figs.py 文件源码

def make_multi_hierarch_cluster_figs(model, field_input):
    def make_multi_cat_clust_fig(cats, freq_thr=500):  # TODO make into config
        """
        Returns fig showing hierarchical clustering of probes from multiple categories
        """
        start = time.time()
        sns.set_style('white')
        # make cat_acts_mat
        acts_mats = []
        cats_probe_list = []
        for cat in cats:
            bool_index = [True if sum(model.term_doc_freq_dict[probe]) > freq_thr else False
                          for probe in model.probe_store.cat_probe_list_dict[cat]]

            cat_probe_acts_df = model.get_single_cat_acts_df(cat)
            filtered_cat_probes_acts_mat = cat_probe_acts_df[bool_index].values
            acts_mats.append(filtered_cat_probes_acts_mat)
            cats_probe_list += [model.probe_store.probe_set[probe_id]
                                for probe_id in cat_probe_acts_df[bool_index].index.tolist()]
        cat_acts_mat = np.vstack((mat for mat in acts_mats))
        # fig
        rcParams['lines.linewidth'] = 2.0
        fig, ax = plt.subplots(figsize=(FigsConfigs.MAX_FIG_WIDTH, 5 * len(cats)), dpi=FigsConfigs.DPI)
        # dendrogram
        dist_matrix = pdist(cat_acts_mat, 'euclidean')
        linkages = linkage(dist_matrix, method='complete')
        dendrogram(linkages,
                   ax=ax,
                   labels=cats_probe_list,
                   orientation='right',
                   leaf_font_size=10)
        ax.tick_params(axis='both', which='both', top='off', right='off', left='off')
        ax.spines['right'].set_visible(False)
        ax.spines['left'].set_visible(False)
        ax.spines['top'].set_visible(False)
        print('{} completed in {:.1f} secs'.format(sys._getframe().f_code.co_name, time.time() - start))
        return fig

    figs = [make_multi_cat_clust_fig(field_input)]
    return figs