def bfs_test_predecessor(self):
assert (dict(nx.bfs_predecessors(self.G, source=1,
depth_limit=3)) ==
{0: 1, 2: 1, 3: 2, 4: 3, 7: 2, 8: 7})
assert (dict(nx.bfs_predecessors(self.D, source=7,
depth_limit=2)) ==
{2: 7, 3: 2, 8: 7, 9: 8})
def bfs_test_predecessor(self):
assert_equal(dict(nx.bfs_predecessors(self.G, source=1,
depth_limit=3)),
{0: 1, 2: 1, 3: 2, 4: 3, 7: 2, 8: 7})
assert_equal(dict(nx.bfs_predecessors(self.D, source=7,
depth_limit=2)),
{2: 7, 3: 2, 8: 7, 9: 8})
def find_lowest_common_ancestor(graph, a, b):
assert nx.is_directed_acyclic_graph(
graph), "Graph has to be acyclic and directed."
preds_1 = nx.bfs_predecessors(graph, a)
preds_2 = nx.bfs_predecessors(graph, b)
common_preds = set([n for n in preds_1
]).intersection(set([n for n in preds_2]))
if len(list(preds_1)) > 0 and len(common_preds) > 0:
min(common_preds, key=lambda n: preds_1[n])
else:
return -1
def test_limited_bfs_predecessor(self):
assert dict(nx.bfs_predecessors(self.G, source=1, depth_limit=3)) == {
0: 1,
2: 1,
3: 2,
4: 3,
7: 2,
8: 7,
}
assert dict(nx.bfs_predecessors(self.D, source=7, depth_limit=2)) == {
2: 7,
3: 2,
8: 7,
9: 8,
}
def subtree(graph: nx.Graph,
nodes: Iterable[Node],
predecessors: Mapping[Node, Node] = None) -> nx.Graph:
"""Extract a subtree with the specified nodes.
Parameters
----------
graph : nx.Graph
Full tree graph.
nodes : Iterable[Node]
Nodes to include.
predecessors : Mapping[Node, Node], optional
Mapping of node to parent node, by default None
Returns
-------
nx.Graph
A proper tree containing at least the specified nodes.
"""
if predecessors is None:
(root, ), nodes = _peek(iter(nodes))
predecessors = dict(nx.bfs_predecessors(graph, root))
base = set(nodes)
parent_nodes = (parents(node, predecessors) for node in base)
result = set(itertools.chain.from_iterable(parent_nodes))
result.update(base)
return graph.subgraph(result)
def get_bfs(source):
edge_local = dict()
for i in g.edges:
edge_local[tuple(sorted(i))] = 0
bfs_output = nx.Graph(nx.bfs_tree(g, source))
leaf_nodes_list = get_leaf_nodes(bfs_output, source)
parent_nodes_list = list(nx.bfs_predecessors(bfs_output, source))
children_nodes_list = list(nx.bfs_successors(bfs_output, source))
# Initializing the queue with the leaf nodes
queue = list(leaf_nodes_list)
while len(queue) > 0:
current_node = queue.pop(0)
if current_node in leaf_nodes_list:
parents = get_parent_nodes(parent_nodes_list, current_node)
#leaf nodes will have only one parent
temp = (current_node, parents[0])
edge_local[tuple(sorted(temp))] = 1
else:
parents = get_parent_nodes(parent_nodes_list, current_node)
children = get_child_nodes(children_nodes_list, current_node)
forward_val = 1
for c in children:
temp = (current_node, c)
forward_val = forward_val + edge_local[tuple(sorted(temp))]
if len(parents) > 0:
temp = (current_node, parents[0])
edge_local[tuple(sorted(temp))] = forward_val
if len(parents) > 0 and parents[0] not in queue:
queue.append(parents[0])
return edge_local
def find_lca(self, term_a, term_b):
"""
Determine the lowest common ancestor for a two terms
:param term_a: HPO term A.
:param term_b: HPO term B.
