refactor

knora/dsplib/utils/xml_upload.py

@@ -211,6 +211,39 @@ def _convert_ark_v0_to_resource_iri(ark: str) -> str:
     return "http://rdfh.ch/" + project_id + "/" + dsp_uuid
 
 
+def parse_xml_file(input_file: str) -> etree.ElementTree:
+    """
+    Parse an XML file with DSP-conform data, remove namespace URI from the elements' names, and transform the special
+    tags <annotation>, <region>, and <link> to their technically correct form <resource restype="Annotation">,
+    <resource restype="Region">, and <resource restype="LinkObj">.
+
+    Args:
+        input_file: path to the XML file
+
+    Returns:
+        the parsed etree.ElementTree
+    """
+    tree = etree.parse(input_file)
+    for elem in tree.getiterator():
+        if not (isinstance(elem, etree._Comment) or isinstance(elem, etree._ProcessingInstruction)):
+            # remove namespace URI in the element's name
+            elem.tag = etree.QName(elem).localname
+        if elem.tag == "annotation":
+            elem.attrib["restype"] = "Annotation"
+            elem.tag = "resource"
+        elif elem.tag == "link":
+            elem.attrib["restype"] = "LinkObj"
+            elem.tag = "resource"
+        elif elem.tag == "region":
+            elem.attrib["restype"] = "Region"
+            elem.tag = "resource"
+
+    # remove unused namespace declarations
+    etree.cleanup_namespaces(tree)
+
+    return tree
+
+
 def xml_upload(input_file: str, server: str, user: str, password: str, imgdir: str, sipi: str, verbose: bool,
                validate_only: bool, incremental: bool) -> bool:
     """
@@ -249,22 +282,7 @@ def xml_upload(input_file: str, server: str, user: str, password: str, imgdir: s
     proj_context = ProjectContext(con=con)
     sipi_server = Sipi(sipi, con.get_token())
 
-    # parse the XML file
-    tree = etree.parse(input_file)
-    for elem in tree.getiterator():
-        if not (isinstance(elem, etree._Comment) or isinstance(elem, etree._ProcessingInstruction)):
-            elem.tag = etree.QName(elem).localname  # remove namespace URI in the element's name
-        if elem.tag == "annotation":
-            elem.attrib["restype"] = "Annotation"
-            elem.tag = "resource"
-        if elem.tag == "link":
-            elem.attrib["restype"] = "LinkObj"
-            elem.tag = "resource"
-        if elem.tag == "region":
-            elem.attrib["restype"] = "Region"
-            elem.tag = "resource"
-    etree.cleanup_namespaces(tree)  # remove unused namespace declarations
-
+    tree = parse_xml_file(input_file)
     root = tree.getroot()
     default_ontology = root.attrib['default-ontology']
     shortcode = root.attrib['shortcode']

test/unittests/test_id_to_iri.py

@@ -3,8 +3,7 @@
 import unittest
 import os
 
-from lxml import etree
-
+from knora.dsplib.utils.xml_upload import parse_xml_file
 from knora.dsplib.utils.id_to_iri import id_to_iri
 
 
@@ -15,7 +14,7 @@ def setUp(self) -> None:
         """Is executed before each test"""
         os.makedirs('testdata/tmp', exist_ok=True)
 
-    def test_invalid_xml_file_name(self):
+    def test_invalid_xml_file_name(self) -> None:
         with self.assertRaises(SystemExit) as cm:
             id_to_iri(xml_file='test.xml',
                       json_file='testdata/test-id2iri-mapping.json',
@@ -24,7 +23,7 @@ def test_invalid_xml_file_name(self):
 
         self.assertEqual(cm.exception.code, 1)
 
-    def test_invalid_json_file_name(self):
+    def test_invalid_json_file_name(self) -> None:
         with self.assertRaises(SystemExit) as cm:
             id_to_iri(xml_file='testdata/test-id2iri-data.xml',
                       json_file='test.json',
@@ -33,22 +32,13 @@ def test_invalid_json_file_name(self):
 
         self.assertEqual(cm.exception.code, 1)
 
-    def test_replace_id_with_iri(self):
+    def test_replace_id_with_iri(self) -> None:
         id_to_iri(xml_file='testdata/test-id2iri-data.xml',
                   json_file='testdata/test-id2iri-mapping.json',
                   out_file=self.out_file,
                   verbose=True)
 
-        tree = etree.parse(self.out_file)
-
-        for elem in tree.getiterator():
-            # skip comments and processing instructions as they do not have namespaces
-            if not (
-                isinstance(elem, etree._Comment)
-                or isinstance(elem, etree._ProcessingInstruction)
-            ):
-                # remove namespace declarations
-                elem.tag = etree.QName(elem).localname
+        tree = parse_xml_file(self.out_file)
 
         resource_elements = tree.xpath("/knora/resource/resptr-prop/resptr")
         result = []

test/unittests/test_xmlupload.py

@@ -4,7 +4,7 @@
 from lxml import etree
 
 from knora.dsplib.models.helpers import BaseError
-from knora.dsplib.utils.xml_upload import _convert_ark_v0_to_resource_iri, _remove_circular_references
+from knora.dsplib.utils.xml_upload import _convert_ark_v0_to_resource_iri, _remove_circular_references, parse_xml_file
 from knora.dsplib.models.xmlresource import XMLResource
 
 
@@ -34,20 +34,7 @@ def test_convert_ark_v0_to_resource_iri(self) -> None:
 
     def test_remove_circular_references(self) -> None:
         # create a list of XMLResources from the test data file
-        tree = etree.parse('testdata/test-data.xml')
-        for elem in tree.getiterator():
-            if not (isinstance(elem, etree._Comment) or isinstance(elem, etree._ProcessingInstruction)):
-                elem.tag = etree.QName(elem).localname  # remove namespace URI in the element's name
-            if elem.tag == "annotation":
-                elem.attrib["restype"] = "Annotation"
-                elem.tag = "resource"
-            if elem.tag == "link":
-                elem.attrib["restype"] = "LinkObj"
-                elem.tag = "resource"
-            if elem.tag == "region":
-                elem.attrib["restype"] = "Region"
-                elem.tag = "resource"
-        etree.cleanup_namespaces(tree)  # remove unused namespace declarations
+        tree = parse_xml_file('testdata/test-data.xml')
         resources = [XMLResource(x, 'testonto') for x in tree.getroot() if x.tag == "resource"]
 
         # get the purged resources and the stashes from the function to be tested

testdata/test-id2iri-mapping.json

@@ -1,4 +1,4 @@
 {
-    "obj_0001": "http://rdfh.ch/082E/ylRvrg7tQI6aVpcTJbVrwg",
-    "obj_0011": "http://rdfh.ch/082E/JK63OpYWTDWNYVOYFN7FdQ"
+    "test_thing_1": "http://rdfh.ch/082E/ylRvrg7tQI6aVpcTJbVrwg",
+    "test_thing_2": "http://rdfh.ch/082E/JK63OpYWTDWNYVOYFN7FdQ"
 }