Merge pull request #22 from GenomicDataInfrastructure/fix-tags-valida…

…tion fix: improve tags validation
GenomicDataInfrastructure · May 27, 2024 · 4a39868 · 4a39868
2 parents 00e7a43 + 2793ec8
commit 4a39868
Show file tree

Hide file tree

Showing 2 changed files with 15 additions and 7 deletions.
diff --git a/ckanext/fairdatapoint/profiles.py b/ckanext/fairdatapoint/profiles.py
@@ -69,14 +69,19 @@ def validate_tags(values_list: List[Dict]) -> List:
     tags = []
     for item in values_list:
         tag_value = item['name']
-        find_illegal = re.search(illegal_pattern, tag_value)
-        if find_illegal:
-            log.warning(f'Tag {tag_value} contains values other than alphanumeric characters, spaces, hyphens, '
-                        f'underscores or dots, they will be replaces with spaces')
-            tag = {'name': re.sub(illegal_pattern, ' ', tag_value)}
-            tags.append(tag)
+        if len(tag_value) < 2:
+            log.warning(f'Tag {tag_value} is shorter than 2 characters and will be removed')
+        elif len(tag_value) > 100:
+            log.warning(f'Tag {tag_value} is longer than 100 characters and will be removed')
         else:
-            tags.append(item)
+            find_illegal = re.search(illegal_pattern, tag_value)
+            if find_illegal:
+                log.warning(f'Tag {tag_value} contains values other than alphanumeric characters, spaces, hyphens, '
+                            f'underscores or dots, they will be replaces with spaces')
+                tag = {'name': re.sub(illegal_pattern, ' ', tag_value)}
+                tags.append(tag)
+            else:
+                tags.append(item)
     return tags
 
 

diff --git a/ckanext/fairdatapoint/tests/test_profiles.py b/ckanext/fairdatapoint/tests/test_profiles.py
@@ -32,6 +32,9 @@
     ([{"name": "CNS/Brain"}], [{"name": "CNS Brain"}]),
     ([{"name": "COVID-19"}, {"name": "3`-DNA"}], [{"name": "COVID-19"}, {"name": "3 -DNA"}]),
     ([{"name": "something-1.1"}, {"name": "breast cancer"}], [{"name": "something-1.1"}, {"name": "breast cancer"}]),
+    ([{"name": "-"}], []),
+    ([{"name": "It is a ridiculously long (more 100 chars) text for a tag therefore it should be removed from the "
+               "result to prevent CKAN harvester from failing"}], []),
     ([], [])
 ])
 def test_validate_tags(input_tags, expected_tags):