def stashskel(name, ids, colors, palette=None, **kws):
if SAVESKELS:
return export_skeleton_json(name,
ids,
colors,
palette=palette,
foldername=FNAME,
**kws)
sf = signal_flow(adj)
# %% [markdown]
# # Compute signal flow marginals for known cell types
signal_flow_marginal(adj, class_labels)
stashfig("known-class-sf-marginal")
# %% [markdown]
# # Write out signal flow as color for jsons
norm = colors.Normalize(vmin=sf.min(), vmax=sf.max())
sm = ScalarMappable(norm=norm, cmap="Reds")
cmap = sm.to_hex(sf)
export_skeleton_json("signal-flow", skeleton_labels, cmap, palette=None)
# # %% [markdown]
# # #
# node_signal_flow = signal_flow(adj)
# mean_sf = np.zeros(k)
# for i in np.unique(pred_labels):
# inds = np.where(pred_labels == i)[0]
# mean_sf[i] = np.mean(node_signal_flow[inds])
# cluster_mean_latent = gmm.model_.means_[:, 0]
# block_probs = SBMEstimator().fit(bin_adj, y=pred_labels).block_p_
# block_prob_df = pd.DataFrame(data=block_probs, index=range(k), columns=range(k))
# block_g = nx.from_pandas_adjacency(block_prob_df, create_using=nx.DiGraph)
# plt.figure(figsize=(10, 10))
# # don't ever let em tell you you're too pythonic
plt.figure()
fg = sns.FacetGrid(data=out_df,
row="Metric",
col="Method",
margin_titles=True,
height=6)
fg.map(sns.lineplot, "K", "ARI")
plt.suptitle("ARI metrics", fontsize=40, y=1.03, verticalalignment="top")
stashfig("overall-ARI")
plt.figure()
fg = sns.FacetGrid(data=out_df,
col="Method",
margin_titles=True,
height=6,
sharey=False)
fg.map(sns.lineplot, x="K", y="Score")
plt.suptitle("Unsupervised score metrics",
fontsize=40,
y=1.05,
verticalalignment="top")
stashfig("overall-score")
# %% [markdown]
# # Try exporting as JSON for CATMAID
out_path = "./maggot_models/notebooks/outs/30.0-BDP-mw-clustering/test_json.json"
export_skeleton_json(out_path, skeleton_labels, pred_labels, "deep")
# %%