def clean_dte_xml_file(input_file_path: str,
output_file_path: str) -> Iterable[bytes]:
with open(input_file_path, mode='rb') as f:
file_bytes = f.read()
xml_doc = xml_utils.parse_untrusted_xml(file_bytes)
xml_doc_cleaned, modified = cl_sii.dte.parse.clean_dte_xml(
xml_doc,
set_missing_xmlns=True,
remove_doc_personalizado=True,
)
# TODO: add exception with a nice message for the caller.
cl_sii.dte.parse.validate_dte_xml(xml_doc_cleaned)
with open(output_file_path, 'w+b') as f:
xml_utils.write_xml_doc(xml_doc_cleaned, f)
with open(output_file_path, mode='rb') as f:
file_bytes_rewritten = f.read()
# note: another way to compute the difference in a similar format is
# `diff -Naur $input_file_path $output_file_path`
file_bytes_diff_gen = difflib.diff_bytes(
dfunc=difflib.unified_diff,
a=file_bytes.splitlines(),
b=file_bytes_rewritten.splitlines())
return file_bytes_diff_gen
def _set_dte_xml_missing_xmlns(xml_doc: XmlElement) -> Tuple[XmlElement, bool]:
# source: name of the XML element without namespace.
# cl_sii/data/ref/factura_electronica/schemas-xml/DTE_v10.xsd#L22 (f57a326)
# cl_sii/data/ref/factura_electronica/schemas-xml/EnvioDTE_v10.xsd#L92 (f57a326)
em_tag_simple = 'DTE'
em_namespace = DTE_XMLNS
em_tag_namespaced = '{%s}%s' % (em_namespace, em_tag_simple)
# Tag of 'DTE' should be ...
assert em_tag_namespaced == '{http://www.sii.cl/SiiDte}DTE'
modified = False
root_em = xml_doc.getroottree().getroot()
root_em_tag = root_em.tag
if root_em_tag == em_tag_namespaced:
pass
elif root_em_tag == em_tag_simple:
modified = True
root_em.set('xmlns', em_namespace)
f = io.BytesIO()
xml_utils.write_xml_doc(xml_doc, f)
new_xml_doc_bytes = f.getvalue()
xml_doc = xml_utils.parse_untrusted_xml(new_xml_doc_bytes)
else:
exc_msg = "XML root element tag does not match the expected simple or namespaced name."
raise Exception(exc_msg, em_tag_simple, em_tag_namespaced, root_em_tag)
return xml_doc, modified
def test_clean_dte_xml_ok_3(self) -> None:
file_bytes = self.dte_bad_xml_3_xml_bytes
xml_doc = xml_utils.parse_untrusted_xml(file_bytes)
self.assertEqual(xml_doc.getroottree().getroot().tag, 'DTE')
with self.assertRaises(xml_utils.XmlSchemaDocValidationError) as cm:
validate_dte_xml(xml_doc)
self.assertSequenceEqual(cm.exception.args, (
"Element 'DTE': No matching global declaration available for the validation root., "
"line 2", ))
xml_doc_cleaned, modified = clean_dte_xml(
xml_doc,
set_missing_xmlns=True,
remove_doc_personalizado=True,
)
self.assertTrue(modified)
