Tree fixes (Critical) (#99)

* deleting old test data

* use solid branch lines only for MS/MS matches

* set taxonomy = unclassified when empty set returned by classyfire

* taxonomy fixes

* unclassified taxonomy when classyfire returns an empty dict

* add case where classyfire returns empty dict

* remove print statement

* check for NaNs in classification, there shouldnt be any

* deconvolute if else condensed form for readability

* remove redundant .keys() call for dict

Author: anupriyatripathi

q2_qemistree/_classyfire.py

@@ -78,14 +78,18 @@ def get_classyfire_taxonomy(feature_data: pd.DataFrame) -> pd.DataFrame:
             response = requests.get(to_classyfire+str(inchikey)+'.json')
             if response.status_code == 200:
                 response = response.json()
-                classyfire_levels_sub = [level for level in classyfire_levels
-                                         if level in response.keys()]
-                if len(classyfire_levels_sub) == 0:
+                sublevels = [level for level in classyfire_levels
+                             if level in response]
+                if len(sublevels) == 0:
+                    classyfire[idx] = 'unclassified'
                     continue
-                taxonomy = [response[level]['name']
-                            if bool(response) and response[level] is not None
-                            else 'unclassified'
-                            for level in classyfire_levels_sub]
+                taxonomy = []
+                for level in classyfire_levels:
+                    if (response and level in sublevels and
+                            response[level] is not None):
+                        taxonomy.append(response[level]['name'])
+                    else:
+                        taxonomy.append('unclassified')
                 classyfire[idx] = taxonomy
             elif response.status_code == 404:
                 classyfire[idx] = 'unclassified'

q2_qemistree/_itol_metadata.py

@@ -81,7 +81,7 @@ def get_itol_visualization(classified_feature_data: str,
                 if fdata.loc[idx, 'annotation_type'] == 'MS2':
                     fh.write(idx + '\t' + 'clade\t' +
                              color + '\tnormal\t6\n')
-                if fdata.loc[idx, 'annotation_type'] == 'CSIFingerID':
+                else:
                     fh.write(idx + '\t' + 'clade\t' +
                              color + '\tdashed\t4\n')
     if label_file_path:

q2_qemistree/tests/data/feature_data_smiles.txt

@@ -1,4 +1,5 @@
 label	#featureID	csi_smiles	ms2_smiles	table_number
-3638d9b9658b202f9c773e5e43206d32	3			1
-c8f6e296c0bacf58cfec06b333fd2015	7	CCCCCCCCCCCCCCCC(=O)OCC(COP(=O)([O-])OCC[N+](C)(C)C)O		1
-a500d26568f130313bdb70e3ae7d2ade	2			1
+featurehash1	3			1
+featurehash2	7	CCCCCCCCCCCCCCCC(=O)OCC(COP(=O)([O-])OCC[N+](C)(C)C)O		1
+featurehash3	2			1
+featurehash4	5	CC(=NC(=O)CC(=NC(=O)C)OOC(=O)C)O		1

q2_qemistree/tests/test_classyfire.py

@@ -54,7 +54,7 @@ def test_malformed_smiles(self):
     def test_classyfire_output(self):
         classified = get_classyfire_taxonomy(self.smiles)
         classified_mols = classified[classified['kingdom'] != 'unclassified']
-        self.assertTrue(len(classified_mols) == 1)
+        self.assertTrue(pd.isna(classified_mols).shape, 0)
         self.assertTrue(classified_mols.loc[1,
                         'kingdom'] == 'Organic compounds')
         self.assertTrue((self.levels.issubset(set(classified.columns))))