From f9c711f77bba992e6e9e1d75929d35e8da828f61 Mon Sep 17 00:00:00 2001
From: Martin Czygan <martin.czygan@gmail.com>
Date: Tue, 7 Jan 2020 15:20:25 +0100
Subject: datacite: adding datacite-specific extra metadata

* attributes.metadataVersion
* attributes.schemaVersion
* attributes.version (source dependent values, follows suggestions in
https://schema.datacite.org/meta/kernel-4.3/doc/DataCite-MetadataKernel_v4.3.pdf#page=26,
but values vary)

Furthermore:

* attributes.types.resourceTypeGeneral
* attributes.types.resourceType
---
 python/fatcat_tools/importers/datacite.py | 28 ++++++++++++++++++++++++++++
 1 file changed, 28 insertions(+)

(limited to 'python/fatcat_tools/importers')

diff --git a/python/fatcat_tools/importers/datacite.py b/python/fatcat_tools/importers/datacite.py
index d7fbd269..c2725aeb 100644
--- a/python/fatcat_tools/importers/datacite.py
+++ b/python/fatcat_tools/importers/datacite.py
@@ -607,6 +607,25 @@ class DataciteImporter(EntityImporter):
         if attributes.get('subjects'):
             extra_datacite['subjects'] = attributes['subjects']
 
+        # Include version information.
+        metadata_version = attributes.get('metadataVersion') or ''
+        schema_version = attributes.get('schemaVersion') or ''
+
+        if metadata_version:
+            extra_datacite['metadataVersion'] = metadata_version
+        if schema_version:
+            extra_datacite['schemaVersion'] = schema_version
+
+        # Include resource types.
+        types = attributes.get('types', {}) or {}
+        resource_type = types.get('resourceType', '') or ''
+        resource_type_general = types.get('resourceTypeGeneral', '') or ''
+
+        if resource_type:
+            extra_datacite['resourceType'] = resource_type
+        if resource_type_general:
+            extra_datacite['resourceTypeGeneral'] = resource_type_general
+
         # Include certain relations from relatedIdentifiers. Keeping the
         # original structure of data here, which is a list of dicts, with
         # relation type, identifer and identifier type (mostly).
@@ -625,6 +644,14 @@ class DataciteImporter(EntityImporter):
 
         extra = dict()
 
+        # "1.0.0", "v1.305.2019", "Final", "v1.0.0", "v0.3.0", "1", "0.19.0",
+        # "3.1", "v1.1", "{version}", "4.0", "10329", "11672", "11555",
+        # "v1.4.5", "2", "V1", "v3.0", "v0", "v0.6", "11124", "v1.0-beta", "1st
+        # Edition", "20191024", "v2.0.0", "v0.9.3", "10149", "2.0", null,
+        # "v0.1.1", "3.0", "1.0", "3", "v1.12.2", "20191018", "v0.3.1", "v1.0",
+        # "10161", "10010691", "10780", # "Presentación"
+        version = attributes.get('version')
+
         # top-level extra keys
         if not container_id and container_name:
             extra['container_name'] = container_name
@@ -666,6 +693,7 @@ class DataciteImporter(EntityImporter):
             refs=refs,
             extra=extra,
             license_slug=license_slug,
+            version=version,
         )
         return re
 
-- 
cgit v1.2.3