def plot_mvmm(mvmm, inches=8, save_dir=None):
"""
Plots loss history and estimated Pi matrix.
"""
if save_dir is not None:
os.makedirs(save_dir, exist_ok=True)
# TODO: maybe add Pi start
loss_vals = mvmm.opt_data_['history']['loss_val']
################
# Loss history #
################
plot_loss_history(loss_vals,
loss_name='Observed data negative log-likelihood')
if save_dir is not None:
fpath = join(save_dir, 'loss_history.png')
save_fig(fpath)
###############
# Pi estimate
################
plt.figure(figsize=(inches, inches))
plot_Pi(mvmm.weights_mat_)
plt.title("Estimated Pi")
if save_dir is not None:
fpath = join(save_dir, 'Pi_est.png')
save_fig(fpath)
def plot_model_selection(mm_gs,
save_dir=None,
name_stub=None,
title='',
inches=8):
n_comps = mm_gs.param_grid['n_components']
# best_n_comp = mm_gs.best_params_['n_components']
select_metric = mm_gs.select_metric
# for sel_metric in mm_gs.model_sel_scores_.columns:
scores = mm_gs.model_sel_scores_[select_metric]
if MEASURE_MIN_GOOD[select_metric]:
sel_n_comp = n_comps[scores.idxmin()]
else:
sel_n_comp = n_comps[scores.idxmax()]
# get selected number of components
# mm_gs.select_metric = sel_metric
# sel_n_comp = mm_gs.best_params_['n_components']
# mm_gs.select_metric = orig_sel_metric
plt.figure(figsize=(inches, inches))
plt.plot(n_comps, scores, marker='.') # , label='bic')
plt.axvline(sel_n_comp,
color='black',
ls='--',
label='Est. number of components = {}'.format(sel_n_comp))
set_xaxis_int_ticks()
plt.legend()
plt.title(title)
plt.ylabel(select_metric)
plt.xlabel("Number of components")
if save_dir is not None:
fname = '{}_model_selection.png'.format(select_metric)
if name_stub is not None:
fname = name_stub + '_' + fname
save_fig(os.path.join(save_dir, fname))
ephys_raw=ephys_raw)
clust_labels = [
'cluster_{}'.format(cl_idx + 1) for cl_idx in range(len(y_cnts))
]
# plot top several clusters
plot_top_clust_ephys_curves(cluster_super_means,
y_cnts=y_cnts,
overall_means=super_data_means,
overall_stds=super_data_stds,
clust_labels=clust_labels,
n_to_show=n_top_clust,
inches=inches)
save_fig(join(ephys_viz_dir, 'ephys_curves_top_clust.png'))
# plot each (non-trival) cluster
# non_trivial_clusters = y_cnts[y_cnts >= 5].index.values
non_trivial_clusters = y_cnts[y_cnts >= 0].index.values
save_dir = make_and_get_dir(ephys_viz_dir, 'cluster_curves')
for cl_idx in non_trivial_clusters:
label = clust_labels[cl_idx]
values = {}
for name in cluster_super_means.keys():
values[name] = cluster_super_means[name][cl_idx]
plt.figure(figsize=(2 * n_datasets * inches, inches))
###############
# joint level #
###############
# cluster prediction data
for k in ['joint_summary', 'y_pred_joint', 'joint_clust_best_samples']:
out[k].to_csv(join(save_dir, '{}.csv'.format(k)))
# metadata comparisons
if 'joint_comparison' in out.keys():
dump(out['joint_comparison'], join(save_dir, 'metadata_comparison'))
plt.figure(figsize=(2 * inches, 2 * inches))
out['joint_comparison'].plot()
save_fig(join(save_dir, 'metadata_comparison.png'))
# survival
if 'joint_survival' in out.keys():
pval = out['joint_survival']['pval']
plt.figure(figsize=(inches, inches))
plot_survival(df=out['joint_survival']['df'], cat_col='cluster')
plt.xlabel("Time (days)")
plt.ylabel("Survival")
plt.title('{} vs. joint label, p={:1.3f}'.format(args.event_col, pval))
save_fig(join(save_dir, 'joint_survival.png'))
