From 87b96b15dc87017d0ff9dc65f877d79f4e598ee4 Mon Sep 17 00:00:00 2001
From: ejohnson643 <eric.johnson643@gmail.com>
Date: Fri, 17 Dec 2021 15:39:34 -0600
Subject: [PATCH] Committing updates to plotting utility.

---
 EMBEDR/embedr.py           | 12 ++++++++++++
 EMBEDR/plotting_utility.py | 26 ++++++++++++++++++++++++++
 2 files changed, 38 insertions(+)

diff --git a/EMBEDR/embedr.py b/EMBEDR/embedr.py
index 78d7630..270022f 100644
--- a/EMBEDR/embedr.py
+++ b/EMBEDR/embedr.py
@@ -1999,6 +1999,18 @@ def plot(self,
                                   cite_EMBEDR=cite_EMBEDR,
                                   **plot_kwds)
 
+        elif plot_type.lower() in ['keff']:
+            from EMBEDR.plots.embedr_scatterplots import Scatterplot
+
+            plotObj = Scatterplot(plot_Y,
+                                  self._kEff,
+                                  fig=fig,
+                                  axis=axis,
+                                  cbar_ax=cbar_ax,
+                                  show_cbar=show_cbar,
+                                  cite_EMBEDR=cite_EMBEDR,
+                                  **plot_kwds)
+
         elif (metadata is not None) and (plot_type in metadata):
 
             if is_categorical:
diff --git a/EMBEDR/plotting_utility.py b/EMBEDR/plotting_utility.py
index f1301f9..ac5ea9c 100644
--- a/EMBEDR/plotting_utility.py
+++ b/EMBEDR/plotting_utility.py
@@ -658,3 +658,29 @@ def process_categorical_label(metadata, label,
 
     return raw_labels, label_counts, long_labels, lab_2_idx_map, label_cmap
 
+
+def get_DBSCAN_clusters(Y, min_samples=10, pwd_perc=1.5):
+    from sklearn.cluster import DBSCAN
+    
+    PWD = pwd(Y, metric='euclidean')
+    PWD_triu = np.triu(PWD, k=1)
+    eps = np.percentile(PWD_triu[PWD_triu.nonzero()], pwd_perc)
+
+    DBObj = DBSCAN(eps=eps, min_samples=min_samples)
+    DBObj.fit(Y)
+
+    db_labels = DBObj.labels_
+
+    ## Count the labels
+    raw_counts = Counter(db_labels)
+    ## Sort in descending order
+    db_lab_counts = sorted(raw_counts.items(), key=lambda item: -item[1])
+    ## Remove -1 label
+    db_lab_counts = {el[0]: el[1] for el in db_lab_counts if el[0] != -1}
+    ## Create mapping from old labels to size-sorted labels
+    db_lab_remap = {old_lab: new_lab for new_lab, old_lab in enumerate(db_lab_counts.keys())}
+    ## Add -1 to map
+    if -1 in raw_counts:
+        db_lab_remap[-1] = -1
+    ## Remap labels
+    return np.asarray([db_lab_remap[old_lab] for old_lab in db_labels])