MRG: add filter_presence

pyproject.toml

@@ -3,11 +3,11 @@ name = "sourmash_plugin_betterplot"
 description = "sourmash plugin for improved plotting/viz and cluster examination."
 readme = "README.md"
 requires-python = ">=3.10"
-version = "0.4.2"
+version = "0.4.3"
 
 dependencies = ["sourmash>=4.8.8,<5", "sourmash_utils>=0.2",
                 "matplotlib", "numpy", "scipy", "scikit-learn",
-                "seaborn", "upsetplot", "matplotlib_venn"]
+                "seaborn", "upsetplot", "matplotlib_venn", "pandas"]
 
 [metadata]
 license = { text = "BSD 3-Clause License" }
@@ -24,3 +24,4 @@ cluster_to_categories_command = "sourmash_plugin_betterplot:Command_ClusterToCat
 tsne_command = "sourmash_plugin_betterplot:Command_TSNE"
 tsne2_command = "sourmash_plugin_betterplot:Command_TSNE2"
 venn = "sourmash_plugin_betterplot:Command_Venn"
+presence_filter = "sourmash_plugin_betterplot:Command_PresenceFilter"

src/sourmash_plugin_betterplot.py

@@ -23,6 +23,7 @@
 from matplotlib.lines import Line2D
 import seaborn as sns
 import upsetplot
+import pandas as pd
 
 import sourmash
 from sourmash import sourmash_args
@@ -846,6 +847,14 @@ def __init__(self, subparser):
             "--no-labels", action="store_true",
             help="disable X & Y axis labels"
         )
+        subparser.add_argument(
+            "--no-x-labels", action="store_true",
+            help="disable X axis labels"
+        )
+        subparser.add_argument(
+            "--no-y-labels", action="store_true",
+            help="disable Y axis labels"
+        )
 
     def main(self, args):
         super().main(args)
@@ -901,12 +910,15 @@ def main(self, args):
         if args.boolean:        # turn off colorbar if boolean.
             kw_args['cbar_pos'] = None
 
+        yticklabels=sample_d_to_idents(query_d_items)
+        xticklabels=sample_d_to_idents(against_d_items)
         if args.no_labels:
-            xticklabels=[]
-            yticklabels=[]
-        else:
-            yticklabels=sample_d_to_idents(query_d_items)
-            xticklabels=sample_d_to_idents(against_d_items)
+            xticklabels = []
+            yticklabels = []
+        elif args.no_x_labels:
+            xticklabels = []
+        elif args.no_y_labels:
+            yticklabels = []
 
         # turn into dissimilarity matrix
         # plot!
@@ -1471,3 +1483,59 @@ def main(self, args):
         if args.output:
             notify(f"saving to '{args.output}'")
             pylab.savefig(args.output)
+
+
+class Command_PresenceFilter(CommandLinePlugin):
+    command = 'presence_filter'
+    description = """\
+Provide a filtered view of 'gather' output, plotting detection or ANI
+against average abund for significant matches.
+"""
+
+    usage = """
+   sourmash scripts presence_filter gather.csv -o presence.png
+"""
+    epilog = epilog
+    formatter_class = argparse.RawTextHelpFormatter
+
+    def __init__(self, subparser):
+        super().__init__(subparser)
+        # add argparse arguments here.
+        subparser.add_argument('gather_csv')
+        subparser.add_argument('-o', '--output', default=None,
+                               help="save image to this file",
+                               required=True)
+        subparser.add_argument('-N', '--min-num-hashes',
+                               default=3, help='threshold (default: 3)')
+        subparser.add_argument('--detection', action="store_true",
+                               default=True)
+        subparser.add_argument('--ani', dest='detection',
+                               action="store_false")
+
+    def main(self, args):
+        df = pd.read_csv(args.gather_csv)
+        notify(f"loaded {len(df)} rows from '{args.gather_csv}'")
+
+        scaled = set(df['scaled'])
+        assert len(scaled) == 1
+        scaled = list(scaled)[0]
+
+        threshold = args.min_num_hashes * scaled
+        df = df[df['unique_intersect_bp'] >= threshold]
+        notify(f"filtered down to {len(df)} rows with unique_intersect_bp >= {threshold}")
+
+        if args.detection:
+            plt.plot(df.f_match_orig, df.average_abund, '.')
+        else:
+            plt.plot(df.match_containment_ani, df.average_abund, '.')
+        ax = plt.gca()
+        ax.set_ylabel('number of copies')
+        ax.set_yscale('log')
+
+        if args.detection:
+            ax.set_xlabel('fraction of genome detected')
+        else:
+            ax.set_xlabel('cANI of match')
+
+        notify(f"saving figure to '{args.output}'")
+        plt.savefig(args.output)