Merge pull request #13 from ccdc-opensource/matplotlib_fix

Matplotlib fix

Author: IsaacSugden

scripts/packing_similarity_dendrogram/ReadMe.md

@@ -72,9 +72,11 @@ similarity.
 -dt 0.25, --dist_tol 0.
 Fractional tolerance for distances (0.0 - 1.0) used by
 packing similarity.
+-pd 5, --pad_length 5
+pad right hand side of plot to make room for structure names
 ```
 
-## Basic usage (in a command prompt):
+## Basic usage (in a command prompt)
 
 ```cmd
 python Packing_Similarity_Dendrogram.py input_file
@@ -137,6 +139,9 @@ clusters at the worst or lowest level connecting them. Figure 2 shows the same d
 which gives more definite clustering but will also hide some similarities due to the hierarchical nature of the
 clustering.
 
+If there are dangling nodes without an associated identifier, it is likely that there multiple structures given the same name
+in the supplied file (searched through multiple databases say)
+
 ![Figure 1](dendogram_figure_1.png)
 **Figure 1** : Example dendrogram based on a database of carbamazepine solid forms using the default single-linkage
 clustering.

scripts/packing_similarity_dendrogram/packing_similarity_dendogram.py

@@ -7,7 +7,7 @@
 #
 # 2016-03-22: created by Anthony Reilly, The Cambridge Crystallographic Data Centre
 # 2016-12-06: updated by Anthony Reilly, The Cambridge Crystallographic Data Centre
-#
+#19/01/2023: edited by Jonas Nyman for instances with large clusters
 
 """Packing_Similarity_Dendrogram.py - Construct a dendrogram for an input set of structures based on packing-similarity
 analysis
@@ -98,7 +98,7 @@ def merge_clusters(c1, c2, level):
     return new_cluster
 
 
-def plot_dendrogram(cluster_list, n_ps_mols, filename):
+def plot_dendrogram(cluster_list, n_ps_mols, filename, pad_length):
     """
     Function for producing a dendrogram from an input cluster hierarchy
     """
@@ -187,43 +187,25 @@ def plot_tree(cluster, x_start_positions, y_start_positions, hs):
 
     # Setup tree plotting by getting each terminal's height
     heights, count = assign_y_positions(cluster_list[0], 0, {})
-
     # Set start of the tree - middle of the plot and 1,1
     xpositions = [1, 1]
     ypositions = [1, get_midpoint(cluster_list[0], heights)]
-
     # Plot tree
     plot_tree(cluster_list[0], xpositions, ypositions, heights)
 
-    # Plot formatting
-    ax = plt.axes()
-    ax.set_frame_on(False)
-    ax.axes.get_yaxis().set_visible(False)
-    ax.spines['top'].set_visible(False)
-    ax.spines['bottom'].set_visible(False)
-    ax.spines['left'].set_visible(False)
-    ax.spines['right'].set_visible(False)
-
-    levels = range(n_ps_mols, -1, -1)
-    highlighted_levels = range(n_ps_mols, -1, -1)
-    #    highlighted_levels.append(0)
-
-    for level in highlighted_levels:
-        plt.plot([level, level], [0, count], "--", linewidth=0.5, color="Gray", zorder=1)
-
     # Pad the plot to have enough space for structure indices
-    plt.xlim(-1, n_ps_mols + 5)
+    plt.xlim(-1, n_ps_mols + pad_length)
     plt.ylim(0, count + 1)
-    ax.set_xticks(levels)
-    ax.tick_params(axis='x', bottom='off', top='off')
+    plt.xticks(np.arange(1,n_ps_mols+2,2)) #  IJS 06/09/22 addition to replace ax.set_xticks(levels)
+
     plt.xlabel('Packing Similarity / ' + str(n_ps_mols) + ' Molecules', fontsize='large')
     # Save output
     plt.savefig(filename + "_packing_similarity_tree.png", dpi=1000, bbox_inches='tight')
     print("Packing tree diagram saved to " + filename + "_packing_similarity_tree.png")
 
 
 def main(input_file, matrix_file, n_ps_mols, output_ps_results, conf_threshold, ps_angles, ps_distances, strip,
-         n_struct, allow_mol_diff, cluster_mode):
+         n_struct, allow_mol_diff, cluster_mode, pad_length):
     # Initialise Packing Similarity
     ps = PackingSimilarity()
     ps.settings.ignore_hydrogen_positions = True
@@ -244,10 +226,9 @@ def main(input_file, matrix_file, n_ps_mols, output_ps_results, conf_threshold,
     ps.settings.packing_shell_size = n_ps_mols
     ps.settings.angle_tolerance = ps_angles
     ps.settings.distance_tolerance = ps_distances
-
     refcodes = []
 
