Merge pull request #46 from fyndata/develop

Release 0.6.2

.bumpversion.cfg

@@ -1,5 +1,5 @@
 [bumpversion]
-current_version = 0.6.1
+current_version = 0.6.2
 commit = True
 tag = True
 

HISTORY.rst

@@ -3,6 +3,12 @@
 History
 -------
 
+0.6.2 (2019-05-15)
++++++++++++++++++++++++
+
+* (PR #45, 2019-05-15) libs.encoding_utils: improve ``clean_base64``
+* (PR #44, 2019-05-15) dte.parse: fix edge case in ``parse_dte_xml``
+
 0.6.1 (2019-05-08)
 +++++++++++++++++++++++
 

cl_sii/__init__.py

@@ -5,4 +5,4 @@
 """
 
 
-__version__ = '0.6.1'
+__version__ = '0.6.2'

cl_sii/dte/parse.py

@@ -21,7 +21,7 @@
 import logging
 import os
 from datetime import date, datetime
-from typing import Tuple
+from typing import Optional, Tuple
 
 from cl_sii.libs import encoding_utils
 from cl_sii.libs import tz_utils
@@ -430,32 +430,37 @@ def parse_dte_xml(xml_doc: XmlElement) -> data_models.DteDataL2:
     # values parsing
     ###########################################################################
 
-    tipo_dte_value = constants.TipoDteEnum(int(tipo_dte_em.text.strip()))
-    folio_value = int(folio_em.text.strip())
-    fecha_emision_value = date.fromisoformat(fecha_emision_em.text.strip())
+    tipo_dte_value = constants.TipoDteEnum(int(_text_strip_or_raise(tipo_dte_em)))
+    folio_value = int(_text_strip_or_raise(folio_em))
+    fecha_emision_value = date.fromisoformat(_text_strip_or_raise(fecha_emision_em))
     fecha_vencimiento_value = None
     if fecha_vencimiento_em is not None:
-        fecha_vencimiento_value = date.fromisoformat(fecha_vencimiento_em.text.strip())
+        fecha_vencimiento_value = date.fromisoformat(
+            _text_strip_or_raise(fecha_vencimiento_em))
 
-    emisor_rut_value = Rut(emisor_rut_em.text.strip())
-    emisor_razon_social_value = emisor_razon_social_em.text.strip()
-    emisor_giro_value = emisor_giro_em.text.strip()
-    emisor_email_value = emisor_email_em.text.strip() if emisor_email_em is not None else None
+    emisor_rut_value = Rut(_text_strip_or_raise(emisor_rut_em))
+    emisor_razon_social_value = _text_strip_or_raise(emisor_razon_social_em)
+    emisor_giro_value = _text_strip_or_raise(emisor_giro_em)
+    emisor_email_value = None
+    if emisor_email_em is not None:
+        emisor_email_value = _text_strip_or_none(emisor_email_em)
 
-    receptor_rut_value = Rut(receptor_rut_em.text.strip())
-    receptor_razon_social_value = receptor_razon_social_em.text.strip()
-    receptor_email_value = receptor_email_em.text.strip() if receptor_email_em is not None else None
+    receptor_rut_value = Rut(_text_strip_or_raise(receptor_rut_em))
+    receptor_razon_social_value = _text_strip_or_raise(receptor_razon_social_em)
+    receptor_email_value = None
+    if receptor_email_em is not None:
+        receptor_email_value = _text_strip_or_none(receptor_email_em)
 
-    monto_total_value = int(monto_total_em.text.strip())
+    monto_total_value = int(_text_strip_or_raise(monto_total_em))
 
     tmst_firma_value = tz_utils.convert_naive_dt_to_tz_aware(
-        dt=datetime.fromisoformat(tmst_firma_em.text),
+        dt=datetime.fromisoformat(_text_strip_or_raise(tmst_firma_em)),
         tz=data_models.DteDataL2.DATETIME_FIELDS_TZ)
 
     signature_signature_value = encoding_utils.decode_base64_strict(
-        signature_signature_value_em.text.strip())
+        _text_strip_or_raise(signature_signature_value_em))
     signature_key_info_x509_cert_der = encoding_utils.decode_base64_strict(
-        signature_key_info_x509_cert_em.text.strip())
+        _text_strip_or_raise(signature_key_info_x509_cert_em))
 
     return data_models.DteDataL2(
         emisor_rut=emisor_rut_value,
@@ -476,6 +481,39 @@ def parse_dte_xml(xml_doc: XmlElement) -> data_models.DteDataL2:
     )
 