:return: Least Common Ancestor for two terms, "HP:0000001"
"""
# if either term is HP:0000001 return it
if any(term == 'HP:0000001' for term in [term_a, term_b]):
return 'HP:0000001'
# if one of the terms is a child of the other return the parent
if self.hpo_network.has_edge(term_a, term_b):
return term_b
if self.hpo_network.has_edge(term_b, term_a):
return term_a
# find common breadth-first-search predecessors
parents = []
for i, term in enumerate([term_a, term_b]):
parents.append(
{p[0] for p in nx.bfs_predecessors(self.hpo_network, term)})
parents[i].add(term)
common_parents = parents[0].intersection(
parents[1])
# lca node
# find the ancestor with the highest IC
# break ties by choosing the node with the greatest depth
return max(common_parents, key=lambda n: (self.hpo_network.nodes[n]['ic'], self.hpo_network.nodes[n]['depth']))
def bfs_edges_modified(tree, source, preference=None):
visited, queue = [], [source]
bfs_tree = nx.bfs_tree(tree, source=source)
predecessors = dict(nx.bfs_predecessors(bfs_tree, source))
edges = []
while queue:
vertex = queue.pop()
if vertex not in visited:
visited.append(vertex)
if (vertex in predecessors.keys()):
edges.append((predecessors[vertex], vertex))
unvisited = set(tree[vertex]) - set(visited)
if (preference != None):
weights = list()
for x in unvisited:
successors = dict(nx.bfs_successors(bfs_tree, source=x))
successors_nodes = list(
itertools.chain.from_iterable(successors.values()))
weights.append(
min([
preference.index(s)
if s in preference else len(preference)
for s in successors_nodes + [x]
]))
unvisited = [
x for _, x in sorted(zip(weights, unvisited),
reverse=True,
key=lambda x: x[0])
]
queue.extend(unvisited)
return edges
def test_predecessor(self):
assert_equal(nx.bfs_predecessors(self.G, source=0), {
1: 0,
2: 1,
3: 1,
4: 2
})
def find_lca(self, term_a, term_b):
"""
Determine the lowest common ancestor for a two terms
:param term_a: HPO term A.
:param term_b: HPO term B.
:return: Least Common Ancestor for two terms, "HP:0000001"
"""
# if either term is HP:0000001 return it
if any(term == 'HP:0000001' for term in [term_a, term_b]):
return 'HP:0000001'
# if one of the terms is a child of the other return the parent
if self.hpo_network.has_edge(term_a, term_b):
return term_b
if self.hpo_network.has_edge(term_b, term_a):
return term_a
# find common breadth-first-search predecessors
parents = []
for term in [term_a, term_b]:
parents.append(
{p[0]
for p in nx.bfs_predecessors(self.hpo_network, term)})
common_parents = parents[0].intersection(parents[1])
# lca node
return max(common_parents,
key=lambda n: self.hpo_network.node[n]['depth'])
def BFS(root, permissions, graph):
common = []
if len(permissions) == 0 or root not in permissions or root in common:
return common
common.append(root)
permissions.remove(root)
nodes = [
u for u, _ in nx.bfs_predecessors(graph, source=root, depth_limit=1)
]
neighbors = sortNeighbor(root, nodes, permissions, graph)
'''
广度优先筛选符合条件的permissions
若满足条件，选取为common
否则， 跳过
'''
#print(neighbors)
for key, values in neighbors:
if freqPermission(graph.graph['name'], key) < 0.05 or weightProportion(
root, key, graph) < 0.2:
break
elif key in permissions:
common.append(key)
permissions.remove(key)
for node in common[1:]:
common.extend(BFS(node, permissions, graph))
return common
def test_predecessor(self):
assert dict(nx.bfs_predecessors(self.G, source=0)) == {
1: 0,
2: 1,
3: 1,
4: 2
}
def calculate_pseudotime(adata):
flat_tree = adata.uns['flat_tree']
dict_edge_len = nx.get_edge_attributes(flat_tree,'len')
adata.obs = adata.obs[adata.obs.columns.drop(list(adata.obs.filter(regex='_pseudotime')))].copy()
# dict_nodes_pseudotime = dict()
for root_node in flat_tree.nodes():
df_pseudotime = pd.Series(index=adata.obs.index)
list_bfs_edges = list(nx.bfs_edges(flat_tree,source=root_node))
dict_bfs_predecessors = dict(nx.bfs_predecessors(flat_tree,source=root_node))
for edge in list_bfs_edges:
list_pre_edges = list()
pre_node = edge[0]
while(pre_node in dict_bfs_predecessors.keys()):
pre_edge = (dict_bfs_predecessors[pre_node],pre_node)
list_pre_edges.append(pre_edge)
pre_node = dict_bfs_predecessors[pre_node]
len_pre_edges = sum([flat_tree.edges[x]['len'] for x in list_pre_edges])
