Fixed #111 : Covering xml:lang on /text

Capitains · Sep 23, 2017 · c861ccb · c861ccb
1 parent db9c094
commit c861ccb
Show file tree

Hide file tree

Showing 6 changed files with 144 additions and 20 deletions.
diff --git a/CHANGES.txt b/CHANGES.txt
@@ -1,3 +1,8 @@
+## 1.1.5 - 2017-09-23
+
+- ( Added a fix for #111 ) Refactored language test unit to cover all cases of guidelines
+- Started refactoring a little some unit tests, with example generations.
+
 ## 1.1.4 - 2017-09-19
 
 - Updated RNG checks for compatibility with _JAVA_OPTIONS environment variable on Travis

diff --git a/HookTest/capitains_units/cts.py b/HookTest/capitains_units/cts.py
@@ -585,27 +585,25 @@ def language(self):
         """ Tests to make sure an xml:lang element is on the correct node
         """
         if self.scheme == "epidoc":
-            try:
-                self.lang = self.xml.xpath('/tei:TEI/tei:text/tei:body/tei:div[@type="edition" or @type="translation" or @type="commentary"]',
-                                           namespaces=TESTUnit.NS)[0].get('{http://www.w3.org/XML/1998/namespace}lang')
-            except:
-                self.lang = ''
-            if self.lang == '' or self.lang is None:
-                self.lang = 'UNK'
-                yield False
-            else:
-                yield True
+            urns_holding_node = self.xml.xpath(
+                "//tei:text/tei:body/tei:div"
+                "[@type='edition' or @type='translation' or @type='commentary']"
+                "[starts-with(@n, 'urn:cts:')]",
+                namespaces=TESTUnit.NS
+            )
         elif self.scheme == "tei":
-            try:
-                self.lang = self.xml.xpath('/tei:TEI/tei:text/tei:body',
-                                           namespaces=TESTUnit.NS)[0].get('{http://www.w3.org/XML/1998/namespace}lang')
-            except:
-                self.lang = ''
-            if self.lang == '' or self.lang is None:
-                self.lang = 'UNK'
-                yield False
-            else:
-                yield True
+            urns_holding_node = self.xml.xpath("//tei:text/tei:body[starts-with(@n, 'urn:cts:')]", namespaces=TESTUnit.NS) + \
+                    self.xml.xpath("//tei:text[starts-with(@xml:base, 'urn:cts:')]", namespaces=TESTUnit.NS)
+
+        try:
+            self.lang = urns_holding_node[0].get('{http://www.w3.org/XML/1998/namespace}lang')
+        except:
+            self.lang = ''
+        if self.lang == '' or self.lang is None:
+            self.lang = 'UNK'
+            yield False
+        else:
+            yield True
 
     def test(self, scheme, inventory=None):
         """ Test a file with various checks

