Deduce ksize from SBT for sourmash watch

sourmash_lib/__main__.py

@@ -860,14 +860,15 @@ def watch(args):
     parser.add_argument('sbt_name', help='name of SBT to search')
     parser.add_argument('inp_file', nargs='?', default='/dev/stdin')
     parser.add_argument('-o', '--output', type=argparse.FileType('wt'))
-    parser.add_argument('-k', '--ksize', type=int, default=DEFAULT_K)
     parser.add_argument('--threshold', default=0.05, type=float)
     parser.add_argument('--input-is-protein', action='store_true')
     sourmash_args.add_moltype_args(parser, default_dna=True)
     parser.add_argument('-n', '--num-hashes', type=int,
                         default=DEFAULT_N,
                         help='number of hashes to use in each sketch (default: %(default)i)')
     parser.add_argument('--name', type=str, default='stdin')
+    sourmash_args.add_ksize_arg(parser, DEFAULT_K)
+
     args = parser.parse_args(args)
 
     if args.input_is_protein and args.dna:
@@ -885,16 +886,27 @@ def watch(args):
         moltype = 'protein'
         is_protein = True
 
-    E = sourmash_lib.Estimators(ksize=args.ksize, n=args.num_hashes,
+    tree = SBT.load(args.sbt_name, leaf_loader=SigLeaf.load)
+
+    def get_ksize(tree):
+        """Walk nodes in `tree` to find out ksize"""
+        for node in tree.nodes:
+            node = node.do_load()
+            if isinstance(node, sourmash_lib.sbtmh.SigLeaf):
+                return node.data.estimator.ksize
+
+    # deduce ksize from the SBT we are loading
+    ksize = args.ksize
+    if ksize is None:
+        ksize = get_ksize(tree)
+
+    E = sourmash_lib.Estimators(ksize=ksize, n=args.num_hashes,
                                 is_protein=is_protein)
     streamsig = sig.SourmashSignature('', E, filename='stdin',
                                       name=args.name)
 
     notify('Computing signature for k={}, {} from stdin',
-           args.ksize, moltype)
-
-
-    tree = SBT.load(args.sbt_name, leaf_loader=SigLeaf.load)
+           ksize, moltype)
 
     def do_search():
         search_fn = SearchMinHashesFindBest().search

sourmash_lib/test_sourmash.py

@@ -1223,6 +1223,30 @@ def test_watch():
         assert 'FOUND: genome-s10.fa.gz, at 1.000' in err
 
 
+def test_watch_deduce_ksize():
+    with utils.TempDirectory() as location:
+        testdata0 = utils.get_test_data('genome-s10.fa.gz')
+        utils.runscript('sourmash',
+                        ['compute', testdata0, '-k', '31', '-o', '1.sig'],
+                        in_directory=location)
+
+        args = ['sbt_index', '--dna', '-k', '31', 'zzz', '1.sig']
+        status, out, err = utils.runscript('sourmash', args,
+                                           in_directory=location)
+
+        cmd = """
+
+             gunzip -c {} | {}/sourmash watch --dna zzz
+
+        """.format(testdata0, utils.scriptpath())
+        status, out, err = utils.run_shell_cmd(cmd, in_directory=location)
+
+        print(out)
+        print(err)
+        assert 'Computing signature for k=31' in err
+        assert 'genome-s10.fa.gz, at 1.000' in err
+
+
 def test_watch_coverage():
     with utils.TempDirectory() as location:
         testdata0 = utils.get_test_data('genome-s10.fa.gz')