def find_current(self, rs):
if isinstance(rs, MergeHistoryNode):
source_node = rs
else:
source_node = self.rsids.get(rs)
if source_node is None:
return None
target = source_node
chain = [target]
# Walk forward from source to sink node
while 1:
next_t = target.child
if next_t is not None:
target = next_t
chain.append(target)
else:
break
# Now walk backward to find first node that isn't deleted
for i in range(len(chain) - 1, 0, -1):
if not chain[i].deleted:
return chain[i].rsid
def load(cls, path_to_cluster):
"""Load from disk.
Args:
path_to_cluster (str): Folder where `ClusterChain` was saved to using `ClusterChain.save`.
Returns:
ClusterChain: The loaded object.
"""
if os.path.isfile(path_to_cluster):
C = smat_util.load_matrix(path_to_cluster)
return cls.from_partial_chain(C)
config_path = os.path.join(path_to_cluster, "config.json")
if not os.path.exists(config_path):
raise ValueError(
f"Cluster config file, {config_path}, does not exist")
with open(config_path, "r", encoding="utf-8") as fin:
config = json.loads(fin.read())
length = config.get("len", None)
if length is None:
raise ValueError(
f'Cluster config file, {config_path}, does not have "len" parameter'
)
chain = []
for i in range(length):
chain.append(
smat_util.load_matrix(
os.path.join(path_to_cluster,
f"C{i}.npz")).tocsc().astype(np.float32))
return cls(chain)
def create_chain(elem):
global errors
bldg_name, ways = elem
if len(ways) < 3:
return (bldg_name, [])
curr = (ways[0][0], ways[0][1])
start = (ways[0][0], ways[0][1])
d = defaultdict(list)
loop = False
for way in ways:
d[(way[0], way[1])].append((way[2], way[3]))
if (way[2], way[3]) == start:
loop = True
if not loop:
return (bldg_name, [])
chain = []
visited = set()
count = 0
while curr != start or count == 0:
chain.append(curr)
next_nodes = d[curr]
curr = None
for next_node in next_nodes:
if next_node not in visited:
curr = next_node
visited.add(next_node)
break
count += 1
if count > len(ways) or curr is None:
return (bldg_name, [])
return (bldg_name, chain)
def find_current(self,rs):
if isinstance(rs,MergeHistoryNode):
source_node = rs
else:
source_node = self.rsids.get(rs)
if source_node is None:
return None
target = source_node
chain = [target]
# Walk forward from source to sink node
while 1:
next_t = target.child
if next_t is not None:
target = next_t
chain.append(target)
else:
break
# Now walk backward to find first node that isn't deleted
for i in range(len(chain)-1,0,-1):
if not chain[i].deleted:
return chain[i].rsid
def getChainsFromConnections(connections,checkConnections=True):
'''Take a list of connections and return a list of connection chains
connections is a dictionary of connections between elements (which must be hashable)
and can be generated using getElementConnections
The checkConnections option tests that there is only one one path
through each point (aka 2 or fewer connections, no branching)
Returns a list of chains (lists of elements)
'''
connections = deepcopy(connections) # Protect the input from modification
if checkConnections: # Check that there is no branching
assert all( len(v)<3 for k,v in viewitems(connections) ), 'Aborting; this network has branching'
chains = []
while len(connections): # loop over possible chains
# Pick a starting point (an end point if possible)
currPt = _firstOrOther([pt for pt,conn in viewitems(connections)
if len(conn)==1],
next(iter(connections))) # was connections.keys()[0]