indices = adata.obs[(adata.obs['branch_id'] == edge) | (adata.obs['branch_id'] == (edge[1],edge[0]))].index
if(edge==flat_tree.edges[edge]['id']):
df_pseudotime[indices] = len_pre_edges + adata.obs.loc[indices,'branch_lam']
else:
df_pseudotime[indices] = len_pre_edges + (flat_tree.edges[edge]['len']-adata.obs.loc[indices,'branch_lam'])
adata.obs[flat_tree.node[root_node]['label']+'_pseudotime'] = df_pseudotime
# dict_nodes_pseudotime[root_node] = df_pseudotime
# nx.set_node_attributes(flat_tree,values=dict_nodes_pseudotime,name='pseudotime')
# adata.uns['flat_tree'] = flat_tree
return None
def DFS(root, permissions, graph):
common = []
common.append(root)
permissions.remove(root)
nodes = [
u for u, _ in nx.bfs_predecessors(graph, source=root, depth_limit=1)
]
#print(nodes)
neighbors = sortNeighbor(root, nodes, permissions, graph)
#print(neighbors)
'''
判断排序后的第一个是否满足要求，
若满足，选取该点作为common，
否则，结束
'''
if freqPermission(graph.graph['name'],
neighbors[0][0]) < 0.05 or weightProportion(
root, neighbors[0][0], graph) < 0.2:
return common
elif neighbors[0][0] in permissions:
common.extend(DFS(neighbors[0][0], permissions, graph))
return common
else:
return common
def evaluate(self, learner):
prufer, err = self.decode(learner.subjects)
# print(prufer)
learner.seq = prufer
if err:
learner.fitness = 1e20
return
tree = self.decode_tree(
prufer) # actually not prufer tho, well whatever :)
learner.tree = tree
# cost1 = 0
# for x1,x2 in combinations(range(self.dim),2):
# route = list(nx.all_simple_paths(tree, source=x1, target=x2))[0]
# route_cost = 0
# for x in range(len(route)-1):
# n1 = route[x]
# n2 = route[x+1]
# route_cost = route_cost + self.distance[n1,n2]
# cost1 = cost1 + route_cost
# learner.fitness = cost1
cost = 0
w = np.ones(self.dim)
k = list(nx.bfs_predecessors(tree, 0))
l = dict(k)
# print(k)
h = [a for a, b in k][::-1]
for z in h:
w[l[z]] = w[l[z]] + w[z]
cost = cost + w[z] * (self.dim - w[z]) * self.distance[z, l[z]]
learner.fitness = cost
# print(len(tree.edges))
return cost
def calculate_shift_distance(adata,root='S0',percentile=95, factor=2.0, preference=None):
flat_tree = adata.uns['flat_tree']
dict_label_node = {value: key for key,value in nx.get_node_attributes(flat_tree,'label').items()}
root_node = dict_label_node[root]
##shift distance for each branch
dict_edge_shift_dist = dict()
max_dist = np.percentile(adata.obs['branch_dist'],percentile) ## maximum distance from cells to branch
leaves = [k for k,v in flat_tree.degree() if v==1]
n_nonroot_leaves = len(list(set(leaves) - set([root_node])))
dict_bfs_pre = dict(nx.bfs_predecessors(flat_tree,root_node))
dict_bfs_suc = dict(nx.bfs_successors(flat_tree,root_node))
#depth first search
if(preference != None):
preference_nodes = [dict_label_node[x] for x in preference]
else:
preference_nodes = None
dfs_nodes = dfs_nodes_modified(flat_tree,root_node,preference=preference_nodes)
dfs_nodes_copy = deepcopy(dfs_nodes)
id_leaf = 0
while(len(dfs_nodes_copy)>1):
node = dfs_nodes_copy.pop()
pre_node = dict_bfs_pre[node]
if(node in leaves):
dict_edge_shift_dist[(pre_node,node)] = factor*max_dist*(id_leaf-(n_nonroot_leaves/2.0))
id_leaf = id_leaf+1
else:
suc_nodes = dict_bfs_suc[node]
dict_edge_shift_dist[(pre_node,node)] = (sum([dict_edge_shift_dist[(node,sn)] for sn in suc_nodes]))/float(len(suc_nodes))
return dict_edge_shift_dist
def get_paths_for_nodes_bfs(g, node_list, sparse=False):
# make sure there are no duplicates in node_list
assert len(node_list) == np.unique(node_list).size
num_nodes = len(node_list)
if sparse:
path_mat = scisparse.csr_matrix((num_nodes, num_nodes))
else:
path_mat = np.zeros((num_nodes, num_nodes))
for i in range(num_nodes):
source = node_list[i]
pred_map = dict(nx.bfs_predecessors(g, source))
for j in range(num_nodes):
# skip self loops
if i == j:
continue
target = node_list[j]
# if target is unreachable, leave path lenght as zero
if target not in pred_map:
continue
# traverse back from each target towards the source and count the
# number of steps
while target != source:
target = pred_map[target]
path_mat[i, j] += 1
return path_mat
def get_layers(graph: nx.Graph,
fractions: Iterable[float],
maximum=True,
weight: WeightSelector = 'weight',
algorithm='prim') -> Iterator[nx.Graph]:
"""Extracts multiple layers from a graph.
