def hist_from_cascade(
transition_probs,
start_nodes,
stop_nodes=[],
allow_loops=False,
n_init=100,
simultaneous=True,
max_hops=10,
):
dispatcher = TraverseDispatcher(
Cascade,
transition_probs,
allow_loops=allow_loops,
n_init=n_init,
max_hops=max_hops,
stop_nodes=stop_nodes,
)
if simultaneous:
hop_hist = dispatcher.start(start_nodes)
else:
n_verts = len(transition_probs)
hop_hist = np.zeros((n_verts, max_hops))
for s in start_nodes:
hop_hist += dispatcher.start(s)
return hop_hist
matrixplot(A, row_sort_class=labels, col_sort_class=labels, cbar=False)
stashfig("sbm")
# %% [markdown]
# ##
p = 0.05
max_hops = 10
n_init = 100
simultaneous = True
transition_probs = to_transmission_matrix(A, p)
start_nodes = np.arange(n_feedforward)
cdispatch = TraverseDispatcher(
Cascade,
transition_probs,
allow_loops=False,
n_init=n_init,
simultaneous=simultaneous,
)
hit_hist = cdispatch.multistart(start_nodes)
fig, axs = plt.subplots(3,
1,
figsize=(10, 10),
gridspec_kw=dict(height_ratios=[0.45, 0.45, 0.1]))
ax = axs[0]
matrixplot(hit_hist.T, ax=ax, col_sort_class=labels, cbar=True)
ax.set_xlabel("Block")
ax.set_yticks(np.arange(1, max_hops + 1) - 0.5)
ax.set_yticklabels(np.arange(1, max_hops + 1))
ax.set_ylabel("Hops")
np.random.seed(888)
max_hops = 10
n_init = 100
p = 0.05
traverse = Cascade
simultaneous = True
transition_probs = to_transmission_matrix(adj, p)
source_inds = meta[meta[class_key].isin(sg)]["inds"].values
out_inds = meta[meta[class_key].isin(og)]["inds"].values
td = TraverseDispatcher(
traverse,
transition_probs,
n_init=n_init,
simultaneous=simultaneous,
stop_nodes=out_inds,
max_hops=max_hops,
allow_loops=False,
)
fwd_hop_hist = td.multistart(source_inds)
fwd_hop_hist = fwd_hop_hist.T
# backward cascade
td = TraverseDispatcher(
traverse,
transition_probs.T,
n_init=n_init,
simultaneous=simultaneous,
stop_nodes=source_inds,
max_hops=max_hops,
matrixplot(post, ax=ax, col_sort_class=labels)
# %% [markdown]
# ##
max_hops = 10
n_init = 100
# random walk
transition_probs = to_markov_matrix(adj) # row normalize!
traverse = RandomWalk
simultaneous = False
td = TraverseDispatcher(
traverse,
transition_probs,
n_init=n_init,
simultaneous=simultaneous,
stop_nodes=np.arange(n_feedforward * (n_blocks - 1),
n_feedforward * (n_blocks)),
max_hops=max_hops,
allow_loops=False,
)
rw_hop_hist = td.multistart(np.arange(n_feedforward))
rw_hop_hist = rw_hop_hist.T
fig, ax = plt.subplots(1, 1, figsize=(10, 5))
matrixplot(rw_hop_hist, ax=ax, col_sort_class=labels)
# forward cascade
traverse = Cascade
simultaneous = True
td = TraverseDispatcher(
traverse,
transition_probs,