# Form a chain and move the current point forward
chain = [currPt]
currPt = connections.pop(currPt)[0]
while currPt: # loop to fill a chain, stop on an invalid
chain.append(currPt)
if len(connections)==0:
break
connections[currPt] = deletecases(connections[currPt], [chain[-2]])
currPt = _firstOrOther(connections.pop(currPt,[]))
chains.append(chain)
return chains
def getChainsFromConnections(connections, checkConnections=True):
'''Take a list of connections and return a list of connection chains
connections is a dictionary of connections between elements (which must be hashable)
and can be generated using getElementConnections
The checkConnections option tests that there is only one one path
through each point (aka 2 or fewer connections, no branching)
Returns a list of chains (lists of elements)
'''
connections = deepcopy(connections) # Protect the input from modification
if checkConnections: # Check that there is no branching
assert all(len(v) < 3 for k, v in viewitems(
connections)), 'Aborting; this network has branching'
chains = []
while len(connections): # loop over possible chains
# Pick a starting point (an end point if possible)
currPt = _firstOrOther(
[pt for pt, conn in viewitems(connections) if len(conn) == 1],
next(iter(connections))) # was connections.keys()[0]
# Form a chain and move the current point forward
chain = [currPt]
currPt = connections.pop(currPt)[0]
while currPt: # loop to fill a chain, stop on an invalid
chain.append(currPt)
if len(connections) == 0:
break
connections[currPt] = deletecases(connections[currPt], [chain[-2]])
currPt = _firstOrOther(connections.pop(currPt, []))
chains.append(chain)
return chains
def walk(self, callback, chain):
res = callback(self, chain)
assert res in (True, False), "Walk function must return boolean, not %r" % (res, )
if not res:
return False
chain.append(self)
self._walk(callback, chain)
chain.pop()
def walk(self, callback, chain):
res = callback(self, chain)
assert res in (True, False), "Walk function must return boolean, not %r" % (res, )
if not res:
return False
chain.append(self)
self._walk(callback, chain)
chain.pop()
def _is_reg_free(self,
addr,
reg,
ignore_current_bb,
level,
total_steps,
debug=False,
prev=None):
if level >= self.cfg_exploration_depth:
raise RegUsed("Max depth %#x %s" % (addr, map(hex, prev)))
if not ignore_current_bb:
if not addr in self.reg_free_map:
# we reached some weird bb
raise RegUsed("Weird block %#x %s" % (addr, map(hex, prev)))
if reg in self.reg_free_map[addr]:
return [addr]
if reg in self.reg_not_free_map[addr]:
raise RegUsed("Not free in bb %#x %s" % (addr, map(hex, prev)))
if prev is None:
prev = []
try:
succ, is_terminate = self.get_all_succ(addr)
# if addr==0x0804B390:
# import IPython; IPython.embed()
if is_terminate:
return [addr]
if len(succ) == 0:
# no successors is weird, the cfg may be incomplete (e.g., NRFIN_00026 original 0x0897F4D5)
raise RegUsed("No successors %#x %s" % (addr, map(hex, prev)))
total_steps[0] += 1
if total_steps[0] >= self.max_cfg_steps:
raise RegUsed("Too many steps %#x %s" %
(addr, map(hex, prev)))
except CfgError:
# something weird is happening in the cfg, let's assume no reg is free
raise RegUsed("CFG error %#x %s" % (addr, map(hex, prev)))
free_regs_in_succ_list = []
chain = []
for s in succ:
if s in prev:
continue # avoid exploring already explored nodes (except the first one).
