Merge pull request #551 from CDLUC3/develop

Develop

impl/datacite.py

@@ -21,6 +21,7 @@
 
 import django.conf
 import lxml.etree
+import xmltodict
 
 import ezidapp.models.shoulder
 import ezidapp.models.validation
@@ -616,6 +617,103 @@ def dcmsRecordToHtml(record):
     except Exception:
         return None
 
+def removeXMLNamespacePrefix(record):
+    """Remove namespace prefixes from XML elements andd attributes.
+
+    Args:
+        record: The record should be unencoded.
+
+    Returns:
+        XML record or None on error.
+    """
+    try:
+        r = lxml.etree.tostring(
+            lxml.etree.XSLT(
+                lxml.etree.parse(
+                    os.path.join(django.conf.settings.PROJECT_ROOT, "profiles", "remove_ns_prefix.xsl")
+                )
+            )(impl.util.parseXmlString(record)),
+            encoding=str,
+        )
+        return r
+    except Exception:
+        return None
+
+def dcmsRecordToDict(record):
+    """Convert a DataCite Metadata Scheme <http://schema.datacite.org/> record
+    to a dict.
+
+    Args:
+        record: The record should be unencoded.
+
+    Returns:
+        A dict or None on error.
+    """
+    record_dict = xmltodict.parse(str(record))
+    return record_dict
+
+def briefDataciteRecord(record):
+    """Convert a DataCite Metadata Scheme <http://schema.datacite.org/> record
+    to a dictionary in simple DataCite format with data fields for Citation Preview.
+
+    Args:
+        record: Datacite record in XML.
+
+    Returns:
+        A dict or None on error. The dict contains the following keys:
+        - 'datacite.creator'
+        - 'datacite.title'
+        - 'datacite.publisher'
+        - 'datacite.publicationyear'
+        - 'datacite.resourcetype'
+    """
+
+    datacite_dict = dcmsRecordToDict(record)
+    briefDcRecord = {}
+    try:
+        if datacite_dict and 'resource' in datacite_dict:
+            if 'creators' in datacite_dict['resource'] and 'creator' in datacite_dict['resource']['creators']:
+                creator = datacite_dict['resource']['creators']['creator']
+                if isinstance(creator, list):
+                    et_al = ''
+                    for i in range(len(creator)):
+                        if 'creatorName' in creator[i]:
+                            if 'datacite.creator' not in briefDcRecord:
+                                briefDcRecord['datacite.creator'] = get_dict_value_by_key(creator[i]['creatorName'], '#text')
+                            else:
+                                et_al = 'et al.'
+                    if briefDcRecord['datacite.creator'] and et_al != '':
+                        briefDcRecord['datacite.creator'] += f' {et_al}'
+                else:
+                    if 'creatorName' in creator:
+                        briefDcRecord['datacite.creator'] = get_dict_value_by_key(creator['creatorName'], '#text')
+
+            if 'titles' in datacite_dict['resource'] and 'title' in datacite_dict['resource']['titles']:
+                title = datacite_dict['resource']['titles']['title']
+                if isinstance(title, list):
+                    if len(title) > 0:
+                        briefDcRecord['datacite.title'] = get_dict_value_by_key(title[0], '#text')
+                else:
+                    briefDcRecord['datacite.title'] = get_dict_value_by_key(title, '#text')
+
+            if 'publisher' in datacite_dict['resource']:
+                briefDcRecord['datacite.publisher'] = get_dict_value_by_key(datacite_dict['resource']['publisher'], '#text')
+
+            if 'publicationYear' in datacite_dict['resource']:
+                briefDcRecord['datacite.publicationyear'] = datacite_dict['resource']['publicationYear']
+
+            if 'resourceType' in datacite_dict['resource']:
+                briefDcRecord['datacite.resourcetype'] = get_dict_value_by_key(datacite_dict['resource']['resourceType'], '@resourceTypeGeneral')
+    except Exception as ex:
+        print(f'error: {ex} - brief record: {briefDcRecord}')
+
+    return briefDcRecord
+
+def get_dict_value_by_key(input_dict, key):
+    if isinstance(input_dict, dict) and key in input_dict:
+        return input_dict.get(key)
+    else:
+        return input_dict
 
 def crossrefToDatacite(record, overrides=None):
     """Convert a Crossref Deposit Schema <http://help.crossref.org/deposit_schema>

impl/ui_manage.py

@@ -425,9 +425,23 @@ def details(request):
         time.time() - float(id_metadata['_updated']) < 60 * 30
     )
     if d['current_profile'].name == 'datacite' and 'datacite' in id_metadata:
-        r = impl.datacite.dcmsRecordToHtml(id_metadata["datacite"])
+        r = impl.datacite.dcmsRecordToHtml(id_metadata['datacite'])
         if r:
             d['datacite_html'] = r
+
+        converted_rd = impl.datacite.removeXMLNamespacePrefix(id_metadata['datacite'])
+        brief_record = impl.datacite.briefDataciteRecord(converted_rd)
+        brief_record_keys = [
+            'datacite.creator',
+            'datacite.title',
+            'datacite.publicationyear',
+            'datacite.publisher',
+            'datacite.resourcetype',
+        ]
+        for key in brief_record_keys:
+            if key in brief_record:
+                d["identifier"][key] = brief_record[key]
+
     if (
         d['current_profile'].name == 'crossref'
         and 'crossref' in id_metadata

profiles/remove_ns_prefix.xsl

@@ -0,0 +1,28 @@
+<?xml version="1.0" encoding="utf-8"?>
+
+<!-- ==========================================================================
+
+transform an XML document by stripping the namespaces from element names and attributes
+
+=========================================================================== -->
+
+<xsl:stylesheet xmlns:xsl="http://www.w3.org/1999/XSL/Transform" version="1.0">
+  <xsl:output method="xml" indent="yes"/>
+
+  <!-- Template for elements -->
+  <xsl:template match="*">
+    <xsl:element name="{local-name()}" namespace="{namespace-uri()}">
+      <xsl:apply-templates select="@* | node()"/>
+    </xsl:element>
+  </xsl:template>
+
+  <!-- Template for attributes -->
+  <xsl:template match="@*">
+    <xsl:attribute name="{local-name()}" namespace="{namespace-uri()}">
+      <xsl:value-of select="."/>
+    </xsl:attribute>
+  </xsl:template>
+
+</xsl:stylesheet>
+
+

requirements-dev.txt

@@ -23,6 +23,7 @@ requests~=2.31.0
 setuptools~=65.5.1
 simplegist~=1.0.1
 toml~=0.10.2
+xmltodict~=0.13.0
 
 PyMySQL~=0.9.3
 configparser~=5.0.2

requirements.txt

@@ -21,6 +21,7 @@ requests~=2.31.0
 setuptools~=65.5.1
 simplegist~=1.0.1
 toml~=0.10.2
+xmltodict~=0.13.0
 
 PyMySQL~=0.9.3
 configparser~=5.0.2

templates/manage/details.html

@@ -59,7 +59,7 @@ <h2 class="heading__icon-4">{% trans "About the Identifier" %}</h2>
     </div -->
   </div>
 
-  {% if not has_block_data and current_profile.name == 'datacite' %}
+  {% if current_profile.name == 'datacite' %}
   <div class="row vertical-buffer-20">
     <div class="call-out">
       <h2 class="call-out__heading">{% trans "Citation Preview" %}</h2>