diff --git a/tests/test_cts/__init__.py b/tests/test_cts/__init__.py
diff --git a/tests/test_cts/mocks/__init__.py b/tests/test_cts/mocks/__init__.py
diff --git a/tests/test_cts/mocks/xmllang_data.py b/tests/test_cts/mocks/xmllang_data.py
@@ -0,0 +1,80 @@
+from MyCapytain.common.utils import xmlparser
+from lxml.etree import tostring
+
+TEMPLATES = """<TEI xmlns="http://www.tei-c.org/ns/1.0">
+<teiHeader>
+    <encodingDesc>
+      <refsDecl n="CTS">
+        <cRefPattern n="chapter" matchPattern="(.+).(.+)" replacementPattern="#xpath(/tei:TEI/tei:text/tei:body/tei:div[@type='edition']/tei:div[@n='$1']/tei:div[@n='$2'])"/>
+        <cRefPattern n="book" matchPattern="(.+)" replacementPattern="#xpath(/tei:TEI/tei:text/tei:body/tei:div[@type='edition']/tei:div[@n='$1'])"/>
+      </refsDecl>
+    </encodingDesc>
+</teiHeader>
+<text><body><div></div></body></text>
+</TEI>"""
+URN = "urn:cts:latinLit:phi1294.phi002.perseus-lat2"
+LANG = "lat"
+
+# The following variable is a list of element with error that they should throw and status (XML_OBJECT, STATUS, ERROR)
+XMLLANG_DOCUMENTS = [
+    ("tei", "<text>", "<text xml:base='{urn}' xml:lang='{lang}'>", True, "TEI works with urn and xml lang on @base/text"),
+    ("tei", "<text>", "<text n='{urn}' xml:lang='{lang}'>", False, "TEI fails with urn and xml lang on @n/text"),
+    ("tei", "<text>", "<text xml:base='{urn}'>", False, "TEI fails with urn without xml lang on @base/text"),
+    ("tei", "<text>", "<text n='{urn}'>", False, "TEI fails with urn without xml lang on @n/text"),
+
+    ("epidoc", "<text>", "<text xml:base='{urn}' xml:lang='{lang}'>", False, "Epidoc fails with urn and xml lang on @base/text"),
+    ("epidoc", "<text>", "<text n='{urn}' xml:lang='{lang}'>", False, "Epidoc fails with urn and xml lang on @n/text"),
+    ("epidoc", "<text>", "<text xml:base='{urn}'>", False, "Epidoc fails with urn without xml lang on @base/text"),
+    ("epidoc", "<text>", "<text n='{urn}'>", False, "Epidoc fails with urn without xml lang on @n/text"),
+
+    ("tei", "<body>", "<body xml:base='{urn}' xml:lang='{lang}'>", False, "TEI fails with urn and xml lang on @base/body"),
+    ("tei", "<body>", "<body n='{urn}' xml:lang='{lang}'>", True, "TEI works with urn and xml lang on @n/body"),
+    ("tei", "<body>", "<body xml:base='{urn}'>", False, "TEI fails with urn without xml lang on @base/body"),
+    ("tei", "<body>", "<body n='{urn}'>", False, "TEI fails with urn without xml lang on @n/body"),
+
+    ("epidoc", "<body>", "<body xml:base='{urn}' xml:lang='{lang}'>", False, "Epidoc fails with urn and xml lang on @base/body"),
+    ("epidoc", "<body>", "<body n='{urn}' xml:lang='{lang}'>", False, "Epidoc fails with urn and xml lang on @n/body"),
+    ("epidoc", "<body>", "<body xml:base='{urn}'>", False, "Epidoc fails with urn without xml lang on @base/body"),
+    ("epidoc", "<body>", "<body n='{urn}'>", False, "Epidoc fails with urn without xml lang on @n/body"),
+
+]
+# Epidocs Tests
+XMLLANG_DOCUMENTS += [
+    (scheme, source, replacement.replace("{epidoc}", type_epidoc), boolean, message.replace("{epidoc}", type_epidoc))
+    for scheme, source, replacement, boolean, message in
+    [
+        ("tei", "<div>", "<div type='{epidoc}' n='{urn}' xml:lang='{lang}'>", False,
+         "TEI fails with urn and xml lang on @n/div-{epidoc}"),
+
+        ("tei", "<div>", "<div type='{epidoc}' xml:base='{urn}' xml:lang='{lang}'>", False,
+         "TEI fails with urn and xml lang on @xml:base/div-{epidoc}"),
+
+        ("tei", "<div>", "<div type='{epidoc}' xml:base='{urn}' xml:lang='{lang}'>", False,
+         "TEI fails with urn and without xml lang on @n/div-{epidoc}"),
+
+        ("tei", "<div>", "<div type='{epidoc}' n='{urn}' xml:lang='{lang}'>", False,
+         "TEI fails with urn and without xml lang on @xml:base/div-{epidoc}"),
+
+        ("epidoc", "<div>", "<div type='{epidoc}' n='{urn}' xml:lang='{lang}'>", True,
+         "Epidoc works with urn and xml lang on @n/div-{epidoc}"),
+
+        ("epidoc", "<div>", "<div type='{epidoc}' xml:base='{urn}' xml:lang='{lang}'>", False,
+         "Epidoc fails with urn and xml lang on @xml:base/div-{epidoc}"),
+
+        ("epidoc", "<div>", "<div type='{epidoc}' xml:base='{urn}'>", False,
+         "Epidoc fails with urn and without xml lang on @n/div-{epidoc}"),
+
+        ("epidoc", "<div>", "<div type='{epidoc}' n='{urn}'>", False,
+         "Epidoc fails with urn and without xml lang on @xml:base/div-{epidoc}")
+    ]
+    for type_epidoc in ["edition", "translation", "commentary"]
+]
+XMLLANG_DOCUMENTS = [
+    (
+        scheme,
+        tostring(xmlparser(TEMPLATES.replace(source, replacement).format(urn=URN, lang=LANG)), encoding=str),
+        boolean,
+        msg + " ("+replacement.format(urn=URN, lang=LANG)+")"
+    )
+    for scheme, source, replacement, boolean, msg in XMLLANG_DOCUMENTS
+]
diff --git a/tests/test_cts/test_unit.py b/tests/test_cts/test_unit.py
@@ -0,0 +1,41 @@
+import unittest
+import HookTest.capitains_units.cts
+import HookTest.units
+import tests.test_cts.mocks.xmllang_data as CAPITAINS_MOCKS
+from lxml import etree
+
+class TestCTS(unittest.TestCase):
+    """ Test the UnitTest for __cts__
+    """
+    pass
+
+
+class TestText(unittest.TestCase):
+    """ Test the UnitTests for Text
+    """
+
+    def setUp(self):
+        self.backup = [x for x in HookTest.capitains_units.cts.CTSText_TestUnit.tests]
+
+    def tearDown(self):
+        HookTest.capitains_units.cts.CTSText_TestUnit.tests = self.backup
+
+    def unit_from_string(self, xml_string, scheme):
+        """ Generates a test unit based on xml_string and scheme
+
+        :param xml_string: String representation of the XML Document
+        :param scheme: Scheme to apply
+        :return: HookTest.capitains_units.cts.CTSText_TestUnit
+        """
+        node = HookTest.capitains_units.cts.CTSText_TestUnit("/false/path")
+        node.scheme = scheme
+        node.xml = etree.fromstring(xml_string, HookTest.units.TESTUnit.PARSER)
+        return node
+
+    def test_xml_lang(self):
+        """ Ensure the XML:LANG test checks against the guideline
+        """
+        for scheme, xml, status, message in CAPITAINS_MOCKS.XMLLANG_DOCUMENTS:
+            unit = self.unit_from_string(xml, scheme)
+            results = [result for result in unit.language()]
+            self.assertEqual(results[0], status, message)