Merge pull request #135 from GSA/numerical-titles-are-ok

Numerical titles are ok

ckanext/datajson/datajson.py

@@ -460,6 +460,9 @@ def import_stage(self, harvest_object):
             return True
 
         dataset = json.loads(harvest_object.content)
+        # Ensure title is a string for munging/manipulation
+        # https://github.com/GSA/data.gov/issues/4172
+        dataset['title'] = str(dataset['title'])
         schema_version = '1.0'  # default to '1.0'
         is_collection = False
         parent_pkg_id = ''

ckanext/datajson/tests/datajson-samples/numerical-title.data.json

@@ -0,0 +1,43 @@
+{
+  "@type": "dcat:Catalog",
+  "describedBy": "https://project-open-data.cio.gov/v1.1/schema/catalog.json",
+  "conformsTo": "https://project-open-data.cio.gov/v1.1/schema",
+  "@context": "https://project-open-data.cio.gov/v1.1/schema/catalog.jsonld",
+  "dataset": [
+    {
+      "identifier": "null-spatial",
+      "accessLevel": "public",
+      "contactPoint": {
+        "hasEmail": "mailto:[email protected]",
+        "@type": "vcard:Contact",
+        "fn": "Nicole Numbi"
+      },
+      "programCode": [
+        "005:059"
+      ],
+      "description": "Sample dataset. Spatial can be null",
+      "title": 707,
+      "distribution": [
+        {
+          "@type": "dcat:Distribution",
+          "downloadURL": "http://www.dm.usda.gov/foia/docs/Copy%20of%20ECM%20Congressional%20Logs%20FY14.xls",
+          "mediaType": "application/vnd.ms-excel",
+          "title": "Congressional Logs for Fiscal Year 2014"
+        }
+      ],
+      "license": "https://creativecommons.org/publicdomain/zero/1.0/",
+      "bureauCode": [
+        "005:12"
+      ],
+      "modified": "2014-10-03",
+      "publisher": {
+        "@type": "org:Organization",
+        "name": "Department of Agriculture"
+      },
+      "spatial": null,
+      "keyword": [
+        "Congressional Logs"
+      ]
+    }
+  ]
+}

ckanext/datajson/tests/mock_datajson_source.py

@@ -41,6 +41,9 @@ def do_GET(self):
         elif self.path == '/null-spatial':
             self.sample_datajson_file = 'null-spatial.data.json'
             self.test_name = 'null-spatial'
+        elif self.path == '/numerical-title':
+            self.sample_datajson_file = 'numerical-title.data.json'
+            self.test_name = 'numerical-title'
         elif self.path == '/text':
             self.test_name = 'test'
             self.respond('abc123', status=200)

ckanext/datajson/tests/test_datajson_ckan_all_harvester.py

@@ -637,3 +637,10 @@ def test_datajson_null_spatial(self):
         dataset = datasets[0]
         expected_title = "Sample Title NUll Spatial"
         assert dataset.title == expected_title
+
+    def test_datajson_numerical_title(self):
+        url = 'http://127.0.0.1:%s/numerical-title' % self.mock_port
+        datasets = self.run_source(url=url)
+        dataset = datasets[0]
+        expected_title = "707"
+        assert dataset.title == expected_title

setup.py

@@ -10,7 +10,7 @@
 
 setup(
     name='ckanext-datajson',
-    version='0.1.11',
+    version='0.1.12',
     description="CKAN extension to generate /data.json",
     long_description=long_description,
     long_description_content_type='text/markdown',