a
li3 = li3[::-1]
path = []
for i in li3:
path.append(self.states[i])
return path, self.proba
def random_walk(self, path_length, alpha=0, rand=random.Random(), start=None):
""" Returns a truncated random walk.
path_length: Length of the random walk.
alpha: probability of restarts.
start: the start node of the random walk.
"""
G = self
if start:
path = [start]
else:
# Sampling is uniform w.r.t V, and not w.r.t E
path = [rand.choice(G.keys())]
while len(path) < path_length:
cur = path[-1]
if len(G[cur]) > 0:
if rand.random() >= alpha:
path.append(rand.choice(G[cur]))
else:
path.append(path[0])
else:
break
return path
def nestTree(parent_node, path, export_data, count_nodes):
### export_data,count_nodes are used for QC only
children = edges[parent_node]
path.append(0)
for child in children.keys():
tuple_path = tuple(path)
#count_nodes+=1
#try: temp = string.join(edges[child].keys(),'|')
#except Exception: temp = ''
#export_data.write(str(tuple_path)+'\t'+child+'\t'+temp+'\n')
p = list(
path
) ### Otherwise, the same path somehow gets used (alternative to copy.deepcopy())
if child in edges:
count_nodes = nestTree(child, p, export_data, count_nodes)
#if count_nodes==1000: kill
path_ontology_db[tuple_path] = child
if child not in built_ontology_paths:
built_ontology_paths[child] = [tuple_path]
elif tuple_path not in built_ontology_paths[child]:
built_ontology_paths[child].append(tuple_path)
path[-1] += 1
return count_nodes
def restore_path(grid, start, end, geo_ref):
increments = [(1, 0), (1, 1), (0, 1), (0, -1), (-1, -1), (-1, 0), (-1, 1),
(1, -1)]
path = []
current = start
time_start = timeit.default_timer()
while True:
path.append(current)
neighbours = [(current[0] + dx, current[1] + dy)
for dx, dy in increments]
vals = []
for x, y in neighbours:
try:
val = grid[x, y]
if val == 0 and not ((x, y) == end): val = 10000000000000
vals.append(val)
except IndexError:
vals.append(10000000000000000000)
minimum = min(vals)
current = neighbours[vals.index(minimum)]
if current == end: break
#if grid[current] <2: break
if timeit.default_timer() - time_start > 0.1: return None
return path[::-1]
def normal_random_walk(self,
path_length,
rand=random.Random(),
start=None):
"""
Define a normal random walk without restart to generate positive training pairs
"""
G = self
pairs = []
if start:
path = [start]
else:
# Sampling is uniform w.r.t V, and not w.r.t E
path = [rand.choice(G.keys())]
while len(path) < path_length:
cur = path[-1]
if len(G[cur]) > 0:
next_node = rand.choice(G[cur])
path.append(next_node)
if path[0] != next_node:
pairs.append((path[0], next_node))
else:
break
return pairs
def get_lowest_latency_path(start, end):
# disjktra's algorithm
distances = {}
predecessors = {}
to_assess = partial_latencies.keys()
for node in partial_latencies:
distances[node] = float("inf")
predecessors[node] = None
sp_set = []
distances[start] = 0
while len(sp_set) < len(to_assess):
still_in = {
node: distances[node]
for node in [node for node in to_assess if node not in sp_set]
}
closest = min(still_in, key=distances.get)
sp_set.append(closest)
for node in partial_latencies[closest]:
if distances[node] > distances[closest] + partial_latencies[
closest][node]:
distances[node] = distances[closest] + partial_latencies[
closest][node]
predecessors[node] = closest
path = [end]
while start not in path:
path.append(predecessors[path[-1]])
