Merge pull request #865 from dandi/fix/species-url-case

species url is case sensitive
dandi · Jan 18, 2022 · 46d1ea2 · 46d1ea2
2 parents 07b0348 + 3a494cf
commit 46d1ea2
Show file tree

Hide file tree

Showing 3 changed files with 22 additions and 3 deletions.
diff --git a/dandi/metadata.py b/dandi/metadata.py
@@ -397,6 +397,7 @@ def parse_purlobourl(url: str, lookup: ty.Optional[ty.Tuple[str, ...]] = None):
     """
 
     req = requests.get(url, allow_redirects=True)
+    req.raise_for_status()
     doc = parseString(req.text)
     for elfound in doc.getElementsByTagName("Class"):
         if (
@@ -432,7 +433,7 @@ def extract_species(metadata):
                 value_id = value_orig
                 lookup = ("rdfs:label", "oboInOwl:hasExactSynonym")
                 try:
-                    result = parse_purlobourl(value, lookup=lookup)
+                    result = parse_purlobourl(value_orig, lookup=lookup)
                 except ConnectionError:
                     value = None
                 else:
@@ -457,7 +458,8 @@ def extract_species(metadata):
                 "(http://www.ontobee.org/ontology/NCBITaxon) into "
                 "your species field in your NWB file. For example: "
                 "http://purl.obolibrary.org/obo/NCBITaxon_9606 is the "
-                "url for the species Homo sapiens."
+                "url for the species Homo sapiens. Please note that "
+                "this url is case sensitive."
             )
         return models.SpeciesType(identifier=value_id, name=value)
     else:

diff --git a/dandi/tests/data/metadata/metadata2asset_3.json b/dandi/tests/data/metadata/metadata2asset_3.json
@@ -85,7 +85,8 @@
       "genotype": "Typical",
       "species": {
         "schemaKey": "SpeciesType",
-        "identifier": "http://purl.obolibrary.org/obo/NCBITaxon_1234175"
+        "identifier": "http://purl.obolibrary.org/obo/NCBITaxon_1234175",
+        "name": "Cyperus bulbosus"
       }
     }
   ]

diff --git a/dandi/tests/test_metadata.py b/dandi/tests/test_metadata.py
@@ -16,6 +16,7 @@
 from .skip import mark
 from ..metadata import (
     extract_age,
+    extract_species,
     get_metadata,
     metadata2asset,
     parse_age,
@@ -338,7 +339,22 @@ def test_time_extract_gest():
                 "oboInOwl:hasExactSynonym": "Norway rat",
             },
         ),
+        (
+            "http://purl.obolibrary.org/obo/NCBITaxon_28584",
+            {
+                "rdfs:label": "Drosophila suzukii",
+            },
+        ),
     ],
 )
 def test_parseobourl(url, value):
     assert parse_purlobourl(url) == value
+
+
+def test_species():
+    m = {"species": "http://purl.obolibrary.org/obo/NCBITaxon_28584"}
+    assert extract_species(m).json_dict() == {
+        "identifier": "http://purl.obolibrary.org/obo/NCBITaxon_28584",
+        "schemaKey": "SpeciesType",
+        "name": "Drosophila suzukii",
+    }