 
+def _text_strip_or_none(xml_em: XmlElement) -> Optional[str]:
+    # note: we need the pair of functions '_text_strip_or_none' and '_text_strip_or_raise'
+    #   because, under certain circumstances, an XML tag:
+    #   - with no content -> `xml_em.text` is None instead of ''
+    #   - with leading and/or trailing whitespace -> `xml_em.text` may or may not include that
+
+    if xml_em is None:
+        raise ValueError("Value must be an XML element, not None.")
+
+    stripped_text: Optional[str] = None
+    if xml_em.text is not None:
+        stripped_text = xml_em.text.strip()
+
+    return stripped_text
+
+
+def _text_strip_or_raise(xml_em: XmlElement) -> str:
+    # note: we need the pair of functions '_text_strip_or_none' and '_text_strip_or_raise'
+    #   because, under certain circumstances, an XML tag:
+    #   - with no content -> `xml_em.text` is None instead of ''
+    #   - with leading and/or trailing whitespace -> `xml_em.text` may or may not include that
+
+    if xml_em is None:
+        raise ValueError("Value must be an XML element, not None.")
+
+    if xml_em.text is None:
+        raise ValueError("Text of XML element is None.")
+    else:
+        stripped_text: str = xml_em.text.strip()
+
+    return stripped_text
+
+
 ###############################################################################
 # helpers
 ###############################################################################

cl_sii/libs/encoding_utils.py

@@ -24,7 +24,13 @@ def clean_base64(value: Union[str, bytes]) -> bytes:
         raise TypeError("Value must be str or bytes.")
 
     # remove line breaks and spaces
-    value_base64_bytes_cleaned = value_base64_bytes.replace(b'\n', b'').replace(b' ', b'')
+    # warning: we may only remove characters that are not part of the standard base-64 alphabet
+    #   (or any of its popular alternatives).
+    value_base64_bytes_cleaned = value_base64_bytes \
+        .replace(b'\n', b'') \
+        .replace(b'\r', b'') \
+        .replace(b'\t', b'') \
+        .replace(b' ', b'')
 
