use only lxml for XML handling (#863)

Author: tomkralidis

.github/workflows/main.yml

@@ -8,9 +8,7 @@ jobs:
     strategy:
       matrix:
         python-version: ['3.7', '3.8', '3.9']
-        lxml: [true, false]
     env:
-        LXML: ${{ matrix.lxml }}
         GITHUB_TOKEN: ${{ secrets.GITHUB_TOKEN }}
         COVERALLS_SERVICE_NAME: github
     steps:
@@ -28,8 +26,6 @@ jobs:
         pip3 install -e .
         pip3 install -r requirements.txt
         pip3 install -r requirements-dev.txt
-        echo "LXML => $LXML"
-        if [ "$LXML" == "true" ]; then pip install lxml; fi
     - name: run tests ⚙️
       run: python3 -m pytest
     - name: run coveralls ⚙️

docs/en/installation.rst

@@ -4,7 +4,7 @@ Installation
 Requirements
 ------------
 
-OWSLib requires a Python interpreter, as well as `ElementTree <https://docs.python.org/2/library/xml.etree.elementtree.html>`_ or `lxml <http://lxml.de>`_ for XML parsing.
+OWSLib requires a Python interpreter, as well as `lxml <https://lxml.de>`_ for XML parsing.
 
 Install
 -------

etc/debian/control

@@ -9,5 +9,5 @@ Homepage: http://geopython.github.com/OWSLib/
 
 Package: python-owslib
 Architecture: all
-Depends: ${misc:Depends}, debconf, python (>=2.7), python-lxml
+Depends: ${misc:Depends}, debconf, python (>=3), python-lxml
 Description: OWSLib is a Python package for client programming with Open Geospatial Consortium (OGC) web service (hence OWS) interface standards, and their related content models.

owslib/catalogue/csw2.py

@@ -226,11 +226,6 @@ def getrecords(self, qtype=None, keywords=[], typenames='csw:Record', propertyna
         else:
             # construct request
             node0 = self._setrootelement('csw:GetRecords')
-            if etree.__name__ != 'lxml.etree':  # apply nsmap manually
-                node0.set('xmlns:ows', namespaces['ows'])
-                node0.set('xmlns:gmd', namespaces['gmd'])
-                node0.set('xmlns:dif', namespaces['dif'])
-                node0.set('xmlns:fgdc', namespaces['fgdc'])
             node0.set('outputSchema', outputschema)
             node0.set('outputFormat', format)
             node0.set('version', self.version)
@@ -354,11 +349,6 @@ def getrecords2(self, constraints=[], sortby=None, typenames='csw:Record', esn='
         else:
             # construct request
             node0 = self._setrootelement('csw:GetRecords')
-            if etree.__name__ != 'lxml.etree':  # apply nsmap manually
-                node0.set('xmlns:ows', namespaces['ows'])
-                node0.set('xmlns:gmd', namespaces['gmd'])
-                node0.set('xmlns:dif', namespaces['dif'])
-                node0.set('xmlns:fgdc', namespaces['fgdc'])
             node0.set('outputSchema', outputschema)
             node0.set('outputFormat', format)
             node0.set('version', self.version)
@@ -622,10 +612,7 @@ def _setidentifierkey(self, el):
             return el
 
     def _setrootelement(self, el):
-        if etree.__name__ == 'lxml.etree':  # apply nsmap
-            return etree.Element(util.nspath_eval(el, namespaces), nsmap=namespaces)
-        else:
-            return etree.Element(util.nspath_eval(el, namespaces))
+        return etree.Element(util.nspath_eval(el, namespaces), nsmap=namespaces)
 
     def _setconstraint(self, parent, qtype=None, propertyname='csw:AnyText', keywords=[], bbox=None, cql=None,
                        identifier=None):