# This will not raise.
validate_dte_xml(xml_doc_cleaned)
f = io.BytesIO()
xml_utils.write_xml_doc(xml_doc_cleaned, f)
file_bytes_rewritten = f.getvalue()
del f
xml_doc_rewritten = xml_utils.parse_untrusted_xml(file_bytes_rewritten)
validate_dte_xml(xml_doc_rewritten)
expected_file_bytes_diff = (
b'--- \n',
b'+++ \n',
b'@@ -1,5 +1,5 @@\n',
b'-<?xml version="1.0" encoding="windows-1252"?>',
b'-<DTE version="1.0">',
b"+<?xml version='1.0' encoding='WINDOWS-1252'?>",
b'+<DTE xmlns="http://www.sii.cl/SiiDte" version="1.0">',
b' <Documento ID="DTE-33-2336600">',
b' <Encabezado>',
b' <IdDoc>',
)
file_bytes_diff_gen = difflib.diff_bytes(
dfunc=difflib.unified_diff,
a=file_bytes.splitlines(),
b=file_bytes_rewritten.splitlines())
self.assertSequenceEqual(
[diff_line for diff_line in file_bytes_diff_gen],
expected_file_bytes_diff)
def test_clean_dte_xml_ok_2(self) -> None:
file_bytes = self.dte_bad_xml_2_xml_bytes
xml_doc = xml_utils.parse_untrusted_xml(file_bytes)
self.assertEqual(
xml_doc.getroottree().getroot().tag,
'DTE')
with self.assertRaises(xml_utils.XmlSchemaDocValidationError) as cm:
validate_dte_xml(xml_doc)
self.assertSequenceEqual(
cm.exception.args,
("Element 'DTE': No matching global declaration available for the validation root., "
"line 2", )
)
xml_doc_cleaned, modified = clean_dte_xml(
xml_doc,
set_missing_xmlns=True,
remove_doc_personalizado=True,
)
self.assertTrue(modified)
# This will not raise.
validate_dte_xml(xml_doc_cleaned)
f = io.BytesIO()
xml_utils.write_xml_doc(xml_doc_cleaned, f)
file_bytes_rewritten = f.getvalue()
del f
xml_doc_rewritten = xml_utils.parse_untrusted_xml(file_bytes_rewritten)
validate_dte_xml(xml_doc_rewritten)
expected_file_bytes_diff = (
b'--- \n',
b'+++ \n',
b'@@ -1,5 +1,5 @@\n',
b'-<?xml version="1.0" encoding="ISO-8859-1"?>',
b'-<DTE version="1.0">',
b"+<?xml version='1.0' encoding='ISO-8859-1'?>",
b'+<DTE xmlns="http://www.sii.cl/SiiDte" version="1.0">',
b' <!-- O Win32 Chrome 73 central VERSION: v20190227 -->',
b' <Documento ID="MiPE76399752-6048">',
b' <Encabezado>',
b'@@ -64,13 +64,13 @@\n',
b' </Documento>',
b' <Signature xmlns="http://www.w3.org/2000/09/xmldsig#">',
b' <SignedInfo>',
b'-<CanonicalizationMethod Algorithm="http://www.w3.org/TR/2001/REC-xml-c14n-20010315" />', # noqa: E501
b'-<SignatureMethod Algorithm="http://www.w3.org/2000/09/xmldsig#rsa-sha1" />',
b'+<CanonicalizationMethod Algorithm="http://www.w3.org/TR/2001/REC-xml-c14n-20010315"/>', # noqa: E501
b'+<SignatureMethod Algorithm="http://www.w3.org/2000/09/xmldsig#rsa-sha1"/>',
b' <Reference URI="#MiPE76399752-6048">',
b' <Transforms>',
b'-<Transform Algorithm="http://www.w3.org/TR/2001/REC-xml-c14n-20010315" />',
b'+<Transform Algorithm="http://www.w3.org/TR/2001/REC-xml-c14n-20010315"/>',
b' </Transforms>',
b'-<DigestMethod Algorithm="http://www.w3.org/2000/09/xmldsig#sha1" />',
b'+<DigestMethod Algorithm="http://www.w3.org/2000/09/xmldsig#sha1"/>',
b' <DigestValue>tk/D3mfO/KtdWyFXYZHe7dtYijg=</DigestValue>',
b' </Reference>',
b' </SignedInfo>',
)
file_bytes_diff_gen = difflib.diff_bytes(
dfunc=difflib.unified_diff,
a=file_bytes.splitlines(),
b=file_bytes_rewritten.splitlines())
self.assertSequenceEqual(
[diff_line for diff_line in file_bytes_diff_gen],
expected_file_bytes_diff
)