Merge 7c70944 into 98d2e37

q2_types/feature_data/__init__.py

@@ -13,21 +13,23 @@
     HeaderlessTSVTaxonomyDirectoryFormat, TSVTaxonomyFormat,
     TSVTaxonomyDirectoryFormat, DNAFASTAFormat, DNASequencesDirectoryFormat,
     PairedDNASequencesDirectoryFormat, AlignedDNAFASTAFormat,
-    AlignedDNASequencesDirectoryFormat)
+    AlignedDNASequencesDirectoryFormat, DifferentialFormat,
+    DifferentialDirectoryFormat)
 from ._type import (
-    FeatureData, Taxonomy, Sequence, PairedEndSequence, AlignedSequence)
+    FeatureData, Taxonomy, Sequence, PairedEndSequence, AlignedSequence,
+    Differential)
 
 # TODO remove these imports when tests are rewritten. Remove from __all__ too
 from ._transformer import DNAIterator, PairedDNAIterator, AlignedDNAIterator
 
 __all__ = [
     'TaxonomyFormat', 'TaxonomyDirectoryFormat', 'HeaderlessTSVTaxonomyFormat',
     'HeaderlessTSVTaxonomyDirectoryFormat', 'TSVTaxonomyFormat',
-    'TSVTaxonomyDirectoryFormat', 'DNAFASTAFormat',
+    'TSVTaxonomyDirectoryFormat', 'DNAFASTAFormat', 'DifferentialFormat',
     'DNASequencesDirectoryFormat', 'PairedDNASequencesDirectoryFormat',
     'AlignedDNAFASTAFormat', 'AlignedDNASequencesDirectoryFormat',
     'FeatureData', 'Taxonomy', 'Sequence', 'PairedEndSequence',
     'AlignedSequence', 'DNAIterator', 'PairedDNAIterator',
-    'AlignedDNAIterator']
+    'AlignedDNAIterator', 'Differential', 'DifferentialDirectoryFormat']
 
 importlib.import_module('q2_types.feature_data._transformer')

q2_types/feature_data/_format.py

@@ -8,7 +8,8 @@
 
 import skbio.io
 import qiime2.plugin.model as model
-
+import qiime2
+from qiime2.plugin import ValidationError
 from ..plugin_setup import plugin
 
 
@@ -185,10 +186,30 @@ def sniff(self):
     AlignedDNAFASTAFormat)
 
 
+class DifferentialFormat(model.TextFileFormat):
+    def validate(self, *args):
+        try:
+            md = qiime2.Metadata.load(str(self))
+        except qiime2.metadata.MetadataFileError as md_exc:
+            raise ValidationError(md_exc) from md_exc
+
+        if md.column_count == 0:
+            raise ValidationError('Format must contain at least 1 column')
+
+        filtered_md = md.filter_columns(column_type='numeric')
+        if filtered_md.column_count != md.column_count:
+            raise ValidationError('Must only contain numeric values.')
+
+
+DifferentialDirectoryFormat = model.SingleFileDirectoryFormat(
+    'DifferentialDirectoryFormat', 'differentials.tsv', DifferentialFormat)
+
+
 plugin.register_formats(
     TSVTaxonomyFormat, TSVTaxonomyDirectoryFormat,
     HeaderlessTSVTaxonomyFormat, HeaderlessTSVTaxonomyDirectoryFormat,
     TaxonomyFormat, TaxonomyDirectoryFormat, DNAFASTAFormat,
     DNASequencesDirectoryFormat, PairedDNASequencesDirectoryFormat,
-    AlignedDNAFASTAFormat, AlignedDNASequencesDirectoryFormat
+    AlignedDNAFASTAFormat, AlignedDNASequencesDirectoryFormat,
+    DifferentialFormat, DifferentialDirectoryFormat
 )

q2_types/feature_data/_transformer.py

@@ -18,7 +18,7 @@
 from ..feature_table import BIOMV210Format
 from . import (TaxonomyFormat, HeaderlessTSVTaxonomyFormat, TSVTaxonomyFormat,
                DNAFASTAFormat, PairedDNASequencesDirectoryFormat,
-               AlignedDNAFASTAFormat)
+               AlignedDNAFASTAFormat, DifferentialFormat)
 