Parameters
----------
graph
The base graph.
fractions
Percentage of nodes in each layer.
maximum
Whether to use a maximum spanning tree.
weight
Key for the weight attribute.
algorithm
Algorithm used to find spanning tree.
Yields
-------
layers
Each layer as its own graph.
"""
nweight, eweight = _weight_selectors(weight)
tree = graph
if algorithm != 'null':
tree = tree_util.spanning_tree(graph, maximum, eweight, algorithm)
sorted_nodes = tree_util.sorted_nodes(tree, nweight, maximum)
root = sorted_nodes[0]
predecessors = dict(nx.bfs_predecessors(tree, root))
level_counts = dict(
(level, int(frac * len(tree))) for level, frac in enumerate(fractions))
level_counts = {
0: 100,
1: 300,
2: 900,
3: 2700,
4: 8100,
5: 24300,
6: len(sorted_nodes)
}
for level in range(len(level_counts)):
count = level_counts[level]
nodes = _take_n(iter(sorted_nodes), min(len(sorted_nodes), count))
subtree = tree_util.subtree(tree, nodes, predecessors)
for _n, data in subtree.nodes.data():
if _n in sorted_nodes[:count]:
data.setdefault('level', level + 1)
for _u, _v, data in subtree.edges.data():
data.setdefault('level', level + 1)
yield subtree
yield tree
def __init__(self, path, name, tree_graph, root):
'''
Input: treegraph of the VascGraph.GeomGraph.DiGraph class
'''
self.path=path
self.name=name
# check tree
if nx.is_tree(tree_graph): pass
else:
print('Cannot wirte non- tree graph!')
return
def fix_indexing(g_old, root):
bfs=list(nx.bfs_predecessors(treegraph, root))
old_indices=[root]
old_indices.extend([i[0] for i in bfs])
new_indices=range(len(old_indices))
mapping={old_indices[i]:new_indices[i] for i in new_indices}
g=DiGraph()
g.add_nodes_from(new_indices)
for i in old_indices:
g.node[mapping[i]]['pos']=g_old.node[i]['pos']
g.node[mapping[i]]['r']=g_old.node[i]['r']
edges_old=g_old.GetEdges()
edges_new=[[mapping[e[0]], mapping[e[1]]] for e in edges_old]
g.add_edges_from(edges_new)
return g
# fix indixing to start with 0 at root node
treegraph=tree_graph.copy()
treegraph=fix_indexing(treegraph, root)
# get info
pos=treegraph.GetNodesPos()
radii=treegraph.GetRadii()
bfs=dict(nx.bfs_predecessors(treegraph, 0))
e1=treegraph.GetNodes()
e2=[]
for i in e1:
if i==0:
e2.append(-1)
else:
e2.append(bfs[i])
self.__write_graph(e1, e2, pos, radii)
def propogate_score(self, source, score, loaner, already_passed):
self.modify_score(source, score, 0, loaner)
bfs = bfs_predecessors(self, source)
for guy in bfs:
if guy in already_passed:
continue
already_passed.add(guy)
dist = shortest_path_length(self, source, guy)
self.modify_score(guy,score,dist, loaner)
def llf_route(G,origin,destination,nodes_proj):
import networkx as nx
max_depth = len(nx.dijkstra_path(G,origin,destination,weight = 'length'))
#print('Max search depth is '+str(max_depth))
edge_list = list(nx.bfs_predecessors(G, origin, depth_limit=max_depth))
#node_list = list(nx.bfs_tree(G, origin, depth_limit=max_depth))
nt_edges = list(nx.bfs_predecessors(G, origin, depth_limit=1))
#print('Initial edges '+str(nt_edges))
nt_nodes = list(nx.bfs_tree(G, origin, depth_limit=1))
for i in range(len(nt_nodes)-1,len(edge_list)):
if edge_list[i][1] in nt_nodes:
deflection = abs(bearing(G,edge_list[i][0],edge_list[i][1],nodes_proj) - bearing(G,edge_list[i][1],origin,nodes_proj))
if deflection < 45.0 :
nt_nodes.append(edge_list[i][0])
nt_edges.append(edge_list[i])
if destination in nt_nodes:
return nx.dijkstra_path(G,origin,destination,weight = 'length')
last_leg = 1000000000.0
for j in range(0,len(nt_nodes)):
leg = nx.dijkstra_path_length(G,destination,nt_nodes[j],weight = 'length')
if leg < last_leg:
last_leg = leg
route_node = nt_nodes[j]
#print('Nodes are '+str(nt_nodes))