-    input_name = input_file.rsplit(".")[0]
+    input_name = os.path.basename(input_file).split(".")[0]
     print("--------------------------------------------------------")
 
     if not matrix_file:
@@ -281,13 +262,11 @@ def main(input_file, matrix_file, n_ps_mols, output_ps_results, conf_threshold,
             if not os.path.exists(overlay_folder):
                 os.makedirs(overlay_folder)
 
-        for i in range(0, structure_size):
-            refcodes.append(str(structure_reader[i].identifier))
-
         for i in range(0, structure_size):
             entry_i = structure_reader[i]
             crystal_i = entry_i.crystal
-            refcodes.append(str(structure_reader[i].identifier))
+            refcodes.append(str(i+1))
+
 
             for j in range(i, structure_size):
                 if i == j:
@@ -395,9 +374,15 @@ def main(input_file, matrix_file, n_ps_mols, output_ps_results, conf_threshold,
     y = np.arange(0, structure_size + 1, 1)
 
     plot = plt.pcolor(x, y, matrix, cmap=plt.get_cmap('rainbow', (n_ps_mols - 1)), vmin=1, vmax=n_ps_mols)
-    plt.xticks(np.arange(0, structure_size + 1, 5) - 0.5, np.arange(0, structure_size + 1, 5))
-    plt.yticks(np.arange(0, structure_size + 1, 5) - 0.5, np.arange(0, structure_size + 1, 5))
-    cb = plt.colorbar(plot, ticks=range(1, 16))
+    if structure_size < 10:
+        plt.xticks(np.arange(0, structure_size + 1) - 0.5, np.arange(0, structure_size + 1))
+        plt.yticks(np.arange(0, structure_size + 1) - 0.5, np.arange(0, structure_size + 1))
+    else:
+        plt.xticks(np.arange(0, structure_size + 1, 5) - 0.5, np.arange(0, structure_size + 1, 5))
+        plt.yticks(np.arange(0, structure_size + 1, 5) - 0.5, np.arange(0, structure_size + 1, 5))
+    cb = plt.colorbar(plot, ticks=range(1, n_ps_mols+1))
+
+
 
     plt.xlim(0, structure_size)
     plt.ylim(0, structure_size)
@@ -407,11 +392,11 @@ def main(input_file, matrix_file, n_ps_mols, output_ps_results, conf_threshold,
     cb.set_label('Packing Similarity  /' + str(n_ps_mols) + ' Molecules', fontsize='x-large')
     plt.savefig(input_name + "_heat_map.png", dpi=300)
     print("Packing similarity heat map saved to " + input_name + "_heat_map.png")
+    ax.clear()
     plt.close()
 
     # Plot a dendrogram
-    plot_dendrogram(cluster_list, n_ps_mols, input_name)
-
+    plot_dendrogram(cluster_list, n_ps_mols, input_name, pad_length)
     print("--------------------------------------------------------")
 
     sys.exit()
@@ -445,6 +430,8 @@ def main(input_file, matrix_file, n_ps_mols, output_ps_results, conf_threshold,
                         help="Tolerance for angles (in degrees) used by packing similarity.")
     parser.add_argument('-dt', '--dist_tol', type=float, default=0.25, metavar="0.25",
                         help="Fractional tolerance for distances (0.0 - 1.0) used by packing similarity.")
+    parser.add_argument('-pd', '--pad_length', type=float, default=5.0, metavar="0.25",
+                        help="padding on right of the plot for listing identifiers") #  IJS 06/09/22 addition, as I hve an example with many refcodes that overspills the plot
     args = parser.parse_args()
     if not os.path.isfile(args.input_file):
         parser.error('%s not found.' % args.input_file)
@@ -454,4 +441,4 @@ def main(input_file, matrix_file, n_ps_mols, output_ps_results, conf_threshold,
 
     main(args.input_file, args.matrix, args.n_molecules, args.o, args.conf_tol, args.angle_tol,
          args.dist_tol, args.strip, args.n_structures, args.allow_molecular_differences,
-         args.clustering_type)
+         args.clustering_type, args.pad_length)