support stdout_lxml for advanced functionality

* lxml is in optional requirements

Signed-off-by: Aleksei Stepanov <penguinolog@gmail.com>

.pylintrc

@@ -3,7 +3,7 @@
 # A comma-separated list of package or module names from where C extensions may
 # be loaded. Extensions are loading into the active Python interpreter and may
 # run arbitrary code.
-extension-pkg-whitelist=
+extension-pkg-whitelist=lxml.etree
 
 # Add files or directories to the blacklist. They should be base names, not
 # paths.
@@ -66,15 +66,13 @@ enable=all
 
 disable=import-star-module-level,
         raw-checker-failed,
-        bad-inline-option,
         locally-disabled,
         file-ignored,
         suppressed-message,
         use-symbolic-message-instead,
         input-builtin,
         round-builtin,
         eq-without-hash,
-        exception-message-attribute,
         exception-escape,
         comprehension-escape,
         similarities,

CI_REQUIREMENTS.txt

@@ -1,2 +1,3 @@
 mock  # no assert_called_once in py35
 -r requirements.txt
+lxml

README.rst

@@ -224,6 +224,8 @@ Execution result object has a set of useful properties:
 
 * `stdout_xml` - STDOUT decoded as XML to `ElementTree` using `defusedxml` library.
 
+* `stdout_lxml` - STDOUT decoded as XML to `ElementTree` using `lxml` library. Accessible only if lxml library installed. Can be insecure.
+
 * `timestamp` -> `typing.Optional(datetime.datetime)`. Timestamp for received exit code.
 
 SSHClient specific

doc/source/ExecResult.rst

@@ -154,6 +154,16 @@ API: ExecResult
         :rtype: xml.etree.ElementTree.Element
         :raises DeserializeValueError: STDOUT can not be deserialized as XML
 
+    .. py:attribute:: stdout_lxml
+
+        XML from stdout using lxml.
+
+        :rtype: lxml.etree.Element
+        :raises DeserializeValueError: STDOUT can not be deserialized as XML
+        :raises AttributeError: lxml is not installed
+
+        .. note:: Can be insecure.
+
     .. py:method:: read_stdout(src=None, log=None, verbose=False)
 
         Read stdout file-like object to stdout.

exec_helpers/exec_result.py

@@ -34,6 +34,11 @@
 from exec_helpers import exceptions
 from exec_helpers import proc_enums
 
+try:
+    import lxml.etree  # type: ignore  # nosec
+except ImportError:
+    lxml = None  # pylint: disable=invalid-name
+
 LOGGER = logging.getLogger(__name__)
 
 
@@ -496,6 +501,8 @@ def __deserialize(self, fmt: str) -> typing.Any:
                 return yaml.safe_load(self.stdout_str)
             if fmt == "xml":
                 return defusedxml.ElementTree.fromstring(bytes(self.stdout_bin))
+            if fmt == "lxml":
+                return lxml.etree.fromstring(bytes(self.stdout_bin))  # nosec
         except Exception as e:
             tmpl = "{{self.cmd}} stdout is not valid {fmt}:\n" "{{stdout!r}}\n".format(fmt=fmt)
             LOGGER.exception(tmpl.format(self=self, stdout=self.stdout_str))
@@ -530,17 +537,32 @@ def stdout_yaml(self) -> typing.Any:
 
     @property
     def stdout_xml(self) -> xml.etree.ElementTree.Element:
-        """YAML from stdout.
+        """XML from stdout.
 
         :rtype: xml.etree.ElementTree.Element
         :raises DeserializeValueError: STDOUT can not be deserialized as XML
         """
         with self.stdout_lock:
             return self.__deserialize(fmt="xml")  # type: ignore
 
+    @property
+    def stdout_lxml(self) -> "lxml.etree.Element":
+        """XML from stdout using lxml.
+
+        :rtype: lxml.etree.Element
+        :raises DeserializeValueError: STDOUT can not be deserialized as XML
+        :raises AttributeError: lxml is not installed
+
+        .. note:: Can be insecure.
+        """
+        if lxml is None:
+            raise AttributeError("lxml is not installed -> attribute is not functional.")
+        with self.stdout_lock:
+            return self.__deserialize(fmt="lxml")
+
     def __dir__(self) -> typing.List[str]:
         """Override dir for IDE and as source for getitem checks."""
-        return [
+        content = [
             "cmd",
             "stdout",
             "stderr",
@@ -558,6 +580,9 @@ def __dir__(self) -> typing.List[str]:
             "stdout_xml",
             "lock",
         ]
+        if lxml is not None:
+            content.append("stdout_lxml")
+        return content
 
     def __getitem__(self, item: str) -> typing.Any:
         """Dict like get data.

requirements.txt

@@ -4,6 +4,6 @@ six>=1.10.0  # MIT
 threaded>=2.0  # Apache-2.0
 PyYAML>=3.12  # MIT
 advanced-descriptors>=1.0  # Apache-2.0
-typing >= 3.6 ; python_version < "3.7"
-psutil >= 5.0
-defusedxml
+typing >= 3.6 ; python_version < "3.7"  # PSF
+psutil >= 5.0  # BSD
+defusedxml  # PSF

setup.py

@@ -255,6 +255,9 @@ def get_simple_vars_from_src(src):
     ],
     use_scm_version=True,
     install_requires=REQUIRED,
+    extras_require={
+        "lxml": ["lxml!=3.7.0"]
+    },
     package_data={"exec_helpers": ["py.typed"]},
 )
 if cythonize is not None:

test/test_exec_result.py

@@ -26,6 +26,11 @@
 import exec_helpers
 from exec_helpers import proc_enums
 
+try:
+    import lxml.etree
+except ImportError:
+    lxml = None
+
 cmd = "ls -la | awk '{print $1}'"
 
 
@@ -305,3 +310,17 @@ def test_stdout_xml(self):
         self.assertEqual(
             xml.etree.ElementTree.tostring(expect), xml.etree.ElementTree.tostring(result.stdout_xml)
         )
+
+    @unittest.skipIf(lxml is None, "no lxml installed")
+    def test_stdout_lxml(self):
+        result = exec_helpers.ExecResult(
+            "test",
+            stdout=[
+                b"<?xml version='1.0'?>\n",
+                b'<data>123</data>\n',
+            ]
+        )
+        expect = lxml.etree.fromstring(b"<?xml version='1.0'?>\n<data>123</data>\n")
+        self.assertEqual(
+            lxml.etree.tostring(expect), lxml.etree.tostring(result.stdout_lxml)
+        )