#print('Edges are '+str(nt_edges))
#print('Node with least distance from destination is '+str(route_node))
prev_node = route_node
route = nx.dijkstra_path(G,destination,route_node,weight = 'length')
#print('Last route segment '+str(route))
while prev_node != origin:
for i in range(0,len(nt_edges)):
if nt_edges[i][0] == prev_node:
prev_node = nt_edges[i][1]
route.append(prev_node)
#print('Updated route '+str(route[::-1]))
break
return route[::-1] #reversing the route
def ShowTree(self, lib='HV'):
import networkx as nx
import inspect
G = self.__G__
# reverse graph used to calculate depth of node
H = nx.MultiDiGraph()
H.add_nodes_from(G)
H.add_edges_from([(e[1], e[0]) for e in G.edges])
for n in H.nodes():
lev = set(
[val[1] for val in dict(nx.bfs_predecessors(H, n)).items()])
lev = len(lev)
G.nodes[n]['depth'] = lev
h = {i: 0 for i in range(100)}
for n in G.nodes():
G.nodes[n]['height'] = h[G.nodes[n]['depth']] * 1.5
G.nodes[n]['pos'] = (float(G.nodes[n]['depth']),
float(G.nodes[n]['height']))
h[G.nodes[n]['depth']] = h[G.nodes[n]['depth']] + 1
if lib.lower() == 'matplotlib':
import matplotlib.pyplot as plt
plt.figure(figsize=(6, 6))
#pos= [key for key in nx.get_node_attributes(G,'pos').keys()] # nx.spring_layout(G,scale=2)
pos = nx.get_node_attributes(G,
'pos') # nx.spring_layout(G,scale=2)
color_map = [G.nodes[g]['color'] for g in G.nodes]
nx.draw(G,
pos,
node_color=color_map,
with_labels=True,
node_size=1000,
connectionstyle='arc3, rad = 0.1')
if lib.lower() == 'hv':
import holoviews as hv
hv.extension('bokeh')
graph = hv.Graph.from_networkx(
G, nx.layout.fruchterman_reingold_layout).opts(
width=800,
height=400,
xaxis=None,
yaxis=None,
legend_position='top_left',
directed=True,
node_size=50,
inspection_policy='edges',
arrowhead_length=0.01,
node_color='color')
labels = hv.Labels(graph.nodes, ['x', 'y'], 'name')
return graph * labels
def straighten_edges(graph):
"""Merges all nodes which only have 1 child with their parent"""
predecessors = dict(nx.bfs_predecessors(graph, "0"))
only_single_successor = [(n, *s) for n, s in nx.bfs_successors(graph, "0")
if len(s) == 1]
nodes_to_be_removed = []
cant_be_removed = {"0"}
for node, successor in only_single_successor:
if node not in cant_be_removed:
predecessor = predecessors[node]
nodes_to_be_removed.append(node)
merge_edges(graph, predecessor, node, successor)
cant_be_removed.add(successor)
remove_nodes(graph, nodes_to_be_removed)
def main():
# create graph from problem statement input
# keeping track of where the solution location should be in the graph
(original_graph, solution_location) = makeGraph("input.txt")
#draw our graph 'corr'
drawVanillaGraph(original_graph,
pos=nx.kamada_kawai_layout(original_graph))
# refurbish graph to better fit our needs
# our needs are to find if a path exists to 'end' from the starting location
# making new graph R(efurbished) U(ltimate) G(raph)
RUG = refurbish(original_graph, solution_location, 0, 1)
drawRefurbishedGraph(RUG, pos=nx.fruchterman_reingold_layout(RUG))
# now print a bfs traversal of the refurbished graph
# bfs predecessors and parents stored in y
# Rocket starts at 0 (A) and Lucky starts at 1 (B)
y = list(nx.bfs_predecessors(RUG, State(R=0, L=1)))
# now traverse bfs tree and output solution (or lack thereof)
traverse_bfs_tree(y, solution_location)
def dependencies(self, symbol):
"""Find all dependencies of a symbol"""
for i in range(len(self) - 1, -1, -1):
if (isinstance(self[i], Assignment) and self[i].symbol == symbol
or isinstance(self[i], ODESystem)
and symbol in self[i].amounts):
break
else:
raise KeyError(f"Could not find symbol {symbol}")
g = self._create_dependency_graph()
symbs = self[i].rhs_symbols
