Release v0.6.4

.bumpversion.cfg

@@ -1,5 +1,5 @@
 [bumpversion]
-current_version = 0.6.3
+current_version = 0.6.4
 commit = True
 tag = True
 

.circleci/config.yml

@@ -28,7 +28,6 @@ jobs:
             pip install --upgrade pip
             pip install --upgrade setuptools wheel
             pip install -r requirements/test.txt
-            pip install -r requirements/extras.txt
 
       - run:
           name: run tests

HISTORY.rst

@@ -3,6 +3,12 @@
 History
 -------
 
+0.6.4 (2019-05-29)
++++++++++++++++++++++++
+
+* (PR #55, 2019-05-29) libs.xml_utils: add ``verify_xml_signature``
+* (PR #54, 2019-05-28) libs: add module ``dataclass_utils``
+
 0.6.3 (2019-05-24)
 +++++++++++++++++++++++
 
@@ -32,7 +38,7 @@ Includes backwards-incompatible changes to data model ``DteDataL2``.
 * (PR #37, 2019-05-08) dte.data_models: alter field ``DteDataL2.firma_documento_dt_naive``
 * (PR #36, 2019-05-08) libs.crypto_utils: add functions
 * (PR #35, 2019-05-07) libs.tz_utils: minor improvements
-* (PR #34, 2019-05-06) docs: Fix `bumpversion` command
+* (PR #34, 2019-05-06) docs: Fix ``bumpversion`` command
 
 0.5.1 (2019-05-03)
 +++++++++++++++++++++++

cl_sii/__init__.py

@@ -5,4 +5,4 @@
 """
 
 
-__version__ = '0.6.3'
+__version__ = '0.6.4'

cl_sii/libs/dataclass_utils.py

@@ -0,0 +1,106 @@
+"""
+Dataclass utils
+===============
+
+Utils for std lib's :class:`dataclasses.Dataclass` classes and instances.
+
+"""
+import dataclasses
+import enum
+
+
+@enum.unique
+class DcDeepComparison(enum.IntEnum):
+
+    """
+    The possible results of a "deep comparison" between 2 dataclass instances.
+
+    .. warning:: The type of both instances must be the same.
+
+    For dataclass instances ``instance_1`` and ``instance_2``, the
+    enum member name should be interpreted as:
+    ``instance_1`` is ``enum_member_name`` of|to|with ``instance_2``
+    e.g. ``instance_1`` is subset of ``instance_2``.
+
+    .. note:: The enum members values are arbitrary.
+
+    """
+
+    EQUAL = 0
+    """
+    For each dataclass attribute A and B have the same value.
+    """
+
+    SUBSET = 11
+    """
+    A and B are not equal, and A's value for each dataclass attribute whose
+    value is not None is equal to B's value for the same attribute.
+    """
+
+    SUPERSET = 12
+    """
+    A and B are not equal, and B's value for each dataclass attribute whose
+    value is not None is equal to A's value for the same attribute.
+    """
+
+    CONFLICTED = -1
+    """
+    For one or more dataclass attributes A and B have a different value.
+    """
+
+
+class DcDeepCompareMixin:
+
+    """
+    Mixin for dataclass instances "deep comparison".
+    """
+
+    def deep_compare_to(self, value: object) -> DcDeepComparison:
+        """
+        Return result of a "deep comparison" against another dataclass instance.
+        """
+        # note: 'is_dataclass' returns True if obj is a dataclass or an instance of a dataclass.
+        if not dataclasses.is_dataclass(self):
+            # TODO: would it be possible to run this check when the **class** is created?
+            raise Exception(
+                "Programming error. Only dataclasses may subclass 'DcDeepCompareMixin'.")
+        # note: 'is_dataclass' returns True if obj is a dataclass or an instance of a dataclass.
+        if not dataclasses.is_dataclass(value):
+            raise TypeError("Value must be a dataclass instance.")
+
+        return _dc_deep_compare_to(self, value)
+
+
+def dc_deep_compare(value_a: object, value_b: object) -> DcDeepComparison:
+    """
+    Return result of a "deep comparison" between dataclass instances.
+    """
+    # note: 'is_dataclass' returns True if obj is a dataclass or an instance of a dataclass.
+    if not dataclasses.is_dataclass(value_a) or not dataclasses.is_dataclass(value_b):
+        raise TypeError("Values must be dataclass instances.")
+
+    return _dc_deep_compare_to(value_a, value_b)
+
+
+def _dc_deep_compare_to(value_a: object, value_b: object) -> DcDeepComparison:
+
+    if type(value_a) != type(value_b):
+        raise TypeError("Values to be compared must be of the same type.")
+
+    if value_a == value_b:
+        return DcDeepComparison.EQUAL
+
+    # Remove dataclass attributes whose value is None.
+    self_dict_clean = {k: v for k, v in dataclasses.asdict(value_a).items() if v is not None}
+    value_dict_clean = {k: v for k, v in dataclasses.asdict(value_b).items() if v is not None}
+
+    if len(self_dict_clean) < len(value_dict_clean):
+        for k, v in self_dict_clean.items():
+            if v != value_dict_clean[k]:
+                return DcDeepComparison.CONFLICTED
+        return DcDeepComparison.SUBSET
+    else:
+        for k, v in value_dict_clean.items():
+            if v != self_dict_clean[k]:
+                return DcDeepComparison.CONFLICTED
+        return DcDeepComparison.SUPERSET