path_ordered = path[::-1]
path_latency = distances[end]
return {"path": path_ordered, "latency": path_latency}
def deserialize_lms_sig(buffer):
q = deserialize_u32(buffer[0:4])
# print "q: " + str(q)
lmots_type = typecode_peek(buffer[4:8])
# print "lmots_type: " + str(lmots_type)
if lmots_type in lmots_params:
pos = 4 + LmotsSignature.bytes(lmots_type)
else:
raise ValueError(err_unknown_typecode, str(lmots_type))
lmots_sig = buffer[4:pos]
lms_type = typecode_peek(buffer[pos:pos + 4])
if lms_type in lms_params:
m, h, LenI = lms_params[lms_type]
else:
raise ValueError(err_unknown_typecode, str(lms_type))
if (q >= 2**h):
raise ValueError(err_bad_value)
pos = pos + 4
path = list()
for i in xrange(0, h):
path.append(buffer[pos:pos + m])
pos = pos + m
# PrintUtl.print_hex("buffer tail", buffer[pos:])
return lms_type, q, lmots_sig, path
def a_star(grid, start, goal):
path = []
path_cost = 0
if (start == goal):
return path, path_cost
queue = PriorityQueue()
queue.put((0, start))
visited = set(start)
branch = {}
found = False
while not queue.empty():
item = queue.get()
current_node = item[1]
if current_node == start:
current_cost = 0.0
else:
current_cost = branch[current_node][0]
if current_node == goal:
print('Found a path.')
found = True
break
else:
for action in valid_actions(grid, current_node):
# get the tuple representation
da = action.delta
next_node = (current_node[0] + da[0], current_node[1] + da[1])
branch_cost = current_cost + action.cost
queue_cost = branch_cost + heuristic(next_node, goal)
if next_node not in visited:
visited.add(next_node)
branch[next_node] = (branch_cost, current_node, action)
queue.put((queue_cost, next_node))
if found:
# retrace steps
n = goal
path_cost = branch[n][0]
path.append(goal)
while branch[n][1] != start:
path.append(branch[n][1])
n = branch[n][1]
path.append(branch[n][1])
else:
print('**********************')
print('Failed to find a path!')
print('**********************')
return None
return path[::-1], path_cost
def show_result(self, initial, goal, duration, length, term, potarr):
line = dict()
line['initial'] = [self.int_float(x) for x in initial]
line['goal'] = [self.int_float(x) for x in goal]
# Omit duration for reproducibility
# line['duration'] = duration
line['length'] = int(length) if length % 1 == 0.0 else float(f"{length:.5f}")
path = list()
# Iterable of poses, handles unitary paths
if term == self.noPath:
poses = []
elif term.getSort() == self.m.findSort('Pose'):
poses = [term]
else:
poses = term.arguments()
for pose in poses:
x, y, t = self.destruct_pose(pose)
path.append([self.int_float(x), self.int_float(y)])
line['path'] = path
if potarr is not None:
line['navfn'] = [float(entry) for row in potarr.arguments() for entry in row.arguments()]
print(json.dumps(line))
def __random_walk__(G, path_length, start, alpha=0, rand=random.Random()):
'''
Returns a truncated random walk.
:param G: networkx graph
:param path_length: Length of the random walk.
:param alpha: probability of restarts.
:param rand: random number generator
:param start: the start node of the random walk.
:return:
'''
path = [start]
while len(path) < path_length:
cur = path[-1]
if len(G.neighbors(cur)) > 0:
if rand.random() >= alpha:
path.append(rand.choice(G.neighbors(cur)))
else:
path.append(path[0])
else:
break
return path
def dsp(n, s, t, C):
'''
DSP algorithm using label correction.