for j, _ in nx.bfs_predecessors(
g, i, sort_neighbors=lambda x: reversed(sorted(x))):
if isinstance(self[j], Assignment):
symbs -= {self[j].symbol}
elif isinstance(self[j], ODESystem):
symbs -= set(self[j].amounts)
symbs |= self[j].rhs_symbols
return symbs
def fix_indexing(g_old, root):
bfs=list(nx.bfs_predecessors(treegraph, root))
old_indices=[root]
old_indices.extend([i[0] for i in bfs])
new_indices=range(len(old_indices))
mapping={old_indices[i]:new_indices[i] for i in new_indices}
g=DiGraph()
g.add_nodes_from(new_indices)
for i in old_indices:
g.node[mapping[i]]['pos']=g_old.node[i]['pos']
g.node[mapping[i]]['r']=g_old.node[i]['r']
edges_old=g_old.GetEdges()
edges_new=[[mapping[e[0]], mapping[e[1]]] for e in edges_old]
g.add_edges_from(edges_new)
return g
def main():
g = nx.Graph()
g.add_edge(0, 1)
g.add_edge(0, 2)
g.add_edge(3, 1)
g.add_edge(4, 2)
print("Start search")
bfs_res = nx.bfs_predecessors(g, 0)
for x in bfs_res:
print(x)
print("Start search")
bfs_res = nx.bfs_successors(g, 0)
for x in bfs_res:
print(x)
print("Start search")
bfs_res = nx.bfs_tree(g, 0, depth_limit=1)
for x in bfs_res:
print(x)
nx.draw(g)
plt.show()
def get_dependent_modules():
print('\n===Dependent Modules===')
for node_name in G.nodes_iter():
dependents = nx.bfs_predecessors(G, node_name)
if len(dependents):
print(dependents)
def tree_part2(graph, pop_col, pop_target, epsilon,node_repeats):
w=graph.copy()
for ed in w.edges():
w.add_edge(ed[0],ed[1],weight=random.random())
T = tree.maximum_spanning_edges(w, algorithm='kruskal', data=False)
ST= nx.Graph()
ST.add_edges_from(list(T))
#nx.draw(ST)
h=ST.copy()
#nx.draw(ST,layout='tree')
#root = random.choice(list(h.nodes()))
root = random.choice([x for x in ST.nodes() if ST.degree(x)>1])#this used to be greater than 2 but failed on small grids:(
#print(root)
predbfs=nx.bfs_predecessors(h, root)#was dfs
pred={}
for ed in predbfs:
pred[ed[0]]=ed[1]
pops={x:[{x},graph.nodes[x][pop_col]] for x in graph.nodes()}
leaves=[]
t=0
layer=0
restarts=0
while 1==1:
if restarts==node_repeats:
w=graph.copy()
for ed in w.edges():
w.add_edge(ed[0],ed[1],weight=random.random())
T = tree.maximum_spanning_edges(w, algorithm='kruskal', data=False)
ST= nx.Graph()
ST.add_edges_from(list(T))
#nx.draw(ST)
h=ST.copy()
#nx.draw(ST,layout='tree')
#root = random.choice(list(h.nodes()))
root = random.choice([x for x in ST.nodes() if ST.degree(x)>1])#this used to be greater than 2 but failed on small grids:(
#print(root)
predbfs=nx.bfs_predecessors(h, root)#was dfs
pred={}
for ed in predbfs:
pred[ed[0]]=ed[1]
pops={x:[{x},graph.nodes[x][pop_col]] for x in graph.nodes()}
leaves=[]
t=0
layer=0
restarts=0
#print("Bad tree -- rebuilding")
if len(list(h.nodes()))==1:
h=ST.copy()
root = random.choice([x for x in ST.nodes() if ST.degree(x)>1])#this used to be greater than 2 but failed on small grids:(
#print(root)
#pred=nx.bfs_predecessors(h, root)#was dfs
predbfs=nx.bfs_predecessors(h, root)#was dfs
pred={}
for ed in predbfs:
pred[ed[0]]=ed[1]
pops={x:[{x},graph.nodes[x][pop_col]] for x in graph.nodes()}
#print("bad root --- restarting",restarts)
restarts+=1
layer=0
leaves=[]
if leaves == []:
leaves = [x for x in h.nodes() if h.degree(x)==1]
layer=layer+1
if len(leaves) == len(list(h.nodes()))-1:
tsum = pops[root][1]
for r in range(2,len(leaves)):
for s in itertools.combinations(leaves,r):
for node in s:
tsum+=pops[node][1]
if abs(tsum-pop_target)<epsilon*pop_target:
print(pops[leaf][1]/pop_target)
clusters={}
clusters[1]=list(pops[leaf][0])
clusters[-1]=[]
for nh in graph.nodes():
if nh not in clusters[1]:
clusters[-1].append(nh)
return clusters
if root in leaves: #this was in an else before but is still apparently necessary?