new_prev = list(prev)
new_prev.append(s)
pchain = self._is_reg_free(s,
reg,
False,
level=level + 1,
total_steps=total_steps,
prev=new_prev,
debug=debug)
chain.append(pchain)
chain.append(addr)
return chain
def chain(self, length=-1, reverse=False, stop_ds=None):
if stop_ds:
# resolve all formats to the same format
stop_ds = Dataset(stop_ds)
chain = []
current = self
while length != len(chain) and current != stop_ds:
chain.append(current)
if not current.previous:
break
current = Dataset(current.previous)
if not reverse:
chain.reverse()
return chain
def aux(node, indent=2, chain=None):
if chain is None:
chain = []
result = ["%s%s" % (" " * indent, node.render())]
children = sorted(self.get_children(node))
children = [aux(c, indent=indent + 2, chain=chain + [c.key]) for c in children if c.key not in chain]
chain.append(node.key)
p = self.packages.get(node.key)
if p:
seen.add(p)
result += list(flatten(children))
return result
def chain(self, length=-1, reverse=False, stop_jobid=None):
if stop_jobid:
# resolve whatever format to the bare jobid
stop_jobid = Dataset(stop_jobid)
if stop_jobid:
stop_jobid = stop_jobid.jobid
chain = []
current = self
while length != len(chain) and current.jobid != stop_jobid:
chain.append(current)
if not current.previous:
break
current = Dataset(current.previous)
if not reverse:
chain.reverse()
return chain
def chain_starting_with(self, ci: CommonInterval) -> List[CommonInterval]:
lvl = self.reverse_index[ci]
chain = []
ci_index = self.nesting_levels[lvl].index(ci)
items_after_ci = self.nesting_levels[lvl][ci_index + 1:]
cur = ci
for other in items_after_ci:
if CommonInterval.intersect(cur, other):
chain.append(other)
cur = other
else:
break
return [ci] + chain if chain else []
def reconstruct_exons(self, path):
# reverse negative stranded data so that all paths go from
# small -> large genomic coords
if self.strand == Strand.NEG:
path.reverse()
# convert from integer node labels to genome (start, end) tuples
path = [self.get_node_interval(nid) for nid in path]
# collapse contiguous nodes along path
newpath = []
chain = [path[0]]
for v in path[1:]:
if chain[-1].end != v.start:
# update path with merge chain node
newpath.append(Exon(chain[0].start, chain[-1].end))
# reset chain
chain = []
chain.append(v)
# add last chain
newpath.append(Exon(chain[0].start, chain[-1].end))
return newpath
def noun_string(data_org):
"""Finds strings of three nouns or more. Returns DataFrame"""
chains = []
tokens = word_tokenize(data_org)
#tokenize to prepare for tagging
w_tag = dict(nltk.pos_tag(tokens))
chain = []
for w, tag in w_tag.items():
#find all nouns based on treebank format
if tag.startswith('N'):
chain.append(w)
else:
if len(chain) >= 3:
chains.append(" ".join(chain))
chain = []
#move information to dataframe for printing to excel
df_noun_string = pd.DataFrame({'Noun Strings (3+ Nouns in a row)': chains},
columns=['Noun Strings (3+ Nouns in a row)'])
return df_noun_string
def get_path_from_parent(self, parent):
"""
Return a list of PathInfos containing the path from the parent
model to the current model, or an empty list if parent is not a
parent of the current model.
"""
if self.model is parent:
return []
model = self.concrete_model
# Get a reversed base chain including both the current and parent
# models.
chain = model._meta.get_base_chain(parent)
chain.reverse()
chain.append(model)
# Construct a list of the PathInfos between models in chain.
path = []
for i, ancestor in enumerate(chain[:-1]):
child = chain[i + 1]
link = child._meta.get_ancestor_link(ancestor)
path.extend(link.get_reverse_path_info())
return path
def reconstruct_exons(self, path):
# reverse negative stranded data so that all paths go from
# small -> large genomic coords
if self.strand == Strand.NEG:
path.reverse()
# convert from integer node labels to genome (start, end) tuples
path = [self.get_node_interval(nid) for nid in path]
# collapse contiguous nodes along path
newpath = []
chain = [path[0]]
for v in path[1:]:
if chain[-1].end != v.start:
# update path with merge chain node
newpath.append(Exon(chain[0].start,
chain[-1].end))
# reset chain
chain = []
chain.append(v)
# add last chain
newpath.append(Exon(chain[0].start, chain[-1].end))
return newpath
def get_path_from_parent(self, parent):
"""
Return a list of PathInfos containing the path from the parent
model to the current model, or an empty list if parent is not a
parent of the current model.