'''
openList = [s]
g = [float('inf') for i in range(n)]
g[s] = 0
parent = [-1 for i in range(n)]
while (len(openList) > 0):
i = openList.pop()
for j in range(n):
if (g[i] + C[i, j]) < g[j] and (g[i] + C[i, j]) < g[t]:
g[j] = g[i] + C[i, j]
parent[j] = i
if j != t:
openList = [j] + openList
end = int(t)
path, cost = [], []
while (end != -1):
path.append(int(end))
cost.append(g[end])
end = parent[end]
return path[::-1], cost
def make_path(fid, sfid=None, cid=None, subid=None):
path = [urllib.unquote(fid)]
if sfid is not None:
path.append(urllib.unquote(sfid))
if cid is not None:
if subid is not None:
path.append(cid + '@' + subid)
else:
path.append(cid)
return '/'.join(path)
def a_starGraph(graph, start, goal):
"""Modified A* to work with NetworkX graphs."""
path = []
path_cost = 0
if (start == goal):
return path, path_cost
queue = PriorityQueue()
queue.put((0, start))
visited = set(start)
branch = {}
found = False
while not queue.empty():
item = queue.get()
current_node = item[1]
if current_node == start:
current_cost = 0.0
else:
current_cost = branch[current_node][0]
if current_node == goal:
print('Found a path.')
found = True
break
else:
for next_node in graph[current_node]:
cost = graph.edges[current_node, next_node]['weight']
branch_cost = current_cost + cost
queue_cost = branch_cost + heuristic(next_node, goal)
if next_node not in visited:
visited.add(next_node)
branch[next_node] = (branch_cost, current_node)
queue.put((queue_cost, next_node))
if found:
# retrace steps
n = goal
path_cost = branch[n][0]
path.append(goal)
while branch[n][1] != start:
path.append(branch[n][1])
n = branch[n][1]
path.append(branch[n][1])
else:
print('**********************')
print('Failed to find a path in graph!')
print('**********************')
return None, None
return path[::-1], path_cost
def _extract_path(tpl: BIP32Template, want_nomatch: bool = False) -> List[int]:
path = []
have_nomatch = False
for s in tpl.sections:
for start, end in s:
if want_nomatch and not have_nomatch:
if (start & HARDENED_INDEX_MASK) != 0:
path.append(start-1)
have_nomatch = True
break
if (end | HARDENED_INDEX_START) != 0xFFFFFFFF:
path.append(end+1)
have_nomatch = True
break
elif random.choice((True, False)):
path.append(random.randint(start, end))
break
else:
path.append(s[0][0])
if want_nomatch and not have_nomatch:
# Could not put non-matching value in any position, that means that
# all sections contain wildcard match. To make a non-matching path,
# just flip the last hardened section to unhardened.
# If there's no hardened sections, flip first section to hardened
have_flipped = False
for i, s in enumerate(tpl.sections):
assert len(s) == 1
start, end = s[0]
assert (start & HARDENED_INDEX_MASK) == 0
assert (end | HARDENED_INDEX_START) == 0xFFFFFFFF
if start >= HARDENED_INDEX_START and not have_flipped:
# Found the hardened section, flip
path[i] = start ^ HARDENED_INDEX_START
have_flipped = True
# do not break so all sections are checked with asserts
if not have_flipped:
# All sections were unhardened, make first section hardened
assert tpl.sections[0][0][0] < HARDENED_INDEX_START
path[0] = tpl.sections[0][0][0] | HARDENED_INDEX_START
return path
def compute_path(cls, col, row, depth):
"""Constructor that takes a col,row of a tile and computes the path.
"""
assert col % 2 == 0
assert row % 4 == 0
xradius = 2**depth
yradius = 2 * 2**depth
colbounds = [-xradius, xradius]
rowbounds = [-yradius, yradius]
path = []
for level in xrange(depth):
# Strategy: Find the midpoint of this level, and determine which
# quadrant this row/col is in. Then set the bounds to that level
# and repeat
xmid = (colbounds[1] + colbounds[0]) // 2
ymid = (rowbounds[1] + rowbounds[0]) // 2
if col < xmid:
if row < ymid:
path.append(0)
colbounds[1] = xmid
rowbounds[1] = ymid
else:
path.append(2)
colbounds[1] = xmid
rowbounds[0] = ymid
else:
if row < ymid:
path.append(1)
colbounds[0] = xmid
rowbounds[1] = ymid
else:
path.append(3)
colbounds[0] = xmid
rowbounds[0] = ymid
return cls(col, row, path)
def random_walk(self, path_length, alpha=0, rand=random.Random(), start=None):
""" Returns a truncated random walk.
path_length: Length of the random walk.
alpha: probability of restarts.
start: the start node of the random walk.
