DATS metadata standards; pre loading JSON to avoid parsing errors aft…

…erwards and to preserve remote @context URIs
pangaea-data-publisher · Nov 11, 2024 · d14a7c4 · d14a7c4
1 parent ed7a857
commit d14a7c4
Show file tree

Hide file tree

Showing 2 changed files with 9 additions and 8 deletions.
diff --git a/fuji_server/controllers/fair_check.py b/fuji_server/controllers/fair_check.py
@@ -371,6 +371,9 @@ def harvest_all_metadata(self):
         # self.retrieve_apis_standards()
         # remove duplicates
         if self.namespace_uri:
+            self.namespace_uri = [
+                ns for ns in self.namespace_uri if not isinstance(ns, list) and not isinstance(ns, dict)
+            ]
             self.namespace_uri = list(set(self.namespace_uri))
 
     def harvest_re3_data(self):

diff --git a/fuji_server/helper/metadata_collector_rdf.py b/fuji_server/helper/metadata_collector_rdf.py
@@ -105,11 +105,6 @@ def set_namespaces(self, graph):
             r"https?:\/\/vocab\.nerc\.ac\.uk\/collection\/[A-Z][0-9]+\/current\/",
             r"https?:\/\/purl\.obolibrary\.org\/obo\/[a-z]+(\.owl|#)",
         ]
-        print(type(graph))
-
-        if isinstance(graph, rdflib.ConjunctiveGraph):
-            for c in graph.contexts():
-                print("CONTEXT: ", c)
         try:
             nm = graph.namespace_manager
             possible = set(graph.predicates()).union(graph.objects(None, RDF.type))
@@ -293,16 +288,19 @@ def parse_metadata(self):
                             % (jsonld_source_url)
                         )
                         try:
-                            print("#####################################", type(rdf_response))
                             # pre-check for valid json
                             json_valid = False
                             try:
                                 json_ = json.loads(rdf_response)
                                 if isinstance(json_, dict):
                                     # add @context URIs to namespaces which are otherwise lost
                                     if isinstance(json_.get("@context"), list):
-                                        self.namespaces.extend(json_.get("@context"))
-                                    elif json_.get("@context"):
+                                        for j_ctx in json_.get("@context"):
+                                            if isinstance(j_ctx, str):
+                                                self.namespaces.append(str(j_ctx))
+                                            elif isinstance(j_ctx, dict):
+                                                self.namespaces.extend(j_ctx.values())
+                                    elif isinstance(json_.get("@context"), str):
                                         self.namespaces.append(str(json_.get("@context")))
                                 json_valid = True
                             except Exception: