Updated ionic_mobility collection with new scripts

ionic_mobility/analyze_db.py

@@ -100,6 +100,7 @@ def build_db_query(criterion=None, value=None, ucq=None, custom=False):
         return eval(ucq)
 
 if __name__ == "__main__":
+    # A simple set of commands to test the script.
     # Let's compare the minimum cation distance for the post-spinels.
     path_id_cm = "mvc-6921-NEB-001"
     path_id_cf = "mvc-12110-NEB-001"

ionic_mobility/crm.py

@@ -0,0 +1,45 @@
+import numpy as np
+import scipy.cluster.vq as vq
+from pylab import plot, show
+from analyze_db import initialize_db, build_db_query, build_feature_vector
+
+"""
+Clustering-ranking-modeling to learn ionic mobility predictors.  We're using
+the X-means clustering algorithm here.
+"""
+
+
+def build_point_list(fv_dict):
+    """build_point_list
+    Builds a list of feature vectors for clustering analysis.
+    :param fv_dict: Dictionary of all needed feature vectors. 
+    :return: [feature vector 1, ..., feature vector N], with N total docs.
+    """
+    point_list = []
+    for pid in fv_dict:
+        vector_columns = [fv_dict[pid][prop] for prop in fv_dict[pid]]
+        point_list.append(vector_columns)
+    return np.array(point_list)
+
+
+if __name__ == "__main__":
+    # We need a list of properties for our feature vector.
+    feature_list = ['activation_energy', 'volume_per_anion', 
+                    'average_bond_length', 'min_cation_distance']
+
+    # We also need a database and some documents from it.
+    yaml_file = "/Users/dchannah/yaml/paper2.yaml"
+    coll = initialize_db(yaml_file)
+
+    # Let's just grab all of the documents in the ionic mobility DB for now.
+    docs = coll.find()
+
+    # Now we build our feature vector for all these docs.
+    fv = build_feature_vector(feature_list, docs)
+
+    # Now let's do some clustering.
+    points = build_point_list(fv)
+    points = vq.whiten(points)  # Need to normalize each column first.
+    centroids, distortion = vq.kmeans(points, 4)
+    indices, _ = vq.vq(points, centroids)
+