ENH: add support for addfragments flag

q2_alignment/_mafft.py

@@ -26,7 +26,7 @@ def run_command(cmd, output_fp, verbose=True):
         subprocess.run(cmd, stdout=output_f, check=True)
 
 
-def _mafft(sequences_fp, alignment_fp, n_threads, parttree):
+def _mafft(sequences_fp, alignment_fp, n_threads, parttree, fragments):
     # Save original sequence IDs since long ids (~250 chars) can be truncated
     # by mafft. We'll replace the IDs in the aligned sequences file output by
     # mafft with the originals.
@@ -92,7 +92,8 @@ def _mafft(sequences_fp, alignment_fp, n_threads, parttree):
         cmd += ['--parttree']
 
     if alignment_fp is not None:
-        cmd += ['--add', sequences_fp, alignment_fp]
+        add_flag = '--addfragments' if fragments else '--add'
+        cmd += [add_flag, sequences_fp, alignment_fp]
     else:
         cmd += [sequences_fp]
 
@@ -124,13 +125,14 @@ def mafft(sequences: DNAFASTAFormat,
           n_threads: int = 1,
           parttree: bool = False) -> AlignedDNAFASTAFormat:
     sequences_fp = str(sequences)
-    return _mafft(sequences_fp, None, n_threads, parttree)
+    return _mafft(sequences_fp, None, n_threads, parttree, False)
 
 
 def mafft_add(alignment: AlignedDNAFASTAFormat,
               sequences: DNAFASTAFormat,
               n_threads: int = 1,
-              parttree: bool = False) -> AlignedDNAFASTAFormat:
+              parttree: bool = False,
+              fragments: bool = False) -> AlignedDNAFASTAFormat:
     alignment_fp = str(alignment)
     sequences_fp = str(sequences)
-    return _mafft(sequences_fp, alignment_fp, n_threads, parttree)
+    return _mafft(sequences_fp, alignment_fp, n_threads, parttree, fragments)

q2_alignment/plugin_setup.py

@@ -46,7 +46,8 @@
     inputs={'alignment': FeatureData[AlignedSequence],
             'sequences': FeatureData[Sequence]},
     parameters={'n_threads': Int % Range(1, None) | Str % Choices(['auto']),
-                'parttree': Bool},
+                'parttree': Bool,
+                'fragments': Bool},
     outputs=[('expanded_alignment', FeatureData[AlignedSequence])],
     input_descriptions={'alignment': 'The alignment to which '
                                      'sequences should be added.',
@@ -55,7 +56,9 @@
         'n_threads': 'The number of threads. (Use `auto` to automatically use '
                      'all available cores)',
         'parttree': 'This flag is required if the number of sequences being '
-                    'aligned are larger than 1000000. Disabled by default'},
+                    'aligned are larger than 1000000. Disabled by default',
+        'fragments': 'This flag indicates that alignment optimized for '
+                     'addition of fragmentary sequences should be used.'},
     output_descriptions={
         'expanded_alignment': 'Alignment containing the provided aligned and '
                               'unaligned sequences.'},

q2_alignment/tests/test_mafft.py

@@ -7,6 +7,7 @@
 # ----------------------------------------------------------------------------
 import os
 import unittest
+from unittest.mock import patch, ANY
 import subprocess
 
 import skbio
@@ -19,7 +20,6 @@
 
 
 class MafftTests(TestPluginBase):
-
     package = 'q2_alignment.tests'
 
     def _prepare_sequence_data(self):
@@ -84,7 +84,6 @@ def test_mafft_parttree_exception(self):
 
 
 class MafftAddTests(TestPluginBase):
-
     package = 'q2_alignment.tests'
 
     def _prepare_sequence_data(self):
@@ -114,6 +113,31 @@ def test_mafft_add(self):
                             constructor=skbio.DNA)
         self.assertEqual(obs, exp)
 
+    def test_mafft_add_fragments(self):
+        alignment, sequences, exp = self._prepare_sequence_data()
+
+        with redirected_stdio(stderr=os.devnull):
+            result = mafft_add(alignment, sequences, fragments=True)
+        obs = skbio.io.read(str(result), into=skbio.TabularMSA,
+                            constructor=skbio.DNA)
+        self.assertEqual(obs, exp)
+
+    def test_mafft_add_flags(self):
+        alignment, sequences, exp = self._prepare_sequence_data()
+
+        with patch('q2_alignment._mafft.run_command') as patched_run_cmd:
+            with patch('q2_alignment._mafft.skbio.TabularMSA.read',
+                       return_value=exp):
+                _ = mafft_add(alignment, sequences)
+                patched_run_cmd.assert_called_with(
+                    ["mafft", "--preservecase", "--inputorder", "--thread",
+                     "1", "--add", ANY, ANY], ANY)
+
+                _ = mafft_add(alignment, sequences, fragments=True)
+                patched_run_cmd.assert_called_with(
+                    ["mafft", "--preservecase", "--inputorder", "--thread",
+                     "1", "--addfragments", ANY, ANY], ANY)
+
     def test_duplicate_input_ids_in_unaligned(self):
         input_fp = self.get_data_path('unaligned-duplicate-ids.fasta')
         sequences = DNAFASTAFormat(input_fp, mode='r')
@@ -181,7 +205,6 @@ def test_long_ids_are_not_truncated_aligned(self):
 
 
 class RunCommandTests(TestPluginBase):
-
     package = 'q2_alignment.tests'
 
     def test_failed_run(self):