Fix lca gather to work on non-DNA databases.

sourmash/lca/command_gather.py

@@ -184,14 +184,19 @@ def gather_main(args):
     """
     set_quiet(args.quiet, args.debug)
 
+    notify("** WARNING: lca gather is deprecated as of sourmash 3.4, and will")
+    notify("**    be removed in sourmash 4.0; use 'gather' instead.")
+    notify('')
+
     if not check_files_exist(args.query, *args.db):
         sys.exit(-1)
 
     # load all the databases
     dblist, ksize, scaled = lca_utils.load_databases(args.db, None)
 
     # for each query, gather all the matches across databases
-    query_sig = sourmash_args.load_query_signature(args.query, ksize, 'DNA')
+    moltype = dblist[0].moltype
+    query_sig = sourmash_args.load_query_signature(args.query, ksize, moltype)
     debug('classifying', query_sig.name())
 
     # make sure we're looking at the same scaled value as database

sourmash/lca/lca_db.py

@@ -519,6 +519,7 @@ def load_databases(filenames, scaled=None, verbose=True):
     "Load multiple LCA databases; return (dblist, ksize, scaled)"
     ksize_vals = set()
     scaled_vals = set()
+    moltype_vals = set()
     dblist = []
 
     # load all the databases
@@ -537,14 +538,19 @@ def load_databases(filenames, scaled=None, verbose=True):
             lca_db.downsample_scaled(scaled)
         scaled_vals.add(lca_db.scaled)
 
+        moltype_vals.add(lca_db.moltype)
+        if len(moltype_vals) > 1:
+            raise Exception('multiple moltypes, quitting')
+
         dblist.append(lca_db)
 
     ksize = ksize_vals.pop()
     scaled = scaled_vals.pop()
+    moltype = moltype_vals.pop()
 
     if verbose:
         notify(u'\r\033[K', end=u'')
-        notify('loaded {} LCA databases. ksize={}, scaled={}', len(dblist),
-               ksize, scaled)
+        notify('loaded {} LCA databases. ksize={}, scaled={} moltype={}',
+               len(dblist), ksize, scaled, moltype)
 
     return dblist, ksize, scaled

sourmash/sbt.py

@@ -57,8 +57,6 @@ def search_transcript(node, seq, threshold):
 import sys
 from tempfile import NamedTemporaryFile
 
-from deprecation import deprecated
-
 from .exceptions import IndexNotSupported
 from .sbt_storage import FSStorage, TarStorage, IPFSStorage, RedisStorage, ZipStorage
 from .logging import error, notify, debug

tests/test_lca.py

@@ -1671,6 +1671,44 @@ def test_incompat_lca_db_scaled(c):
     assert 'new scaled 10000 is lower than current sample scaled 10000' in str(e.value)
 
 
+@utils.in_thisdir
+def test_lca_gather_protein(c):
+    # test lca gather on protein foo
+    testquery = utils.get_test_data('prot/protein/GCA_001593925.1_ASM159392v1_protein.faa.gz.sig')
+    db1 = utils.get_test_data('prot/protein.lca.json.gz')
+
+    c.run_sourmash('lca', 'gather', testquery, db1)
+
+    assert c.last_result.status == 0
+    assert 'loaded 1 LCA databases. ksize=57, scaled=100 moltype=protein' in c.last_result.err
+    assert '340.9 kbp   100.0%  100.0%      s__B26-1 sp001593925 sp.' in c.last_result.out
+
+
+@utils.in_thisdir
+def test_lca_gather_deprecated_message(c):
+    # lca gather is deprecated for 4.0; check message
+    testquery = utils.get_test_data('prot/protein/GCA_001593925.1_ASM159392v1_protein.faa.gz.sig')
+    db1 = utils.get_test_data('prot/protein.lca.json.gz')
+
+    c.run_sourmash('lca', 'gather', testquery, db1)
+
+    assert c.last_result.status == 0
+    assert 'WARNING: lca gather is deprecated as of sourmash 3.4' in c.last_result.err
+
+
+@utils.in_thisdir
+def test_incompat_lca_db_moltype(c):
+    # test load of incompatible LCA DBs
+    testquery = utils.get_test_data('prot/protein/GCA_001593925.1_ASM159392v1_protein.faa.gz.sig')
+    db1 = utils.get_test_data('prot/protein.lca.json.gz')
+    db2 = utils.get_test_data('prot/dayhoff.lca.json.gz')
+
+    with pytest.raises(ValueError) as e:
+        c.run_sourmash('lca', 'gather', testquery, db1, db2)
+
+    assert 'Exception: multiple moltypes, quitting' in str(e.value)
+
+
 @utils.in_tempdir
 def test_incompat_lca_db_ksize(c):
     # create a database with ksize of 25