owslib/catalogue/csw3.py

@@ -242,11 +242,6 @@ def getrecords(self, constraints=[], sortby=None, typenames='csw30:Record', esn=
         else:
             # construct request
             node0 = self._setrootelement('csw30:GetRecords')
-            if etree.__name__ != 'lxml.etree':  # apply nsmap manually
-                node0.set('xmlns:ows110', namespaces['ows110'])
-                node0.set('xmlns:gmd', namespaces['gmd'])
-                node0.set('xmlns:dif', namespaces['dif'])
-                node0.set('xmlns:fgdc', namespaces['fgdc'])
             node0.set('outputSchema', outputschema)
             node0.set('outputFormat', format)
             node0.set('version', self.version)
@@ -516,10 +511,7 @@ def _setidentifierkey(self, el):
             return el
 
     def _setrootelement(self, el):
-        if etree.__name__ == 'lxml.etree':  # apply nsmap
-            return etree.Element(util.nspath_eval(el, namespaces), nsmap=namespaces)
-        else:
-            return etree.Element(util.nspath_eval(el, namespaces))
+        return etree.Element(util.nspath_eval(el, namespaces), nsmap=namespaces)
 
     def _setconstraint(self, parent, qtype=None, propertyname='csw30:AnyText', keywords=[], bbox=None, cql=None,
                        identifier=None):

owslib/etree.py

@@ -4,37 +4,33 @@
 # Contact email: sgillies@frii.com
 # =============================================================================
 
+
+from lxml import etree
+from lxml.etree import ParseError
+ElementType = etree._Element
+
 from owslib.namespaces import Namespaces
 
 
-def patch_well_known_namespaces(etree_module):
-    """Monkey patches the etree module to add some well-known namespaces."""
+def patch_well_known_namespaces():
+    """Monkey patches lxml.etree to add some well-known namespaces."""
 
     ns = Namespaces()
 
     try:
-        register_namespace = etree_module.register_namespace
+        register_namespace = etree.register_namespace
     except AttributeError:
-        etree_module._namespace_map
+        etree._namespace_map
 
         def register_namespace(prefix, uri):
-            etree_module._namespace_map[uri] = prefix
+            etree._namespace_map[uri] = prefix
 
     for k, v in list(ns.get_namespaces().items()):
         register_namespace(k, v)
 
+    etree.set_default_parser(
+        parser=etree.XMLParser(resolve_entities=False)
+    )
 
-# try to find lxml or elementtree
-try:
-    from lxml import etree
-    from lxml.etree import ParseError
-    ElementType = etree._Element
-except ImportError:
-    import xml.etree.ElementTree as etree
-    ElementType = etree.Element
-    try:
-        from xml.etree.ElementTree import ParseError
-    except ImportError:
-        from xml.parsers.expat import ExpatError as ParseError
 
-patch_well_known_namespaces(etree)
+patch_well_known_namespaces()

owslib/feature/schema.py

@@ -13,7 +13,7 @@
 from urllib.parse import urlencode, parse_qsl
 from owslib.etree import etree
 from owslib.namespaces import Namespaces
-from owslib.util import which_etree, findall, Authentication, openURL
+from owslib.util import findall, Authentication, openURL
 
 MYNS = Namespaces()
 XS_NAMESPACE = MYNS.get_namespace("xs")

owslib/util.py

@@ -277,11 +277,8 @@ def nspath_eval(xpath, namespaces):
 
 def cleanup_namespaces(element):
     """ Remove unused namespaces from an element """
-    if etree.__name__ == 'lxml.etree':
-        etree.cleanup_namespaces(element)
-        return element
-    else:
-        return etree.fromstring(etree.tostring(element))
+    etree.cleanup_namespaces(element)
+    return element
 
 
 def add_namespaces(root, ns_keys):
@@ -292,35 +289,34 @@ def add_namespaces(root, ns_keys):
 
     ns_keys = [(x, namespaces.get_namespace(x)) for x in ns_keys]
 