     return value_base64_bytes_cleaned
 

tests/test_data/sii-dte/DTE--76354771-K--33--170--cleaned-mod-empty-emails.xml

@@ -0,0 +1,123 @@
+<?xml version='1.0' encoding='ISO-8859-1'?>
+<DTE xmlns="http://www.sii.cl/SiiDte" version="1.0">
+  <!-- O Win32 Chrome 73 central VERSION: v20190227 -->
+<Documento ID="MiPE76354771-13419">
+    <Encabezado>
+      <IdDoc>
+        <TipoDTE>33</TipoDTE>
+        <Folio>170</Folio>
+        <FchEmis>2019-04-01</FchEmis>
+        <TpoTranCompra>1</TpoTranCompra>
+        <TpoTranVenta>1</TpoTranVenta>
+        <FmaPago>2</FmaPago>
+      </IdDoc>
+      <Emisor>
+        <RUTEmisor>76354771-K</RUTEmisor>
+        <RznSoc>INGENIERIA ENACON SPA</RznSoc>
+        <GiroEmis>Ingenieria y Construccion</GiroEmis>
+        <CorreoEmisor></CorreoEmisor>
+        <Acteco>421000</Acteco>
+        <CdgSIISucur>078525666</CdgSIISucur>
+        <DirOrigen>MERCED 753  16 ARBOLEDA DE QUIILOTA</DirOrigen>
+        <CmnaOrigen>QUILLOTA</CmnaOrigen>
+        <CiudadOrigen>QUILLOTA</CiudadOrigen>
+      </Emisor>
+      <Receptor>
+        <RUTRecep>96790240-3</RUTRecep>
+        <RznSocRecep>MINERA LOS PELAMBRES</RznSocRecep>
+        <GiroRecep>EXTRACCION Y PROCESAMIENTO DE COBRE</GiroRecep>
+        <CorreoRecep></CorreoRecep>
+        <Contacto>Felipe Barria</Contacto>
+        <DirRecep>Av. Apoquindo 4001  1802</DirRecep>
+        <CmnaRecep>LAS CONDES</CmnaRecep>
+        <CiudadRecep>SANTIAGO</CiudadRecep>
+      </Receptor>
+      <Totales>
+        <MntNeto>2517900</MntNeto>
+        <TasaIVA>19.00</TasaIVA>
+        <IVA>478401</IVA>
+        <MntTotal>2996301</MntTotal>
+      </Totales>
+    </Encabezado>
+    <Detalle>
+      <NroLinDet>1</NroLinDet>
+      <NmbItem>Tableros electricos 3 tom</NmbItem>
+      <DscItem>as 3p + t; 380v; 50 hz; 32a; 3 tomas monofasicas 2p + t; 240v; 50 hz; 16a; proteccion ip, segun orden de compra de la referencia.-</DscItem>
+      <QtyItem>2.00</QtyItem>
+      <UnmdItem>Unid</UnmdItem>
+      <PrcItem>1258950.00</PrcItem>
+      <MontoItem>2517900</MontoItem>
+    </Detalle>
+    <Referencia>
+      <NroLinRef>1</NroLinRef>
+      <TpoDocRef>801</TpoDocRef>
+      <FolioRef>4510083633</FolioRef>
+      <FchRef>2019-03-22</FchRef>
+    </Referencia>
+<TED version="1.0"><DD><RE>76354771-K</RE><TD>33</TD><F>170</F><FE>2019-04-01</FE><RR>96790240-3</RR><RSR>MINERA LOS PELAMBRES</RSR><MNT>2996301</MNT><IT1>Tableros electricos 3 tom</IT1><CAF version="1.0"><DA><RE>76354771-K</RE><RS>INGENIERIA ENACON SPA</RS><TD>33</TD><RNG><D>170</D><H>170</H></RNG><FA>2019-04-01</FA><RSAPK><M>uv7BUO3yg/7RoMjh1mPXXG/8YIwjtXsu7kcOq7dZQj66QCiY4FVz2fIhF1jaU0GSikq/jq26IFGylGus92OnPQ==</M><E>Aw==</E></RSAPK><IDK>300</IDK></DA><FRMA algoritmo="SHA1withRSA">PI7bw8y0RNUJrGxyhb2gr6BjFtv/Ikyo/6g69wycoXTHSoRML3xvZvOBytreN7REw9JF0Ldoj91RRtaZbH38bA==</FRMA></CAF><TSTED>2019-04-01T01:36:40</TSTED></DD><FRMT algoritmo="SHA1withRSA">DKFS7bNYRpVYLNEII+eyLcBHmNwQIHVkbqgR96wKcnDEcU6NsHQUMUyXpr7ql7xD9iuGkZDmNxHuY+Mq913oSA==</FRMT></TED>
+    <TmstFirma>2019-04-01T01:36:40</TmstFirma>
+
+  </Documento>
+<Signature xmlns="http://www.w3.org/2000/09/xmldsig#">
+<SignedInfo>
+<CanonicalizationMethod Algorithm="http://www.w3.org/TR/2001/REC-xml-c14n-20010315"/>