"""
G = self
if start:
path = [start]
else:
# Sampling is uniform w.r.t V, and not w.r.t E
path = [rand.choice(list(G.keys()))]
while len(path) < path_length:
cur = path[-1]
if len(G[cur]) > 0:
if rand.random() >= alpha:
###############
# RANDOM WALK #
###############
# path.append(rand.choice(G[cur]))
###############
# RANDOM WALK #
###############
################################
# NON-BACKTRACKING RANDOM WALK #
################################
neighbors = G[cur]
if len(path) > 1:
previous = path[-2]
try:
neighbors.remove(previous)
except ValueError:
pass
if neighbors:
path.append(rand.choice(neighbors))
else:
path.append(path[0])
################################
# NON-BACKTRACKING RANDOM WALK #
################################
else:
path.append(path[0])
else:
break
return [str(node) for node in path]
def get_resource(self, action=None, token=True):
path = []
if self.settings.prefix is not None:
path.append(self.settings.prefix)
if action is None:
path.append('')
if action is not None:
path.append(action)
if token == 'one-time':
query = {'token': self.gen_one_time_token()}
elif token:
query = {'token': self.auth_token}
else:
query = None
return url.Resource('/'.join(path), query)
def bgpa_customers_routes(AS, br_tuples):
if self.paths_from_customers == {}:
return
for br_id, br_addr in brs:
nn = br_addr[0]
msg_dst = br_addr[0] + ' ' + br_addr[1]
for path_dst, path_tuple in self.paths_from_customers:
path = path_tuple[0]
path = path.append(self.AS)
networks = self.networks_in_as[path[-1]]
networks = ' '.join(networks)
path = ' '.join(path)
for ip_addr in self.addrs:
if ip_addr[0] == nn:
msg_src = ip_addr[0] + ' ' + ip_addr[1]
bgpa_msg = msg_src + ' ' + msg_dst + ' ' + self.ID + ' PATHADV ' + path + ' NETWKS ' + networks
print bgpa_msg
def get_parameter(param_path, template_path, db):
"""returns string value of parameter, given its path, which can contain
wildcards,template path, from which wildcards will be substituted and
structure, from which it should be fetched"""
# path to parameter to be built
path = []
# TODO optimize this
if param_path.startswith('.'):
path = template_path[:-1] + [param_path[1:]]
else:
paths_diverged = False
all_steps = param_path.split('.')
for j in range(len(all_steps)):
step = all_steps[j]
# we need to substitute all asterisks for their values
if step == '*':
if not paths_diverged:
path.append(template_path[j])
else:
sys.stderr.write("""
Paths already diverged,wildcard cant be
substituted {}""".format(path))
# special case if we want to return key not value of dict
# only for assurance that ^ is at the end and paths has not
# diverged
elif step == '^':
if j == len(all_steps) - 1 and not paths_diverged:
path.append('^')
else:
sys.stderr.write("""
Paths already diverged or ^ in parameter path\n""")
else:
if step != template_path[j]:
paths_diverged = True
path.append(step)
# now we get value for built path
val = db
for i in range(len(path)):
step = path[i]
if step == '^':
val = template_path[i]
else:
val = val[step]
return str(val)
def get_blocks(self):
'''
Compares two binary strings under the assumption that y is the result of
applying the following transformations onto x:
* change single bytes in x (likely)
* expand single bytes in x to two bytes (less likely)
* drop single bytes in x (even less likely)
Returns a generator that yields elements of the form (unmodified, xdiff, ydiff),
where each item represents a binary chunk with "unmodified" denoting whether the
chunk is the same in both strings, "xdiff" denoting the size of the chunk in x
and "ydiff" denoting the size of the chunk in y.