 
 # Taxonomy format transformers
@@ -359,3 +359,21 @@ def _19(data: AlignedDNAIterator) -> AlignedDNAFASTAFormat:
 @plugin.register_transformer
 def _33(ff: AlignedDNAFASTAFormat) -> qiime2.Metadata:
     return _dnafastaformats_to_metadata(ff)
+
+
+# differential types
+@plugin.register_transformer
+def _222(ff: DifferentialFormat) -> pd.DataFrame:
+    return qiime2.Metadata.load(str(ff)).to_dataframe()
+
+
+@plugin.register_transformer
+def _223(ff: DifferentialFormat) -> qiime2.Metadata:
+    return qiime2.Metadata.load(str(ff))
+
+
+@plugin.register_transformer
+def _224(data: pd.DataFrame) -> DifferentialFormat:
+    ff = DifferentialFormat()
+    qiime2.Metadata(data).save(str(ff))
+    return ff

q2_types/feature_data/_type.py

@@ -11,7 +11,8 @@
 from ..plugin_setup import plugin
 from . import (TSVTaxonomyDirectoryFormat, DNASequencesDirectoryFormat,
                PairedDNASequencesDirectoryFormat,
-               AlignedDNASequencesDirectoryFormat)
+               AlignedDNASequencesDirectoryFormat,
+               DifferentialDirectoryFormat)
 
 
 FeatureData = SemanticType('FeatureData', field_names='type')
@@ -26,8 +27,13 @@
 AlignedSequence = SemanticType('AlignedSequence',
                                variant_of=FeatureData.field['type'])
 
+Differential = SemanticType('Differential',
+                            variant_of=FeatureData.field['type'])
+
 plugin.register_semantic_types(FeatureData, Taxonomy, Sequence,
-                               PairedEndSequence, AlignedSequence)
+                               PairedEndSequence, AlignedSequence,
+                               Differential)
+
 
 plugin.register_semantic_type_to_format(
     FeatureData[Taxonomy],
@@ -41,3 +47,5 @@
 plugin.register_semantic_type_to_format(
     FeatureData[AlignedSequence],
     artifact_format=AlignedDNASequencesDirectoryFormat)
+plugin.register_semantic_type_to_format(
+    FeatureData[Differential], DifferentialDirectoryFormat)

q2_types/feature_data/tests/data/bad_differential.tsv

@@ -0,0 +1,32 @@
+featureid	effect	bad_effect
+F0	-0.910182258821735	a
+F1	1.01418002973925	0
+F2	1.02456128258909	0
+F3	-0.74363992043225	0
+F4	1.29823896534823	0
+F5	-1.12965055281585	0
+F6	-0.522401797448688	0
+F7	0.327560711072239	0
+F8	-1.3738693465664802	0
+F9	-0.7847891526325621	0
+F10	-0.280063201878434	0
+F11	-0.251269847578052	0
+F12	1.24602780723028	0
+F13	0.665734866338239	0
+F14	-0.889042985114811	0
+F15	-0.811956802515126	0
+F16	0.11591764582945001	0
+F17	-0.156195990858492	0
+F18	-0.965770633683909	0
+F19	0.8035240337800391	0
+F20	0.680306950765235	0
+F21	-0.688388077896823	0
+F22	0.7949653982850671	0
+F23	-1.11912925367142	0
+F24	-1.1059356352774599	0
+F25	0.678934047810573	0
+F26	-0.937189288219405	0
+F27	-0.9997301931164799	0
+F28	-0.31799033232181	0
+F29	-0.66141741897569	0
+F30	0.550511528315366	0