+<SignatureMethod Algorithm="http://www.w3.org/2000/09/xmldsig#rsa-sha1"/>
+<Reference URI="#MiPE76354771-13419">
+<Transforms>
+<Transform Algorithm="http://www.w3.org/TR/2001/REC-xml-c14n-20010315"/>
+</Transforms>
+<DigestMethod Algorithm="http://www.w3.org/2000/09/xmldsig#sha1"/>
+<DigestValue>ij2Qn6xOc2eRx3hwyO/GrzptoBk=</DigestValue>
+</Reference>
+</SignedInfo>
+<SignatureValue>
+fsYP5p/lNfofAz8POShrJjqXdBTNNtvv4/TWCxbvwTIAXr7BLrlvX3C/Hpfo4viqaxSu1OGFgPnk
+ddDIFwj/ZsVdbdB+MhpKkyha83RxhJpYBVBY3c+y9J6oMfdIdMAYXhEkFw8w63KHyhdf2E9dnbKi
+wqSxDcYjTT6vXsLPrZk=
+</SignatureValue>
+<KeyInfo>
+<KeyValue>
+<RSAKeyValue>
+<Modulus>
+pB4Bs0Op+L0za/zpFQYBiCrVlIOKgULo4uvRLCI5picuxI6X4rE7f3g9XBIZrqtmTUSshmifKLXl
+9T/ScdkuLyIcsHj0QHkbe0LCHSzw1+pH1yTT/dn5NeFVR2InIkL/PzHkjmVJR/M0R50lGJ1W+nqN
+Uavs/9J+gR9BBMs/eYE=
+</Modulus>
+<Exponent>AQAB</Exponent>
+</RSAKeyValue>
+</KeyValue>
+<X509Data>
+<X509Certificate>
+MIIGVDCCBTygAwIBAgIKMUWmvgAAAAjUHTANBgkqhkiG9w0BAQUFADCB0jELMAkGA1UEBhMCQ0wx
+HTAbBgNVBAgTFFJlZ2lvbiBNZXRyb3BvbGl0YW5hMREwDwYDVQQHEwhTYW50aWFnbzEUMBIGA1UE
+ChMLRS1DRVJUQ0hJTEUxIDAeBgNVBAsTF0F1dG9yaWRhZCBDZXJ0aWZpY2Fkb3JhMTAwLgYDVQQD
+EydFLUNFUlRDSElMRSBDQSBGSVJNQSBFTEVDVFJPTklDQSBTSU1QTEUxJzAlBgkqhkiG9w0BCQEW
+GHNjbGllbnRlc0BlLWNlcnRjaGlsZS5jbDAeFw0xNzA5MDQyMTExMTJaFw0yMDA5MDMyMTExMTJa
+MIHXMQswCQYDVQQGEwJDTDEUMBIGA1UECBMLVkFMUEFSQUlTTyAxETAPBgNVBAcTCFF1aWxsb3Rh
+MS8wLQYDVQQKEyZTZXJ2aWNpb3MgQm9uaWxsYSB5IExvcGV6IHkgQ2lhLiBMdGRhLjEkMCIGA1UE
+CwwbSW5nZW5pZXLDrWEgeSBDb25zdHJ1Y2Npw7NuMSMwIQYDVQQDExpSYW1vbiBodW1iZXJ0byBM
+b3BleiAgSmFyYTEjMCEGCSqGSIb3DQEJARYUZW5hY29ubHRkYUBnbWFpbC5jb20wgZ8wDQYJKoZI
+hvcNAQEBBQADgY0AMIGJAoGBAKQeAbNDqfi9M2v86RUGAYgq1ZSDioFC6OLr0SwiOaYnLsSOl+Kx
+O394PVwSGa6rZk1ErIZonyi15fU/0nHZLi8iHLB49EB5G3tCwh0s8NfqR9ck0/3Z+TXhVUdiJyJC
+/z8x5I5lSUfzNEedJRidVvp6jVGr7P/SfoEfQQTLP3mBAgMBAAGjggKnMIICozA9BgkrBgEEAYI3
+FQcEMDAuBiYrBgEEAYI3FQiC3IMvhZOMZoXVnReC4twnge/sPGGBy54UhqiCWAIBZAIBBDAdBgNV
+HQ4EFgQU1dVHhF0UVe7RXIz4cjl3/Vew+qowCwYDVR0PBAQDAgTwMB8GA1UdIwQYMBaAFHjhPp/S
+ErN6PI3NMA5Ts0MpB7NVMD4GA1UdHwQ3MDUwM6AxoC+GLWh0dHA6Ly9jcmwuZS1jZXJ0Y2hpbGUu
+Y2wvZWNlcnRjaGlsZWNhRkVTLmNybDA6BggrBgEFBQcBAQQuMCwwKgYIKwYBBQUHMAGGHmh0dHA6
+Ly9vY3NwLmVjZXJ0Y2hpbGUuY2wvb2NzcDAjBgNVHREEHDAaoBgGCCsGAQQBwQEBoAwWCjEzMTg1
+MDk1LTYwIwYDVR0SBBwwGqAYBggrBgEEAcEBAqAMFgo5NjkyODE4MC01MIIBTQYDVR0gBIIBRDCC
+AUAwggE8BggrBgEEAcNSBTCCAS4wLQYIKwYBBQUHAgEWIWh0dHA6Ly93d3cuZS1jZXJ0Y2hpbGUu
+Y2wvQ1BTLmh0bTCB/AYIKwYBBQUHAgIwge8egewAQwBlAHIAdABpAGYAaQBjAGEAZABvACAARgBp
+AHIAbQBhACAAUwBpAG0AcABsAGUALgAgAEgAYQAgAHMAaQBkAG8AIAB2AGEAbABpAGQAYQBkAG8A
+IABlAG4AIABmAG8AcgBtAGEAIABwAHIAZQBzAGUAbgBjAGkAYQBsACwAIABxAHUAZQBkAGEAbgBk
+AG8AIABoAGEAYgBpAGwAaQB0AGEAZABvACAAZQBsACAAQwBlAHIAdABpAGYAaQBjAGEAZABvACAA
+cABhAHIAYQAgAHUAcwBvACAAdAByAGkAYgB1AHQAYQByAGkAbzANBgkqhkiG9w0BAQUFAAOCAQEA
+mxtPpXWslwI0+uJbyuS9s/S3/Vs0imn758xMU8t4BHUd+OlMdNAMQI1G2+q/OugdLQ/a9Sg3clKD
+qXR4lHGl8d/Yq4yoJzDD3Ceez8qenY3JwGUhPzw9oDpg4mXWvxQDXSFeW/u/BgdadhfGnpwx61Un
++/fU24ZgU1dDJ4GKj5oIPHUIjmoSBhnstEhIr6GJWSTcDKTyzRdqBlaVhenH2Qs6Mw6FrOvRPuud
+B7lo1+OgxMb/Gjyu6XnEaPu7Vq4XlLYMoCD2xrV7WEADaDTm7KcNLczVAYqWSF1WUqYSxmPoQDFY
++kMTThJyCXBlE0NADInrkwWgLLygkKI7zXkwaw==
+</X509Certificate>
+</X509Data>
+</KeyInfo>
+</Signature></DTE>