# TODO: label survival variable!
##############
# View level #
##############
transc_sel_markers = transc[select_markers]
transc_sel_markers.\
to_csv(join(pro_data_dir, 'transcriptomic_select_markers.csv'))
# PCA
pca = PCA(n_components='rmt_threshold', rank_sel_kws={'thresh_method': 'dg'})
pca.fit(transc_sel_markers.values)
# save scores
UD = pca.unnorm_scores_
UD = pd.DataFrame(UD,
index=transc_sel_markers.index,
columns=['pc_{}'.format(k + 1) for k in range(UD.shape[1])])
UD.index.name = transc_sel_markers.index.name
UD.name = 'transcriptomic_select_markers'
UD.to_csv(
join(pro_data_dir,
'{}_pca_feats.csv'.format('transcriptomic_select_markers')))
# save scree plot
plt.figure(figsize=(8, 8))
scree_plot(pca.all_svals_, color='black')
plt.axvline(pca.n_components_,
label='{}'.format(pca.n_components_),
color='red')
plt.legend()
save_fig(
join(MouseETPaths().results_dir,
'{}_rank_selection.png'.format('transcriptomic_select_markers')))
best_row['n_blocks_est'],
best_row['n_view_comp']))
mvmm_sel_df.to_csv(os.path.join(bd_sel_dir, 'bd_mvmm_model_sel.csv'))
# metrics2compute = mvmm_results['fit_data'][0]['models']['bd_mvmm'].metrics2compute
metrics2compute = [args.select_metric]
for metric in metrics2compute:
plt.figure(figsize=(inches, inches))
# plot_bd_mvmm_model_sel(mvmm_sel_df, select_metric=metric)
plot_mvmm_model_selection(mvmm_sel_df, group_var='n_blocks_est',
group_var_label="Number of blocks",
select_metric=metric)
# plt.title('Multiview models')
save_fig(join(bd_sel_dir, 'bd_mvmm_model_sel_{}.png'.format(metric)))
# fitting history
# if isinstance(estimator, TwoStage):
if estimator.start_.max_n_steps > 0:
bd_start_save_dir = make_and_get_dir(opt_diag_dir,
'bd_mvmm', 'start')
bd_final_save_dir = make_and_get_dir(opt_diag_dir,
'bd_mvmm', 'final')
plot_mvmm(estimator.start_, inches=inches,
save_dir=bd_start_save_dir)
else:
bd_final_save_dir = make_and_get_dir(opt_diag_dir, 'bd_mvmm')
plt.axhline(thresh_dg, label='DG theshold ({:1.2f})'.format(thresh_dg),
color='blue', ls='--')
plt.axhline(thresh_mpe, label='MPE theshold ({:1.2f})'.format(thresh_mpe),
color='orange', ls='--')
plt.axvline(rank_est_dg,
label='DG rank estimate ({})'.format(rank_est_dg),
color='blue')
plt.axvline(rank_est_mpe,
label='MPE rank estimate ({})'.format(rank_est_mpe),
color='orange')
plt.legend()
plt.title(k)
save_fig(join(diagnostics_dir, '{}_scree_plot.png'.format(k)))
# # save scree plot
# plt.figure(figsize=(10, 10))
# scree_plot(pca.all_svals_, color='black')
# plt.axvline(pca.n_components_,
# label='{}'.format(pca.n_components_),
# color='red')
# plt.legend()
# plt.ylim(0)
# save_fig(join(diagnostics_dir, '{}_rank_selection.png'.format(k)))
# # for bi cross validation save MSE curve
# if pca.n_components == 'bi_cv':
# plt.figure(figsize=(10, 10))
# mse = pca.rank_sel_out_['errors']
def plot_log_pen_mvmm(mvmm, inches=8, save_dir=None):