cl_sii/libs/xml_utils.py

@@ -19,20 +19,25 @@
 
 
 """
+import io
 import logging
 import os
-from typing import IO
+from typing import IO, Tuple, Union
 
 import defusedxml
 import defusedxml.lxml
 import lxml.etree
+import signxml
+import signxml.exceptions
 import xml.parsers.expat
 import xml.parsers.expat.errors
 from lxml.etree import ElementBase as XmlElement  # noqa: F401
 # note: 'lxml.etree.ElementTree' is a **function**, not a class.
 from lxml.etree import _ElementTree as XmlElementTree  # noqa: F401
 from lxml.etree import XMLSchema as XmlSchema  # noqa: F401
 
+from . import crypto_utils
+
 
 logger = logging.getLogger(__name__)
 
@@ -111,6 +116,29 @@ class XmlSchemaDocValidationError(Exception):
     """
 
 
+class XmlSignatureInvalid(Exception):
+
+    """
+    XML signature is invalid, for any reason.
+    """
+
+
+class XmlSignatureUnverified(XmlSignatureInvalid):
+
+    """
+    XML signature verification (i.e. digest validation) failed.
+
+    This means the signature is not to be trusted.
+    """
+
+
+class XmlSignatureInvalidCertificate(XmlSignatureInvalid):
+
+    """
+    Certificate validation failed on XML signature processing.
+    """
+
+
 ###############################################################################
 # functions
 ###############################################################################
@@ -323,3 +351,111 @@ def write_xml_doc(xml_doc: XmlElement, output: IO[bytes]) -> None:
         # default: True.
         with_tail=True,
     )
+
+
+def verify_xml_signature(
+    xml_doc: XmlElement,
+    trusted_x509_cert: Union[crypto_utils.X509Cert, crypto_utils._X509CertOpenSsl] = None,
+) -> Tuple[bytes, XmlElementTree, XmlElementTree]:
+    """
+    Verify the XML signature in ``xml_doc``.
+
+    .. note::
+        XML document with more than one signature is not supported.
+
+    If the inputs are ok but the XML signature does not verify,
+    raises :class:`XmlSignatureUnverified`.
+
+    If ``trusted_x509_cert`` is None, it requires that the signature in
+    ``xml_doc`` includes a a valid X.509 **certificate chain** that
+    validates against the *known certificate authorities*.
+
+    If ``trusted_x509_cert`` is given, it must be a **trusted** external
+    X.509 certificate, and the verification will be of whether the XML
+    signature in ``xml_doc`` was signed by ``trusted_x509_cert`` or not;
+    thus **it overrides** any X.509 certificate information included
+    in the signature.
+
+    .. note::
+        It is strongly recommended to validate ``xml_doc`` beforehand
+        (against the corresponding XML schema, using :func:`validate_xml_doc`).
+
+    :param xml_doc:
+    :param trusted_x509_cert: a trusted external X.509 certificate, or None
+    :raises :class:`XmlSignatureInvalidCertificate`:
+        certificate validation failed
+    :raises :class:`XmlSignatureInvalid`:
+        signature is invalid
+    :raises :class:`XmlSchemaDocValidationError`:
+        XML doc is not valid
+    :raises :class:`ValueError`:
+
+    """
+    if not isinstance(xml_doc, XmlElement):
+        raise TypeError("'xml_doc' must be an XML document/element.")
+
+    n_signatures = (
+        len(xml_doc.findall('.//ds:Signature', namespaces=XML_DSIG_NS_MAP))
+        + len(xml_doc.findall('.//dsig11:Signature', namespaces=XML_DSIG_NS_MAP))
+        + len(xml_doc.findall('.//dsig2:Signature', namespaces=XML_DSIG_NS_MAP)))
+
+    if n_signatures > 1:
+        raise NotImplementedError("XML document with more than one signature is not supported.")
+
+    xml_verifier = signxml.XMLVerifier()
+
+    if isinstance(trusted_x509_cert, crypto_utils._X509CertOpenSsl):
+        trusted_x509_cert_open_ssl = trusted_x509_cert
+    elif isinstance(trusted_x509_cert, crypto_utils.X509Cert):
+        trusted_x509_cert_open_ssl = crypto_utils._X509CertOpenSsl.from_cryptography(
+            trusted_x509_cert)
+    elif trusted_x509_cert is None:
+        trusted_x509_cert_open_ssl = None
+    else:
+        # A 'crypto_utils._X509CertOpenSsl' is ok but we prefer 'crypto_utils.X509Cert'.
+        raise TypeError("'trusted_x509_cert' must be a 'crypto_utils.X509Cert' instance, or None.")
+
+    # warning: performance issue.
+    # note: 'signxml.XMLVerifier.verify()' calls 'signxml.util.XMLProcessor.get_root()',
+    #   which converts the data to string, and then reparses it using the same function we use
+    #   in 'parse_untrusted_xml()' ('defusedxml.lxml.fromstring'), but without all the precautions
+    #   we have there. See:
+    #      https://github.com/XML-Security/signxml/blob/v2.6.0/signxml/util/__init__.py#L141-L151
+    #   Considering that, we'd rather write to bytes ourselves and control the process.
+    f = io.BytesIO()
+    write_xml_doc(xml_doc, f)
+    tmp_bytes = f.getvalue()
+
+    try:
+        # note: by passing 'x509_cert' we override any X.509 certificate information supplied
+        #   by the signature itself.
+        result: signxml.VerifyResult = xml_verifier.verify(
+            data=tmp_bytes, require_x509=True, x509_cert=trusted_x509_cert_open_ssl)
+
+    except signxml.exceptions.InvalidDigest as exc:
+        # warning: catch before 'InvalidSignature' (it is the parent of 'InvalidDigest').
+        raise XmlSignatureUnverified(str(exc)) from exc
+
+    except signxml.exceptions.InvalidCertificate as exc:
+        # warning: catch before 'InvalidSignature' (it is the parent of 'InvalidCertificate').
+        raise XmlSignatureInvalidCertificate(str(exc)) from exc
+
+    except signxml.exceptions.InvalidSignature as exc:
+        logger.exception(
+            "Unexpected exception (it should have been an instance of subclass of "
+            "'InvalidSignature'). Error: %s",
+            str(exc))
+        raise XmlSignatureInvalid(str(exc)) from exc
+
+    except signxml.exceptions.InvalidInput as exc:
+        raise ValueError("Invalid input.", str(exc)) from exc
+
+    except lxml.etree.DocumentInvalid as exc:
+        # Simplest and safest way to get the error message (see 'validate_xml_doc()').
+        # Error example:
+        #   "Element '{http://www.w3.org/2000/09/xmldsig#}X509Certificate': '\nabc\n' is not a
+        #   valid value of the atomic type 'xs:base64Binary'., line 30"
+        validation_error_msg = str(exc)
+        raise XmlSchemaDocValidationError(validation_error_msg) from exc
+
+    return result.signed_data, result.signed_xml, result.signature_xml

