From 9725815e43159078557534eed7e410a844a72e5d Mon Sep 17 00:00:00 2001
From: Christian Tacke <58549698+ChristianTackeGSI@users.noreply.github.com>
Date: Thu, 3 Aug 2023 11:25:53 +0200
Subject: [PATCH] fix: Only handle "Person"s in contacts

Codemeta also allows "Organizations" in some contact fields
(author, contributor, creator, ...).  And for example
"maintainer" will also allow Organizations in
Codemeta 4.0 [1].

The current conversion code (to zenodo) only needs to
handle Persons. So make sure that only Persons are handled
and everything else is ignored.

[1]: https://github.com/codemeta/codemeta/pull/250#issuecomment-1660417541
---
 eossr/metadata/codemeta2zenodo.py                        | 9 +++++++--
 .../tests/samples/codemeta_contributors_sample.json      | 5 +++++
 2 files changed, 12 insertions(+), 2 deletions(-)

diff --git a/eossr/metadata/codemeta2zenodo.py b/eossr/metadata/codemeta2zenodo.py
index 4233e33..61fd2d8 100644
--- a/eossr/metadata/codemeta2zenodo.py
+++ b/eossr/metadata/codemeta2zenodo.py
@@ -38,6 +38,9 @@ def parse_person_schema_property(person_property, contributor_field):
         dictionary with the correct zenodo syntax for all {author, contributor,
         maintainer}.
     """
+    if person_property["@type"] != "Person":
+        return None
+
     zenodo_person = {}
 
     name = person_property['familyName']
@@ -93,10 +96,12 @@ def add_author_metadata(zenodo_file, codemt_person_entry, person_field):
             zenodo_person = parse_person_schema_property(person_property, person_field)
             # 'name' is the only key that MUST be contained in a
             # person_property at least
-            full_contacts[zenodo_person['name']] = zenodo_person
+            if zenodo_person is not None:
+                full_contacts[zenodo_person['name']] = zenodo_person
     else:
         zenodo_person = parse_person_schema_property(codemt_person_entry, person_field)
-        full_contacts[zenodo_person['name']] = zenodo_person
+        if zenodo_person is not None:
+            full_contacts[zenodo_person['name']] = zenodo_person
 
     # then save each person by field and avoid duplicates
     for person, value in full_contacts.items():
diff --git a/eossr/metadata/tests/samples/codemeta_contributors_sample.json b/eossr/metadata/tests/samples/codemeta_contributors_sample.json
index 6a27134..f4b73a1 100644
--- a/eossr/metadata/tests/samples/codemeta_contributors_sample.json
+++ b/eossr/metadata/tests/samples/codemeta_contributors_sample.json
@@ -38,6 +38,11 @@
     "@id": "https://orcid.org/0000-0000-0000-0000"
   },
   "contributor": [
+    {
+      "@type": "ResearchOrganization",
+      "@id": "https://ror.org/",
+      "name:": "Some Research Org"
+    },
     {
       "@type": "Person",
       "givenName": "Contributor-name",
-- 
GitLab