Merge pull request #171 from Capitains/empty_reference_master

EmptyReference exception along with unit tests

CHANGES.md

@@ -1,3 +1,7 @@
+### 2018-06-22 2.0.7 @sonofmun
+
+- Added exception for empty references
+
 ### 2017-11-16 2.0.6 @ponteineptique
 
 - Added a way to specify the joining string for Plaintext export in TEIResource (Fixed #146)

MyCapytain/__init__.py

@@ -9,4 +9,4 @@
 
 """
 
-__version__ = "2.0.6"
+__version__ = "2.0.7"

MyCapytain/errors.py

@@ -60,3 +60,7 @@ class UndispatchedTextError(Exception, MyCapytainException):
 class UnknownCollection(KeyError, MyCapytainException):
     """ A collection is unknown to its ancestor
     """
+
+class EmptyReference(SyntaxWarning, MyCapytainException):
+    """ Error generated when a duplicate is found in Reference
+    """

MyCapytain/resources/texts/local/capitains/cts.py

@@ -11,7 +11,7 @@
 
 import warnings
 
-from MyCapytain.errors import DuplicateReference, MissingAttribute, RefsDeclError
+from MyCapytain.errors import DuplicateReference, MissingAttribute, RefsDeclError, EmptyReference
 from MyCapytain.common.utils import copyNode, passageLoop, normalizeXpath
 from MyCapytain.common.constants import XPATH_NAMESPACES, RDF_NAMESPACES
 from MyCapytain.common.reference import URN, Citation, Reference
@@ -248,6 +248,10 @@ def getValidReff(self, level=None, reference=None, _debug=False):
                 message = ", ".join(duplicates)
                 warnings.warn(message, DuplicateReference)
             del duplicates
+            empties = [n for n in passages if n.strip('.') != n or n == '']
+            if len(empties) > 0:
+                message = '{} empty reference(s) at citation level {}'.format(len(empties), level)
+                warnings.warn(message, EmptyReference)
 
         return passages
 
@@ -664,4 +668,4 @@ def textObject(self):
 
         :rtype: CapitainsCtsText
         """
-        return self.__text__
+        return self.__text__

tests/resources/texts/local/commonTests.py

@@ -211,6 +211,22 @@ def test_warning(self):
                          "Warning should be DuplicateReference")
         self.assertEqual(str(w[0].message), "1", "Warning message should be list of duplicate")
 
+    def test_empty_ref_warning(self):
+        with open("tests/testing_data/texts/empty_references.xml") as xml:
+            text = CapitainsCtsText(resource=xml)
+        with warnings.catch_warnings(record=True) as w:
+            # Cause all warnings to always be triggered.
+            warnings.simplefilter("always")
+            for i in [1, 2, 3]:
+                text.getValidReff(level=i, _debug=True)
+
+        self.assertEqual(len(w), 3, "There should be warning on each level")
+        self.assertEqual(issubclass(w[-1].category, MyCapytain.errors.EmptyReference), True,
+                         "Warning should be EmptyReference")
+        self.assertEqual(str(w[0].message), "1 empty reference(s) at citation level 1",
+                         "Warning message should indicate number of references and the level at which they occur")
+
+
     def test_wrong_main_scope(self):
         with open("tests/testing_data/texts/sample2.xml", "rb") as f:
             with self.assertRaises(MyCapytain.errors.RefsDeclError):
@@ -836,4 +852,4 @@ def test_first_list(self):
         self.assertEqual(
             str(passage.lastId), "2.40.8",
             "Last reff should be the last"
-        )
+        )