requirements/release.txt

@@ -6,7 +6,7 @@
 bumpversion==0.5.3
 setuptools==41.0.1
 twine==1.13.0
-wheel==0.33.1
+wheel==0.33.4
 
 # Packages dependencies:
 #   - twine:
@@ -18,6 +18,6 @@ wheel==0.33.1
 #       - tqdm
 pkginfo==1.5.0.1
 readme-renderer==24.0
-requests==2.21.0
+requests==2.22.0
 requests-toolbelt==0.9.1
-tqdm==4.31.1
+tqdm==4.32.1

requirements/test.txt

@@ -1,5 +1,6 @@
 # note: it is mandatory to register all dependencies of the required packages.
 -r base.txt
+-r extras.txt
 
 # Required packages:
 codecov==2.0.15

tests/test_data/xml/trivial-doc.xml

@@ -0,0 +1,24 @@
+<?xml version="1.0"?>
+<data>
+    <!-- This comment is very important! -->
+    <country name="Liechtenstein">
+        <rank>1</rank>
+        <year>2008</year>
+        <gdppc>141100</gdppc>
+        <neighbor name="Austria" direction="E"/>
+        <neighbor name="Switzerland" direction="W"/>
+    </country>
+    <country name="Singapore">
+        <rank>4</rank>
+        <year>2011</year>
+        <gdppc>59900</gdppc>
+        <neighbor name="Malaysia" direction="N"/>
+    </country>
+    <country name="Panama">
+        <rank>68</rank>
+        <year>2011</year>
+        <gdppc>13600</gdppc>
+        <neighbor name="Costa Rica" direction="W"/>
+        <neighbor name="Colombia" direction="E"/>
+    </country>
+</data>