if save_dir is not None:
os.makedirs(save_dir, exist_ok=True)
# info = get_log_pen_mvmm_info(mvmm)
# if save_dir is not None:
# # TODO: save this
# save_dir
# else:
# print(info)
Pi = mvmm.weights_mat_
zero_thresh = 1e-10 # not sure if we need this
summary, Pi_comm = community_summary(Pi, zero_thresh=zero_thresh)
Pi_symlap_spec = eigh_Lsym_bp(Pi)[0]
if 'init_params' in mvmm.opt_data_['history']:
Pi_init = mvmm.opt_data_['history']['weights'].reshape(
Pi.shape) # TODO: check this
Pi_init_symlap_spec = eigh_Lsym_bp(Pi_init)[0]
else:
Pi_init = None
obs_nll = mvmm.opt_data_['history']['obs_nll']
loss_vals = mvmm.opt_data_['history']['loss_val']
####################
# Initial weights #
###################
if Pi_init is not None:
plt.figure(figsize=(inches, inches))
plot_Pi(Pi_init)
plt.title('weights initial value')
if save_dir is not None:
fpath = join(save_dir, 'weights_init.png')
save_fig(fpath)
######################
# Estimated weights #
######################
plt.figure(figsize=(2 * inches, inches))
plt.subplot(1, 2, 1)
plot_Pi(Pi)
plt.title('weights estimate, n_blocks={}'.format(summary['n_communities']))
plt.subplot(1, 2, 2)
plot_Pi(Pi, mask=Pi_comm < zero_thresh)
plt.title('weights estimate, block diagonal perm')
if save_dir is not None:
fpath = join(save_dir, 'weights_est.png')
save_fig(fpath)
##########################
# Spectrum of BD weights #
##########################
plt.figure(figsize=(inches, inches))
idxs = np.arange(1, len(Pi_symlap_spec) + 1)
plt.plot(idxs, Pi_symlap_spec, marker='.', label='Estimate')
if Pi_init is not None:
plt.plot(idxs, Pi_init_symlap_spec, marker='.', label="Initial")
plt.title('weights estimate spectrum')
plt.ylim(0)
plt.legend()
if save_dir is not None:
fpath = join(save_dir, 'weights_spectrum.png')
save_fig(fpath)
###########################
# Obs NLL for entire path #
###########################
plt.figure(figsize=[inches, inches])
plot_loss_history(obs_nll, loss_name="Obs NLL")
if save_dir is not None:
fpath = join(save_dir, 'obs_nll.png')
save_fig(fpath)
plt.figure(figsize=[inches, inches])
plot_loss_history(loss_vals, loss_name="log penalized obs nll")
if save_dir is not None:
fpath = join(save_dir, 'loss_vals.png')
save_fig(fpath)
def plot_bd_mvmm(mvmm, inches=8, save_dir=None):
"""
Initial BD weights, Estimated BD weights, spectrums of both
Number of steps in each adaptive stage
Evals of entire path
Loss history for each segment
"""
if save_dir is not None:
os.makedirs(save_dir, exist_ok=True)
info = get_bd_mvmm_info(mvmm)
if save_dir is not None:
# TODO: save this
save_dir
else:
print(info)
# BD weight estimate
bd_weights = mvmm.bd_weights_
zero_thresh = mvmm.zero_thresh
summary, Pi_comm = community_summary(bd_weights, zero_thresh=zero_thresh)
bd_weights_symlap_spec = eigh_Lsym_bp(bd_weights)[0]
# initial BD weights
bd_weights_init = mvmm.opt_data_['adpt_opt_data']['adapt_pen_history'][