-    if etree.__name__ != 'lxml.etree':
-        # We can just add more namespaces when not using lxml.
-        # We can't re-add an existing namespaces.  Get a list of current
-        # namespaces in use
-        existing_namespaces = set()
-        for elem in root.iter():
-            if elem.tag[0] == "{":
-                uri, tag = elem.tag[1:].split("}")
-                existing_namespaces.add(namespaces.get_namespace_from_url(uri))
-        for key, link in ns_keys:
-            if link is not None and key not in existing_namespaces:
-                root.set("xmlns:%s" % key, link)
-        return root
-    else:
-        # lxml does not support setting xmlns attributes
-        # Update the elements nsmap with new namespaces
-        new_map = root.nsmap
-        for key, link in ns_keys:
-            if link is not None:
-                new_map[key] = link
-        # Recreate the root element with updated nsmap
-        new_root = etree.Element(root.tag, nsmap=new_map)
-        # Carry over attributes
-        for a, v in list(root.items()):
-            new_root.set(a, v)
-        # Carry over children
-        for child in root:
-            new_root.append(deepcopy(child))
-        return new_root
+    # lxml does not support setting xmlns attributes
+    # Update the elements nsmap with new namespaces
+    new_map = root.nsmap
+    for key, link in ns_keys:
+        if link is not None:
+            new_map[key] = link
+    # Recreate the root element with updated nsmap
+    new_root = etree.Element(root.tag, nsmap=new_map)
+    # Carry over attributes
+    for a, v in list(root.items()):
+        new_root.set(a, v)
+    # Carry over children
+    for child in root:
+        new_root.append(deepcopy(child))
+    return new_root
+
+    # We can just add more namespaces when not using lxml.
+    # We can't re-add an existing namespaces.  Get a list of current
+    # namespaces in use
+    existing_namespaces = set()
+    for elem in root.iter():
+        if elem.tag[0] == "{":
+            uri, tag = elem.tag[1:].split("}")
+            existing_namespaces.add(namespaces.get_namespace_from_url(uri))
+    for key, link in ns_keys:
+        if link is not None and key not in existing_namespaces:
+            root.set("xmlns:%s" % key, link)
+    return root
 
 
 def getXMLInteger(elem, tag):
@@ -519,21 +515,14 @@ def element_to_string(element, encoding=None, xml_declaration=False):
     if encoding is None:
         encoding = "ISO-8859-1"
 
-    if etree.__name__ == 'lxml.etree':
-        if xml_declaration:
-            if encoding in ['unicode', 'utf-8']:
-                output = '<?xml version="1.0" encoding="utf-8" standalone="no"?>\n{}'.format(
-                    etree.tostring(element, encoding='unicode'))
-            else:
-                output = etree.tostring(element, encoding=encoding, xml_declaration=True)
+    if xml_declaration:
+        if encoding in ['unicode', 'utf-8']:
+            output = '<?xml version="1.0" encoding="utf-8" standalone="no"?>\n{}'.format(
+                etree.tostring(element, encoding='unicode'))
         else:
-            output = etree.tostring(element)
+            output = etree.tostring(element, encoding=encoding, xml_declaration=True)
     else:
-        if xml_declaration:
-            output = '<?xml version="1.0" encoding="{}" standalone="no"?>\n{}'.format(
-                encoding, etree.tostring(element, encoding=encoding))
-        else:
-            output = etree.tostring(element)
+        output = etree.tostring(element)
 
     return output
 
@@ -777,21 +766,6 @@ def bind_url(url):
 log.addHandler(NullHandler())
 
 
-def which_etree():
-    """decipher which etree library is being used by OWSLib"""
-
-    which_etree = None
-
-    if 'lxml' in etree.__file__:
-        which_etree = 'lxml.etree'
-    elif 'xml/etree' in etree.__file__:
-        which_etree = 'xml.etree'
-    elif 'elementree' in etree.__file__:
-        which_etree = 'elementtree.ElementTree'
-
-    return which_etree
-
-
 def findall(root, xpath, attribute_name=None, attribute_value=None):
     """Find elements recursively from given root element based on
     xpath and possibly given attribute

requirements.txt

@@ -1,5 +1,6 @@
+dataclasses; python_version < '3.7'
+lxml
 python-dateutil>=1.5
 pytz
-requests>=1.0
 pyyaml
-dataclasses; python_version < '3.7'
+requests>=1.0