Example:
>>> x = "abcdefghijklm"
>>> y = "mmmcdefgHIJZklm"
>>> list(MemoryComparator(x, y).get_blocks())
[(False, 2, 3), (True, 5, 5),
(False, 3, 4), (True, 3, 3)]
'''
x, y = self.x, self.y
_, moves = self.get_grid()
# walk the grid
path = []
i, j = 0, 0
while True:
dy, dx = self.move_to_gradient[moves[j][i]]
if dy == dx == 0: break
path.append((dy == 1 and x[i] == y[j], dy, dx))
j, i = j + dy, i + dx
for i, j in zip(range(i, len(x)), itertools.count(j)):
if j < len(y): path.append((x[i] == y[j], 1, 1))
else: path.append((False, 0, 1))
i = j = 0
for unmodified, subpath in itertools.groupby(path, itemgetter(0)):
ydiffs = list(map(itemgetter(1), subpath))
dx, dy = len(ydiffs), sum(ydiffs)
yield unmodified, dx, dy
i += dx
j += dy
def find_best_path(k, n, curr_path, actual_path, min_cost, cost):
if k == n:
c = 0
path = []
path.append(0)
for i in curr_path:
path.append(i)
path.append(n + 1)
for j in range(len(path) - 1):
c += cost[path[j], path[j + 1]]
if min_cost[0] == -1 or c < min_cost[0]:
while actual_path.__len__() > 0:
actual_path.pop()
for tmp in path:
actual_path.append(tmp)
min_cost[0] = c
else:
for i in range(k -1, n):
curr_path[k - 1], curr_path[i] = curr_path[i], curr_path[k - 1]
find_best_path(k + 1, n, curr_path, actual_path, min_cost, cost)
curr_path[k - 1], curr_path[i] = curr_path[i], curr_path[k - 1]
def random_walk(G, path_length, alpha=0, rand=random.Random(), start=None, temporal=False):
""" Returns a truncated random walk.
path_length: Length of the random walk.
alpha: probability of restarts.
start: the start node of the random walk.
temporal: whether to use "time" edge attribute to only generate a temporal walk
"""
if start:
path = [start]
else:
# Sampling is uniform w.r.t V, and not w.r.t E
path = [rand.choice(G.nodes())]
if temporal:
u = path[-1]
times = [G[u][v]['time'] for v in G.neighbors(u)]
if times:
t = rand.choice(times)
else:
return path
while len(path) < path_length:
cur = path[-1]
if len(G[cur]) > 0:
if rand.random() >= alpha:
if temporal:
node_and_time = [(v,G[cur][v]['time']) for v in G.neighbors(cur) if G[cur][v]['time'] > t]
if node_and_time:
next_node, t = rand.choice(node_and_time)
path.append(next_node)
else:
break
else:
path.append(rand.choice(list(G[cur].keys())))
else:
path.append(path[0])
else:
break
return path
def run_trial(policy,
problem_server,
limit=1000,
det_sample=False,
show_time=False):
"""Run policy on problem. Returns (cost, path), where cost may be None if
goal not reached before horizon."""
env = RemoteEnv(problem_server)
obs = env.reset()
# total cost of this run
cost = 0
path = []
if det_sample:
bak_sample = policy.action_space.weighted_sample
policy.action_space.weighted_sample = det_sampler
try:
for action_num in range(1, limit):
if show_time:
start = time()
action, _ = policy.get_action(obs)
if show_time:
elapsed = time() - start
print('get_action took %fs' % elapsed)
path.append(env.action_name(action))
obs, reward, done, step_info = env.step(action)
cost += step_info['step_cost']
if step_info['goal_reached']:
path.append('GOAL! :D')
return cost, True, path
# we can run out of time or run out of actions to take
if done:
break
path.append('FAIL! D:')
finally:
if det_sample:
policy.action_space.weighted_sample = bak_sample
return cost, False, path