leaves.remove(root)
#if layer %10==0:
#print("Layer",layer)
for leaf in leaves:
if layer>1 and abs(pops[leaf][1]-pop_target) < pop_target * epsilon:
#print(pops[leaf][1]/pop_target)
#ST.remove_edge(leaf,pred[leaf])#One option but slow
#parts=list(nx.connected_components(h)) #ST here too
#print(layer, len(parts))
#part=parts[random.random()<.5]
clusters={}
clusters[1]=list(pops[leaf][0])
clusters[-1]=[]
for nh in graph.nodes():
if nh not in clusters[1]:
clusters[-1].append(nh)
return clusters
parent = pred[leaf]
pops[parent][1]+=pops[leaf][1]
pops[parent][0]=pops[parent][0].union(pops[leaf][0])
#h = nx.contracted_edge(h,(parent,leaf),self_loops = False)#too slow on big graphs
h.remove_node(leaf)
leaves.remove(leaf)
t=t+1
def _run_bfswpf(ppc, options, **kwargs):
"""
SPARSE version of distribution power flow solution according to [1]
:References:
[1] Jen-Hao Teng, "A Direct Approach for Distribution System Load Flow Solutions", IEEE Transactions on Power Delivery, vol. 18, no. 3, pp. 882-887, July 2003.
:param ppc: matpower-style case data
:return: results (pypower style), success (flag about PF convergence)
"""
time_start = time() # starting pf calculation timing
tap_shift = ppc['branch'][:, SHIFT].real
enforce_q_lims, tolerance_kva, max_iteration, calculate_voltage_angles, numba = _get_options(
options)
numba, makeYbus = _import_numba_extensions_if_flag_is_true(numba)
ppci = ppc
baseMVA, bus, gen, branch = \
ppci["baseMVA"], ppci["bus"], ppci["gen"], ppci["branch"]
nbus = bus.shape[0]
# generate results for original bus ordering
Ybus, Yf, Yt = makeYbus(baseMVA, bus, branch)
# get bus index lists of each type of bus
ref, pv, pq = bustypes(bus, gen)
# creating networkx graph from list of branches
G = nx.Graph()
G.add_edges_from((int(fb), int(tb), {
"shift": float(shift)
}) for fb, tb, shift in list(
zip(branch[:, F_BUS].real, branch[:, T_BUS].real, tap_shift)))
if not nx.is_connected(G):
Graphs = list(nx.connected_component_subgraphs(G))
else:
Graphs = [G]
V_final = np.zeros(nbus, dtype=complex)
V_tapshifts = np.zeros(nbus)
for subi, G in enumerate(Graphs):
ppci_sub = _cut_ppc(ppci, G.nodes())
nbus_sub = len(G)
# depth-first-search bus ordering and generating Direct Load Flow matrix DLF = BCBV * BIBC
DLF, ppc_bfsw, buses_ordered_bfsw = _bibc_bcbv(ppci_sub, G)
ppc_bfsw['branch'][:, SHIFT] = 0
baseMVA_bfsw, bus_bfsw, gen_bfsw, branch_bfsw, ref_bfsw, pv_bfsw, pq_bfsw,\
on, gbus, V0 = _get_pf_variables_from_ppci(ppc_bfsw)
Sbus_bfsw = makeSbus(baseMVA_bfsw, bus_bfsw, gen_bfsw)
Ybus_bfsw, Yf_bfsw, Yt_bfsw = makeYbus(baseMVA_bfsw, bus_bfsw,
branch_bfsw)
# #----- run the power flow -----
V_res, success = _bfswpf(DLF, bus_bfsw, gen_bfsw, branch_bfsw, baseMVA,
Ybus_bfsw, buses_ordered_bfsw, Sbus_bfsw, V0,
ref_bfsw, pv_bfsw, pq_bfsw, enforce_q_lims,
tolerance_kva, max_iteration, **kwargs)
V_final[
buses_ordered_bfsw] = V_res # return bus voltages in original bus order
# TODO: find the better way to consider transformer phase shift and remove this workaround
if calculate_voltage_angles:
predecessors = nx.bfs_predecessors(G, ref[subi])
branches = list(zip(branch[:, F_BUS].real, branch[:, T_BUS].real))
for bus_start in predecessors.iterkeys():
bus_pred = bus_start
bus_next = bus_start
while predecessors.get(bus_next) is not None:
bus_next = predecessors.get(bus_pred)
shift_angle = G.get_edge_data(bus_pred, bus_next)['shift']
if (bus_pred, bus_next) in branches:
V_tapshifts[bus_start] += shift_angle
else:
V_tapshifts[bus_start] -= shift_angle
bus_pred = bus_next
V_final *= np.exp(1j * np.pi / 180 * V_tapshifts)
# #----- output results to ppc ------
ppci["et"] = time() - time_start # pf time end
bus, gen, branch = pfsoln(baseMVA, bus, gen, branch, Ybus, Yf, Yt, V_final,
ref, pv, pq)
ppci["success"] = success
ppci["bus"], ppci["gen"], ppci["branch"] = bus, gen, branch
return ppci, success
def run(self, G): s = random.randint(0, G.number_of_nodes()) networkx.bfs_predecessors(G, s)
def _all_predecessors(graph, nodes):
return [k for node in nodes for k in nx.bfs_predecessors(graph, node).keys()]
def predecessors(h, root):
return {a: b for a, b in nx.bfs_predecessors(h, root)}
def get_dependent_modules():
print('\n===Dependent Modules===')
for node_name in G.nodes():
dependents = nx.bfs_predecessors(G, node_name)
if len(dependents):
print(dependents)
def bfs(graph, root, max_depth):
"""
Perform breadth-first search to compute the shortest paths from a root node to all
other nodes in the graph. To reduce running time, the max_depth parameter ends
the search after the specified depth.