'opt_data'][0]['history']['init_params']['bd_weights']
bd_weights_init_symlap_spec = eigh_Lsym_bp(bd_weights_init)[0]
# optimization history
adpt_history = mvmm.opt_data_['adpt_opt_data']['adapt_pen_history'][
'opt_data']
if 'raw_eval_sum' in adpt_history[0]['history']:
n_steps = [
len(adpt_history[i]['history']['raw_eval_sum'])
for i in range(len(adpt_history))
]
n_steps_cumsum = np.cumsum(n_steps)
raw_eval_sum = \
np.concatenate([adpt_history[i]['history']['raw_eval_sum']
for i in range(len(adpt_history))])
else:
raw_eval_sum = None
n_steps = None
n_steps_cumsum = None
obs_nll = np.concatenate([
adpt_history[i]['history']['obs_nll'] for i in range(len(adpt_history))
])
if mvmm.opt_data_['ft_opt_data'] is not None:
fine_tune_obs_nll = mvmm.opt_data_['ft_opt_data']['history']['obs_nll']
else:
fine_tune_obs_nll = None
######################
# Initial BD weights #
######################
plt.figure(figsize=(inches, inches))
plot_Pi(bd_weights_init)
plt.title('BD weights initial value')
if save_dir is not None:
fpath = join(save_dir, 'BD_weights_init.png')
save_fig(fpath)
########################
# Estimated BD weights #
########################
plt.figure(figsize=(2 * inches, inches))
plt.subplot(1, 2, 1)
plot_Pi(bd_weights)
plt.title('BD weights estimate, n_blocks={}'.format(
summary['n_communities']))
plt.subplot(1, 2, 2)
plot_Pi(bd_weights, mask=Pi_comm < zero_thresh)
plt.title('BD weights estimate, block diagonal perm')
if save_dir is not None:
fpath = join(save_dir, 'BD_weights_est.png')
save_fig(fpath)
##########################
# Spectrum of BD weights #
##########################
plt.figure(figsize=(inches, inches))
idxs = np.arange(1, len(bd_weights_symlap_spec) + 1)
plt.plot(idxs, bd_weights_symlap_spec, marker='.', label='Estimate')
plt.plot(idxs, bd_weights_init_symlap_spec, marker='.', label="Initial")
plt.title('BD weights estimate spectrum')
plt.ylim(0)
plt.legend()
if save_dir is not None:
fpath = join(save_dir, 'BD_weights_spectrum.png')
save_fig(fpath)
##################################
# Number of steps for each stage #
##################################
if n_steps is not None:
plt.figure(figsize=(inches, inches))
idxs = np.arange(1, len(n_steps) + 1)
plt.plot(idxs, n_steps, marker='.')
plt.ylim(0)
plt.ylabel("Number of steps")
plt.xlabel("Adaptive stage")
if save_dir is not None:
fpath = join(save_dir, 'n_steps.png')
save_fig(fpath)
###########################
# Obs NLL for entire path #
###########################
plt.figure(figsize=[inches, inches])
plot_loss_history(obs_nll, loss_name="Obs NLL (entire path)")
if save_dir is not None:
fpath = join(save_dir, 'path_obs_nll.png')
save_fig(fpath)
#########################
# Evals for entire path #
#########################
if raw_eval_sum is not None:
plt.figure(figsize=[inches, inches])
plt.plot(np.log10(raw_eval_sum), marker='.')