"""
if self.model is parent:
return []
model = self.concrete_model
# Get a reversed base chain including both the current and parent
# models.
chain = model._meta.get_base_chain(parent)
chain.reverse()
chain.append(model)
# Construct a list of the PathInfos between models in chain.
path = []
for i, ancestor in enumerate(chain[:-1]):
child = chain[i + 1]
link = child._meta.get_ancestor_link(ancestor)
path.extend(link.get_reverse_path_info())
return path
def coref(xml, start_id=1):
soup = _soup(xml)
token_by_ids = _token_by_ids(soup)
docs_e = soup.findall('document')
assert len(docs_e) == 1
docs_e = docs_e[0]
# Despite the name, this element contains conferences (note the "s")
corefs_e = docs_e.findall('coreference')
if not corefs_e:
# No coreferences to process
raise StopIteration
assert len(corefs_e) == 1
corefs_e = corefs_e[0]
curr_id = start_id
for coref_e in corefs_e:
if corefs_e.tag != 'coreference':
# To be on the safe side
continue
# This tag is now a full corference chain
chain = []
for mention_e in coref_e.getiterator('mention'):
# Note: There is a "representative" attribute signalling the most
# "suitable" mention, we are currently not using this
# Note: We don't use the head information for each mention
sentence_id = int(mention_e.find('sentence').text)
start_tok_id = int(mention_e.find('start').text)
end_tok_id = int(mention_e.find('end').text) - 1
mention_id = 'T%s' % (curr_id, )
chain.append(mention_id)
curr_id += 1
yield TextBoundAnnotation(
((token_by_ids[sentence_id][start_tok_id].start,
token_by_ids[sentence_id][end_tok_id].end), ),
mention_id, 'Mention', '')
yield EquivAnnotation('Coreference', chain, '')
def calculate(self):
chain = []
path = []
path_cluster = []
cluster = []
dfs = DFS(self.mol)
for bonds in Chem.FindAllSubgraphsOfLengthN(self.mol, self._order):
dfs.reset(bonds)
typ = dfs()
nodes = dfs.nodes
if typ == ChiType.chain:
chain.append(nodes)
elif typ == ChiType.path:
path.append(nodes)
elif typ == ChiType.path_cluster:
path_cluster.append(nodes)
else:
cluster.append(nodes)
return ChiBonds(chain, path, path_cluster, cluster)
def calculate(self):
chain = []
path = []
path_cluster = []
cluster = []
dfs = DFS(self.mol)
for bonds in Chem.FindAllSubgraphsOfLengthN(self.mol, self._order):
dfs.reset(bonds)
typ = dfs()
nodes = dfs.nodes
if typ == ChiType.chain:
chain.append(nodes)
elif typ == ChiType.path:
path.append(nodes)
elif typ == ChiType.path_cluster:
path_cluster.append(nodes)
else:
cluster.append(nodes)
return ChiBonds(chain, path, path_cluster, cluster)
CHANGE_I = [-1, -1, -1, # y or rows
0, 0,
1, 1, 1]
# the border, later drawn around the image
border = []
# saving chain_code as array
chain = []
# iterator for edges
curr_point = start_point
for direction in DIRECTIONS:
idx = DIR_2_IDX[direction]
new_point = (start_point[0]+CHANGE_I[idx], start_point[1]+CHANGE_J[idx])
if img[new_point] != 0: # if is ROI
border.append(new_point)
chain.append(direction)
curr_point = new_point
break
count = 0
while curr_point != start_point:
#figure direction to start search
b_direction = (direction + 5) % 8
dirs_1 = range(b_direction, 8)
dirs_2 = range(0, b_direction)
dirs = []
dirs.extend(dirs_1)
dirs.extend(dirs_2)
for direction in dirs:
idx = DIR_2_IDX[direction]
new_point = (curr_point[0]+CHANGE_I[idx], curr_point[1]+CHANGE_J[idx])