q2_types/feature_data/tests/data/differentials.tsv

@@ -0,0 +1,32 @@
+featureid	effect
+F0	-0.910182258821735
+F1	1.01418002973925
+F2	1.02456128258909
+F3	-0.74363992043225
+F4	1.29823896534823
+F5	-1.12965055281585
+F6	-0.522401797448688
+F7	0.327560711072239
+F8	-1.3738693465664802
+F9	-0.7847891526325621
+F10	-0.280063201878434
+F11	-0.251269847578052
+F12	1.24602780723028
+F13	0.665734866338239
+F14	-0.889042985114811
+F15	-0.811956802515126
+F16	0.11591764582945001
+F17	-0.156195990858492
+F18	-0.965770633683909
+F19	0.8035240337800391
+F20	0.680306950765235
+F21	-0.688388077896823
+F22	0.7949653982850671
+F23	-1.11912925367142
+F24	-1.1059356352774599
+F25	0.678934047810573
+F26	-0.937189288219405
+F27	-0.9997301931164799
+F28	-0.31799033232181
+F29	-0.66141741897569
+F30	0.550511528315366

q2_types/feature_data/tests/data/empty_differential.tsv

@@ -0,0 +1,32 @@
+featureid
+F0
+F1
+F2
+F3
+F4
+F5
+F6
+F7
+F8
+F9
+F10
+F11
+F12
+F13
+F14
+F15
+F16
+F17
+F18
+F19
+F20
+F21
+F22
+F23
+F24
+F25
+F26
+F27
+F28
+F29
+F30

q2_types/feature_data/tests/data/inf_differential.tsv

@@ -0,0 +1,32 @@
+featureid	effect
+F0	inf
+F1	1.01418002973925
+F2	1.02456128258909
+F3	-0.74363992043225
+F4	1.29823896534823
+F5	-1.12965055281585
+F6	-0.522401797448688
+F7	0.327560711072239
+F8	-1.3738693465664802
+F9	-0.7847891526325621
+F10	-0.280063201878434
+F11	-0.251269847578052
+F12	1.24602780723028
+F13	0.665734866338239
+F14	-0.889042985114811
+F15	-0.811956802515126
+F16	0.11591764582945
+F17	-0.156195990858492
+F18	-0.965770633683909
+F19	0.8035240337800391
+F20	0.680306950765235
+F21	-0.688388077896823
+F22	0.7949653982850671
+F23	-1.11912925367142
+F24	-1.10593563527746
+F25	0.678934047810573
+F26	-0.937189288219405
+F27	-0.99973019311648
+F28	-0.31799033232181
+F29	-0.66141741897569
+F30	0.550511528315366

q2_types/feature_data/tests/data/not_differential.tsv

@@ -0,0 +1 @@
+asdfasdfasdfasdf asdfasdfa asdfasdf asdf asdfas dfa

q2_types/feature_data/tests/test_format.py

@@ -16,7 +16,7 @@
     HeaderlessTSVTaxonomyDirectoryFormat, TSVTaxonomyFormat,
     TSVTaxonomyDirectoryFormat, DNAFASTAFormat, DNASequencesDirectoryFormat,
     PairedDNASequencesDirectoryFormat, AlignedDNAFASTAFormat,
-    AlignedDNASequencesDirectoryFormat
+    AlignedDNASequencesDirectoryFormat, DifferentialDirectoryFormat
 )
 from qiime2.plugin.testing import TestPluginBase
 from qiime2.plugin import ValidationError
@@ -203,5 +203,58 @@ def test_aligned_dna_sequences_directory_format(self):
         format.validate()
 