plt.ylabel('log10(sum smallest evals)')
plt.xlabel('step')
plt.title('Eigenvalue history (entire path)')
for s in n_steps_cumsum:
plt.axvline(s - 1, color='grey')
if save_dir is not None:
fpath = join(save_dir, 'path_evals.png')
save_fig(fpath)
###########################
# Losses for each segment #
###########################
if save_dir is not None:
segment_dir = join(save_dir, 'segments')
os.makedirs(segment_dir, exist_ok=True)
for i in range(len(adpt_history)):
loss_vals = adpt_history[i]['history']['loss_val']
plot_loss_history(loss_vals,
'loss val, adapt segment {}'.format(i + 1))
if save_dir is not None:
fpath = join(segment_dir, 'loss_history_{}.png'.format(i + 1))
save_fig(fpath)
##########################
# fine tune loss history #
##########################
if fine_tune_obs_nll is not None:
plot_loss_history(fine_tune_obs_nll, 'fine tune obs NLL')
if save_dir is not None:
fpath = join(segment_dir, 'fine_tune_loss_history.png')
save_fig(fpath)
# additional results #
######################
n_blocks_true = community_summary(Pi_true)[0]['n_communities']
n_comp_true = (Pi_true > 0).sum()
###########
# Pi true #
###########
plt.figure()
sns.heatmap(Pi_true, cmap='Blues', square=True, cbar=True, vmin=0)
plt.xlabel('View 1 clusters')
plt.ylabel('View 2 clusters')
plt.title("Pi True")
save_fig(os.path.join(save_dir, 'Pi_true.png'))
if 'bd_mvmm' not in to_exclude:
#########################
# Block diagonal Pi Viz #
#########################
bd_mvmm = models['bd_mvmm'].final_
print('\n\n\n BD summary\n\n\n')
summarize_bd(D=bd_mvmm.bd_weights_,
n_blocks=n_blocks_true,
zero_thresh=zero_thresh)
plt.title('BD estimate')
save_fig(os.path.join(save_dir, 'Pi_est_bd_mvmm.png'))
'fit_time': None,
'n_comp_est': None,
'bic': None,
'aic': None,
'silhouette': None,
'calinski_harabasz': None,
'davies_bouldin': None,
'dunn': None
}
###########################
# classification accuracy #
###########################
plt.figure(figsize=[inches, inches])
plot_clf_results(results)
save_fig(join(results_save_dir, 'clf_acc'))
###########
# Pi True #
###########
Pi_true = pi_true_summary['Pi']
Pi_true_mask = Pi_true == 0
# non-paper version of Pi_true
plt.figure(figsize=[inches, inches])
sns.heatmap(Pi_true.T, cmap='Blues', square=True, cbar=False, vmin=0)
plt.xlabel('View 1 clusters')
plt.ylabel('View 2 clusters')
title = 'Pi True\nNumber of blocks={}\nBlock shapes: {}'.\
format(pi_true_summary['summary']['n_communities'],
pi_true_summary['summary']['comm_shapes'])
pd.DataFrame(out[k]).to_csv(join(save_dir, '{}.csv'.format(k)))
if 'comparison' in out.keys():
dump(out['comparison'], join(save_dir, 'variable_comparison'))
if 'cl_super_means' in out.keys():
out['cl_super_means'].to_csv(join(save_dir, 'cl_super_means.csv'))
out['stand_cl_super_means'].\
to_csv(join(save_dir, 'stand_cl_super_means.csv'))
############
# Plot PCA #
############
plt.figure(figsize=(inches, inches))
plot_gmm_pcs(gmm=model, X=X.values)
save_fig(join(save_dir, 'pca_projections.png'))
######################
# Plot cluster means #
######################
mean_save_dir = make_and_get_dir(save_dir, 'cl_means')
scatter_save_dir = make_and_get_dir(mean_save_dir, 'scatter')
if 'stand_cl_super_means' in out.keys():
stand_cl_means = out['stand_cl_super_means']
else:
stand_cl_means = out['stand_cl_means']
cluster_labels = out['info']['cluster_labels']
for cl_idx in range(stand_cl_means.shape[0]):
est_n_comp = estimator.n_components
res_writer.write("Cat data GMM estimated number of components: {}".
format(est_n_comp))
# _model_sel_dir = make_and_get_dir(model_sel_dir, 'cat_gmm')
# plt.figure(figsize=(inches, inches))
plot_model_selection(cat_model, save_dir=model_sel_dir,
name_stub='cat_gmm',
title='GMM on concatenated data',
inches=inches)
# optimization history
loss_history = estimator.opt_data_['history']['obs_nll']
plot_loss_history(loss_history,
loss_name='Observed data negative log-likelihood')
save_fig(os.path.join(opt_diag_dir, 'cat_best_model_opt_history.png'))
else:
print("cat isnt fit...")
########################
# view marginal models #
########################
n_views = len(view_models)
for v in range(n_views):
estimator = view_models[v].best_estimator_
sel_models['view_' + str(v)] = estimator