3 Md @s$ddZddZddZddZdS) cCs|j||}|S)aPropagate labels by one step Parameters ---------- P : scipy sparse matrix, shape = [n_samples, n_samples] Propagation matrix F : numpy array, shape = [n_samples, n_classes] Label matrix B : numpy array, shape = [n_samples, n_classes] Base matrix Returns ---------- F_new : array, shape = [n_samples, n_classes] Label matrix )dot)PFBZF_newr_/var/www/html/virt/lib/python3.6/site-packages/networkx/algorithms/node_classification/utils.py _propagatesrc Csddl}g}i}d}x^t|jddD]J\}}||dkr&|d|}||kr^|||<|d7}|j|||gq&W|j|}|jddt|jdd d D} || fS) aGet and return information of labels from the input graph Parameters ---------- G : Network X graph label_name : string Name of the target label Returns ---------- labels : numpy array, shape = [n_labeled_samples, 2] Array of pairs of labeled node ID and label ID label_dict : numpy array, shape = [n_classes] Array of labels i-th element contains the label corresponding label ID `i` NT)datacSsg|] \}}|qSrr).0label_rrr 5sz#_get_label_info..cSs|dS)Nr r)xrrr5sz!_get_label_info..)key)numpy enumerateZnodesappendarraysorteditems) GZ label_namenplabelsZ label_to_idZlidinr label_dictrrr_get_label_infos    rcCsddl}|j||f}|S)aRCreate and return zero matrix Parameters ---------- n_samples : integer The number of nodes (samples) on the input graph n_classes : integer The number of classes (distinct labels) on the input graph Returns ---------- F : numpy array, shape = [n_samples, n_classes] Label matrix rN)rZzeros)Z n_samplesZ n_classesrrrrr_init_label_matrix:srcCs&ddl}|j|dd}||j}|S)aPredict labels by learnt label matrix Parameters ---------- F : numpy array, shape = [n_samples, n_classes] Learnt (resulting) label matrix label_dict : numpy array, shape = [n_classes] Array of labels i-th element contains the label corresponding label ID `i` Returns ---------- predicted : numpy array, shape = [n_samples] Array of predicted labels rNr )Zaxis)rZargmaxtolist)rrrZpredicted_label_idsZ predictedrrr_predictOs r!N)rrrr!rrrrs$