 
+class TestDifferentialFormat(TestPluginBase):
+    package = 'q2_types.feature_data.tests'
+
+    def test_differential_format(self):
+        filepath = self.get_data_path('differentials.tsv')
+        temp_dir = self.temp_dir.name
+        shutil.copy(filepath,
+                    os.path.join(temp_dir, 'differentials.tsv'))
+        format = DifferentialDirectoryFormat(temp_dir, mode='r')
+        format.validate()
+        self.assertTrue(True)
+
+    def test_differential_format_empty(self):
+        filepath = self.get_data_path('empty_differential.tsv')
+        temp_dir = self.temp_dir.name
+        shutil.copy(filepath,
+                    os.path.join(temp_dir, 'differentials.tsv'))
+
+        with self.assertRaisesRegex(ValidationError, 'least 1 column'):
+            format = DifferentialDirectoryFormat(temp_dir, mode='r')
+            format.validate()
+
+    def test_differential_format_not(self):
+        filepath = self.get_data_path('not_differential.tsv')
+        temp_dir = self.temp_dir.name
+        shutil.copy(filepath,
+                    os.path.join(temp_dir, 'differentials.tsv'))
+
+        with self.assertRaises(ValidationError):
+            format = DifferentialDirectoryFormat(temp_dir, mode='r')
+            format.validate()
+
+    def test_differential_format_inf(self):
+        filepath = self.get_data_path('inf_differential.tsv')
+        temp_dir = self.temp_dir.name
+        shutil.copy(filepath,
+                    os.path.join(temp_dir, 'differentials.tsv'))
+
+        with self.assertRaisesRegex(ValidationError, 'numeric'):
+            format = DifferentialDirectoryFormat(temp_dir, mode='r')
+            format.validate()
+
+    def test_differential_format_bad_type(self):
+        filepath = self.get_data_path('bad_differential.tsv')
+        temp_dir = self.temp_dir.name
+        shutil.copy(filepath,
+                    os.path.join(temp_dir, 'differentials.tsv'))
+
+        with self.assertRaisesRegex(ValidationError, 'numeric'):
+            format = DifferentialDirectoryFormat(temp_dir, mode='r')
+            format.validate()
+
+
 if __name__ == '__main__':
     unittest.main()

q2_types/feature_data/tests/test_transformer.py

@@ -21,7 +21,7 @@
     TaxonomyFormat, HeaderlessTSVTaxonomyFormat, TSVTaxonomyFormat,
     DNAFASTAFormat, DNAIterator, PairedDNAIterator,
     PairedDNASequencesDirectoryFormat, AlignedDNAFASTAFormat,
-    AlignedDNAIterator
+    DifferentialFormat, AlignedDNAIterator
 )
 from q2_types.feature_data._transformer import (
     _taxonomy_formats_to_dataframe, _dataframe_to_tsv_taxonomy_format)
@@ -656,5 +656,42 @@ def test_aligned_dnafasta_format_to_metadata(self):
         self.assertEqual(exp, obs)
 
 
+class TestDifferentialTransformer(TestPluginBase):
+    package = 'q2_types.feature_data.tests'
+
+    def test_differential_to_df(self):
+
+        _, obs = self.transform_format(DifferentialFormat, pd.DataFrame,
+                                       filename='differentials.tsv')
+
+        # sniff to see if the first 4 feature ids are the same
+        exp = ['F0', 'F1', 'F2', 'F3']
+        obs = list(obs.index[:4])
+        self.assertListEqual(exp, obs)
+
+    def test_differential_to_md(self):
+
+        _, obs = self.transform_format(DifferentialFormat, qiime2.Metadata,
+                                       filename='differentials.tsv')
+        obs = obs.to_dataframe()
+        # sniff to see if the first 4 feature ids are the same
+        exp = ['F0', 'F1', 'F2', 'F3']
+        obs = list(obs.index[:4])
+        self.assertListEqual(exp, obs)
+
+    def test_df_to_differential(self):
+        transformer = self.get_transformer(pd.DataFrame, DifferentialFormat)
+
+        index = pd.Index(['SEQUENCE1', 'SEQUENCE2', 'SEQUENCE3'])
+        index.name = 'featureid'
+        input = pd.DataFrame(
+            [-1.3, 0.1, 1.2], index=index, columns=['differential'],
+            dtype=float)
+
+        obs = transformer(input)
+
+        self.assertIsInstance(obs, DifferentialFormat)
+
+
 if __name__ == '__main__':
     unittest.main()