tests/test_dte_parse.py

@@ -291,6 +291,8 @@ def setUpClass(cls) -> None:
             'test_data/sii-dte/DTE--76354771-K--33--170--cleaned.xml')
         cls.dte_clean_xml_2_xml_bytes = read_test_file_bytes(
             'test_data/sii-dte/DTE--76399752-9--33--25568--cleaned.xml')
+        cls.dte_clean_xml_1b_xml_bytes = read_test_file_bytes(
+            'test_data/sii-dte/DTE--76354771-K--33--170--cleaned-mod-empty-emails.xml')
 
         cls.dte_clean_xml_1_cert_pem_bytes = encoding_utils.clean_base64(
             crypto_utils.remove_pem_cert_header_footer(
@@ -364,6 +366,32 @@ def test_parse_dte_xml_ok_1(self) -> None:
                 receptor_email=None,
             ))
 
+    def test_parse_dte_xml_ok_1b(self) -> None:
+        xml_doc = xml_utils.parse_untrusted_xml(self.dte_clean_xml_1b_xml_bytes)
+
+        parsed_dte = parse_dte_xml(xml_doc)
+        self.assertDictEqual(
+            dict(parsed_dte.as_dict()),
+            dict(
+                emisor_rut=Rut('76354771-K'),
+                tipo_dte=cl_sii.dte.constants.TipoDteEnum.FACTURA_ELECTRONICA,
+                folio=170,
+                fecha_emision_date=date(2019, 4, 1),
+                receptor_rut=Rut('96790240-3'),
+                monto_total=2996301,
+                emisor_razon_social='INGENIERIA ENACON SPA',
+                receptor_razon_social='MINERA LOS PELAMBRES',
+                fecha_vencimiento_date=None,
+                firma_documento_dt=tz_utils.convert_naive_dt_to_tz_aware(
+                    dt=datetime(2019, 4, 1, 1, 36, 40),
+                    tz=DteDataL2.DATETIME_FIELDS_TZ),
+                signature_value=self._TEST_DTE_1_SIGNATURE_VALUE,
+                signature_x509_cert_der=self.dte_clean_xml_1_cert_der,
+                emisor_giro='Ingenieria y Construccion',
+                emisor_email=None,
+                receptor_email=None,
+            ))
+
     def test_parse_dte_xml_ok_2(self) -> None:
         xml_doc = xml_utils.parse_untrusted_xml(self.dte_clean_xml_2_xml_bytes)