Commit b7530e26935afb3b75313d776f3795b5b4b2ecb7
1 parent
29644ae6c3
Exists in
master
allow you to plot count matrix. For each cluster, the number of element belongin…
…g to each class is plot into a matrix.
Showing 1 changed file with 106 additions and 0 deletions Side-by-side Diff
bin/plot-count-matrix.py
| 1 | +''' | |
| 2 | +This script aims to plot matrix count. | |
| 3 | +''' | |
| 4 | +import argparse | |
| 5 | +import numpy as np | |
| 6 | +from data import read_file, index_by_id | |
| 7 | +from sklearn import preprocessing | |
| 8 | +import matplotlib.pyplot as plt | |
| 9 | + | |
| 10 | + | |
| 11 | +# TODO: Avoir la liste des personnages | |
| 12 | +# TODO: liste des clusters | |
| 13 | +parser = argparse.ArgumentParser(description="Plot count matrix") | |
| 14 | +parser.add_argument("clustering", type=str, | |
| 15 | + help="clustering file") | |
| 16 | +parser.add_argument("classlst", type=str, | |
| 17 | + help="List used for its classes.") | |
| 18 | +parser.add_argument("lst", type=str, | |
| 19 | + help="list") | |
| 20 | +parser.add_argument("--outfile", type=str, default="out.pdf", | |
| 21 | + help="output file path") | |
| 22 | + | |
| 23 | +args = parser.parse_args() | |
| 24 | +CLUSTERING = args.clustering | |
| 25 | +CLASS_LST = args.classlst | |
| 26 | +LST = args.lst | |
| 27 | +OUTFILE = args.outfile | |
| 28 | + | |
| 29 | +# -- READ FILES | |
| 30 | +clustering = read_file(CLUSTERING) | |
| 31 | +clustering_ind = index_by_id(clustering) | |
| 32 | + | |
| 33 | +class_lst = read_file(CLASS_LST) | |
| 34 | +class_lst_ind = index_by_id(class_lst) | |
| 35 | + | |
| 36 | +lst = read_file(LST) | |
| 37 | + | |
| 38 | +# -- GET CLASSES AND CLUSTERS | |
| 39 | +classes = np.asarray([class_lst_ind[x[0][0]][x[0][3]][0][1] for x in lst]) | |
| 40 | +clusters = np.asarray([clustering_ind[x[0][0]][x[0][3]][0][1] for x in lst]) | |
| 41 | + | |
| 42 | +def generate_count_matrix(classes, clusters): | |
| 43 | + ''' | |
| 44 | + Generate matrices for the given set | |
| 45 | + Lines are clusters and columns are classes. | |
| 46 | + A cell is contains the number of character occurence | |
| 47 | + on a specific cluster. | |
| 48 | + ''' | |
| 49 | + | |
| 50 | + # Index Classes | |
| 51 | + classe_unique = np.unique(classes) | |
| 52 | + #all_classes = np.unique(np.concatenate((classe_unique))) | |
| 53 | + all_classes = classe_unique | |
| 54 | + | |
| 55 | + # Label Encoder for classes | |
| 56 | + le = preprocessing.LabelEncoder() | |
| 57 | + le.fit(all_classes) | |
| 58 | + | |
| 59 | + # Index | |
| 60 | + cluster_unique = np.unique(clusters) | |
| 61 | + | |
| 62 | + #all_clusters = np.unique(np.concatenate((cluster_unique))) | |
| 63 | + all_clusters = cluster_unique | |
| 64 | + # Create matrix lin(clust) col(class) | |
| 65 | + counts_matrix = np.zeros((np.max(np.asarray(all_clusters, dtype=np.int)) + 1, len(all_classes))) | |
| 66 | + | |
| 67 | + for cluster in all_clusters: | |
| 68 | + | |
| 69 | + # Il faut d'abord extraire les classes présentes dans ce cluster | |
| 70 | + cc = np.extract(np.asarray(clusters) == cluster, np.asarray(classes)) | |
| 71 | + | |
| 72 | + cc_unique, cc_counts = np.unique(cc, return_counts=True) | |
| 73 | + cc_ind = dict(zip(cc_unique, cc_counts)) | |
| 74 | + | |
| 75 | + for class_ in all_classes: | |
| 76 | + class_id = le.transform([class_])[0] | |
| 77 | + if class_ in cc_ind: | |
| 78 | + counts_matrix[int(cluster)][int(class_id)] = cc_ind[class_] | |
| 79 | + return (counts_matrix, all_classes, all_clusters) | |
| 80 | + | |
| 81 | +count_matrix, all_classes, all_clusters = generate_count_matrix(classes, clusters) | |
| 82 | + | |
| 83 | +fig, ax = plt.subplots() | |
| 84 | +fig.set_size_inches(10, len(all_clusters) + 1) | |
| 85 | +im = ax.imshow(count_matrix) | |
| 86 | + | |
| 87 | +ax.set_xticks(np.arange(len(all_classes))) | |
| 88 | +ax.set_yticks(np.arange(len(all_clusters))) | |
| 89 | + | |
| 90 | +ax.set_xticklabels(all_classes) | |
| 91 | +ax.set_yticklabels(all_clusters) | |
| 92 | + | |
| 93 | +fig.colorbar(im, ax=ax) | |
| 94 | + | |
| 95 | +plt.setp(ax.get_xticklabels(), rotation=45, ha="right", rotation_mode="anchor") | |
| 96 | + | |
| 97 | +# Loop over data dimensions and create text annotations. | |
| 98 | +for i in range(count_matrix.shape[0]): | |
| 99 | + for j in range(count_matrix.shape[1]): | |
| 100 | + text = ax.text(j, i, int(count_matrix[i, j]), | |
| 101 | + ha="center", va="center", color="w") | |
| 102 | + | |
| 103 | + | |
| 104 | +ax.set_title("Count Matrix") | |
| 105 | +fig.tight_layout() | |
| 106 | +plt.savefig(OUTFILE, bbox_inches='tight') |