Merge pull request #18 from galaxy-genome-annotation/gff_go

Fix loading of GO terms from GFF

README.md

@@ -87,6 +87,9 @@ $ chakin feature load_fasta \
 
 ## History
 
+- 2.3.6
+    - Fix loading of GO terms from GFF
+
 - 2.3.5
     - Fix has_table() calls with recent sqlalchemy versions
 

chado/feature/__init__.py

@@ -567,6 +567,10 @@ def load_gff(self, gff, analysis_id, organism_id, landmark_type=None, re_protein
             if lm.feature_id not in self._landmark_cache[lm.uniquename]:
                 self._landmark_cache[lm.uniquename].append(lm.feature_id)
 
+        # Preload GO terms
+        db = 'GO'
+        self.ci._preload_dbxref2cvterms(db)
+
         examiner = GFF.GFFExaminer()
         gff_handle = open(gff)
         gff_limits = examiner.available_limits(gff_handle)

setup.py

@@ -6,7 +6,7 @@
 
 setup(
     name="chado",
-    version='2.3.5',
+    version='2.3.6',
     description="Chado library",
     author="Anthony Bretaudeau",
     author_email="[email protected]",

test-data/annot.gff

@@ -169,4 +169,4 @@ scaffold00001	phytozome6	CDS	4061250	4061345	.	+	0	Parent=PAC:18136225;PACid=181
 scaffold00001	phytozome6	CDS	4061417	4061500	.	+	0	Parent=PAC:18136225;PACid=18136225
 scaffold00001	phytozome6	CDS	4061617	4061719	.	+	0	Parent=PAC:18136225;PACid=18136225
 scaffold00001	phytozome6	CDS	4061823	4061905	.	+	2	Parent=PAC:18136225;PACid=18136225
-scaffold00001	phytozome6	three_prime_UTR	4061906	4062210	.	+	.	ID="some_special_cds";Parent=PAC:18136225;PACid=18136225;Derives_from=PAC:18136217;Ontology_term=GO:000001,GO:00002;Target=scaffold00001 120 320 -
+scaffold00001	phytozome6	three_prime_UTR	4061906	4062210	.	+	.	ID="some_special_cds";Parent=PAC:18136225;PACid=18136225;Derives_from=PAC:18136217;Ontology_term=GO:000001,GO:00002,GO:0006812;Target=scaffold00001 120 320 -

test/gff_test.py

@@ -272,11 +272,13 @@ def test_load_gff(self):
 
         terms = {cvt.cvterm.name: cvt.cvterm.dbxref.db_id for cvt in derivesfrom[0].feature_cvterm_collection}
 
-        assert len(terms) == 2, "gff>ontology_term loaded correctly"
+        assert len(terms) == 3, "gff>ontology_term loaded correctly"
         assert '000001' in terms, "gff>ontology_term loaded correctly"
         assert '00002' in terms, "gff>ontology_term loaded correctly"
+        assert 'cation transport' in terms, "gff>ontology_term loaded correctly"
         assert terms['000001'] == dbs['GO'], "gff>ontology_term loaded correctly"
         assert terms['00002'] == dbs['GO'], "gff>ontology_term loaded correctly"
+        assert terms['cation transport'] == dbs['GO'], "gff>ontology_term loaded correctly"
 
         # Target location
         assert len(derivesfrom[0].featureloc_collection) == 2, "gff>target loc ok"
@@ -618,11 +620,13 @@ def test_load_gff_twice(self):
 
         terms = {cvt.cvterm.name: cvt.cvterm.dbxref.db_id for cvt in derivesfrom[0].feature_cvterm_collection}
 
-        assert len(terms) == 2, "gff>ontology_term loaded correctly"
+        assert len(terms) == 3, "gff>ontology_term loaded correctly"
         assert '000001' in terms, "gff>ontology_term loaded correctly"
         assert '00002' in terms, "gff>ontology_term loaded correctly"
+        assert 'cation transport' in terms, "gff>ontology_term loaded correctly"
         assert terms['000001'] == dbs['GO'], "gff>ontology_term loaded correctly"
         assert terms['00002'] == dbs['GO'], "gff>ontology_term loaded correctly"
+        assert terms['cation transport'] == dbs['GO'], "gff>ontology_term loaded correctly"
 
         # Target location
         assert len(derivesfrom[0].featureloc_collection) == 2, "gff>target loc ok"
@@ -1301,11 +1305,13 @@ def test_load_gff_addonly(self):
 
         terms = {cvt.cvterm.name: cvt.cvterm.dbxref.db_id for cvt in derivesfrom[0].feature_cvterm_collection}
 
-        assert len(terms) == 2, "gff>ontology_term loaded correctly"
+        assert len(terms) == 3, "gff>ontology_term loaded correctly"
         assert '000001' in terms, "gff>ontology_term loaded correctly"
         assert '00002' in terms, "gff>ontology_term loaded correctly"
+        assert 'cation transport' in terms, "gff>ontology_term loaded correctly"
         assert terms['000001'] == dbs['GO'], "gff>ontology_term loaded correctly"
         assert terms['00002'] == dbs['GO'], "gff>ontology_term loaded correctly"
+        assert terms['cation transport'] == dbs['GO'], "gff>ontology_term loaded correctly"
 
         # Target location
         assert len(derivesfrom[0].featureloc_collection) == 2, "gff>target loc ok"