E.g., if max_depth=2, only paths of length 2 or less will be considered.
This means that nodes greather than max_depth distance from the root will not
appear in the result.
You may use these two classes to help with this implementation:
https://docs.python.org/3.5/library/collections.html#collections.defaultdict
https://docs.python.org/3.5/library/collections.html#collections.deque
Params:
graph.......A networkx Graph
root........The root node in the search graph (a string). We are computing
shortest paths from this node to all others.
max_depth...An integer representing the maximum depth to search.
Returns:
node2distances...dict from each node to the length of the shortest path from
the root node
node2num_paths...dict from each node to the number of shortest paths from the
root node that pass through this node.
node2parents.....dict from each node to the list of its parents in the search
tree
In the doctests below, we first try with max_depth=5, then max_depth=2.
>>> node2distances, node2num_paths, node2parents = bfs(example_graph(), 'E', 5)
>>> sorted(node2distances.items())
[('A', 3), ('B', 2), ('C', 3), ('D', 1), ('E', 0), ('F', 1), ('G', 2)]
>>> sorted(node2num_paths.items())
[('A', 1), ('B', 1), ('C', 1), ('D', 1), ('E', 1), ('F', 1), ('G', 2)]
>>> sorted((node, sorted(parents)) for node, parents in node2parents.items())
[('A', ['B']), ('B', ['D']), ('C', ['B']), ('D', ['E']), ('F', ['E']), ('G', ['D', 'F'])]
>>> node2distances, node2num_paths, node2parents = bfs(example_graph(), 'E', 2)
>>> sorted(node2distances.items())
[('B', 2), ('D', 1), ('E', 0), ('F', 1), ('G', 2)]
>>> sorted(node2num_paths.items())
[('B', 1), ('D', 1), ('E', 1), ('F', 1), ('G', 2)]
>>> sorted((node, sorted(parents)) for node, parents in node2parents.items())
[('B', ['D']), ('D', ['E']), ('F', ['E']), ('G', ['D', 'F'])]
"""
###TODO
node2distances = {}
node2num_paths = {}
node2parents = {}
q = deque()
q.append(root)
seen = set()
res = []
noOf_shortest_path = []
count = 0
while len(q) > 0:
n = q.popleft()
if n not in seen:
res.append(n)
seen.add(n)
count += 1
for nn in graph.neighbors(n):
if nn not in seen:
q.append(nn)
paths = []
for i in range(len(res)):
paths.append(nx.single_source_shortest_path(graph, res[i]))
for i in range(len(paths)):
for k, v in paths[i].items():
if k == root:
node2distances[res[i]] = len(v)
node2distances = sorted(node2distances.items())
#print(node2distances)
n_spaths = dict.fromkeys(graph, 0.0)
s_paths = nx.all_pairs_shortest_path(graph)
for source in graph:
for path in s_paths[source].values():
for node in path[1:]: # ignore firs element (source == node)
n_spaths[node] += 1 # this path passes through `node`
j = 0
bak = []
for i in range(len(res)):
for k, v in s_paths.items():
if k == res[i]:
node2num_paths[res[i]] = len(v)
node2num_paths = sorted(node2num_paths.items())
#print(node2num_paths)
l = []
for i in range(len(res)):
pred = nx.bfs_predecessors(graph, res[i])
for i in range(len(res)):
for k, v in pred.items():
if k == res[i]:
node2parents[res[i]] = v
l.append(v)
blah = []
for k, v in node2parents.items():
temp = tuple((k, list([v])))
blah.append(temp)
node2parents = blah
#print(node2parents)
gaandu = bottom_up(root, node2distances, node2num_paths, node2parents)
return gaandu
pass
def test_predecessor(self):
assert_equal(nx.bfs_predecessors(self.G,source=0),
{1: 0, 2: 1, 3: 1, 4: 2})