tests/testing_data/texts/empty_references.xml

@@ -0,0 +1,88 @@
+<?xml-model href="http://www.stoa.org/epidoc/schema/8.19/tei-epidoc.rng"
+  schematypens="http://relaxng.org/ns/structure/1.0"?><?xml-model href="http://www.stoa.org/epidoc/schema/8.19/tei-epidoc.rng"
+  schematypens="http://purl.oclc.org/dsdl/schematron"?>
+<TEI xmlns="http://www.tei-c.org/ns/1.0">
+    <teiHeader type="text">
+        <fileDesc>
+            <titleStmt>
+                <title>Epigrammata</title>
+                <author>Martial</author>
+                <sponsor>Perseus Project, Tufts University</sponsor>
+                <principal>Gregory Crane</principal>
+            </titleStmt>
+            <publicationStmt>
+                <publisher>Trustees of Tufts University</publisher>
+                <pubPlace>Medford, MA</pubPlace>
+                <authority>Perseus Project</authority>
+            </publicationStmt>
+            <sourceDesc default="false">
+                <bibl default="false">
+                    <title>M. Valerii Martialis Epigrammaton libri / recognovit W. Heraeus</title>
+                    <author>Martial</author>
+                    <editor role="editor">Wilhelm Heraeus<!--, 1862-1938 (main editor)--></editor>
+                    <editor role="editor">Jacobus Borovskij<!-- (added corrections in 1976, more than 30 years before 2011)--></editor>
+                    <pubPlace>Leipzig</pubPlace>
+                    <date>1925/1976</date>
+                </bibl>
+            </sourceDesc>
+        </fileDesc>
+        <encodingDesc>
+            <refsDecl n="CTS">
+                <cRefPattern n="line"
+                    matchPattern="(\w+).(\w+).(\w+)"
+                    replacementPattern="#xpath(/tei:TEI/tei:text/tei:body/tei:div/tei:div[@n='$1']/tei:div[@n='$2']/tei:l[@n='$3'])">
+                    <p>This pointer pattern extracts book and poem and line</p>
+                </cRefPattern>
+                <cRefPattern n="poem"
+                    matchPattern="(\w+).(\w+)"
+                    replacementPattern="#xpath(/tei:TEI/tei:text/tei:body/tei:div/tei:div[@n='$1']/tei:div[@n='$2'])">
+                    <p>This pointer pattern extracts book and poem</p>
+                </cRefPattern>
+                <cRefPattern n="book"
+                    matchPattern="(\w+)"
+                    replacementPattern="#xpath(/tei:TEI/tei:text/tei:body/tei:div/tei:div[@n='$1'])">
+                    <p>This pointer pattern extracts book</p>
+                </cRefPattern>
+            </refsDecl>
+            <refsDecl xml:id="TEI.2">
+                <refState unit="book"/>
+                <refState unit="poem"/>
+                <refState unit="line"/>
+            </refsDecl>
+        </encodingDesc>
+        <profileDesc>
+            <langUsage default="false">
+                <language ident="eng">English</language>
+                <language ident="grc">Latin</language>
+            </langUsage>
+        </profileDesc>
+        <revisionDesc>
+            <change when="2008-01-01" who="Gregory Crane">Initial Markup</change>
+            <change when="2015-03-06" who="Thibault Clérice">converted to EpiDoc/CTS</change>
+        </revisionDesc>
+    </teiHeader>
+    <text n="urn:cts:latinLit:phi1294.phi002.perseus-lat2"
+
+        xml:id="stoa0045.stoa0">
+        <body>
+            <div type="edition" n="urn:cts:latinLit:phi1294.phi002.perseus-lat2" xml:lang="lat">
+                <div type="textpart" subtype="book" n="1">
+                    <div type="textpart" subtype="poem" n="pr">
+                        <l n="">Spero me secutum in libellis meis tale temperamen-</l>
+                    </div>
+                </div>
+                <div type="textpart" subtype="book" n="2">
+                    <div type="textpart" subtype="poem" n="">
+                        <l n="1">Spero me secutum in libellis meis tale temperamen-</l>
+                    </div>
+                </div>
+                <div type="textpart" subtype="book" n="">
+                    <div type="textpart" subtype="poem" n="1">
+                        <l n="1">Spero me secutum in libellis meis tale temperamen-</l>
+                    </div>
+                </div>
+            </div>
+        </body>
+    </text>
+</TEI>
+