Merge pull request #6 from belerico/latex2asciimath

Latex2asciimath fix minor issues
belerico · Apr 22, 2020 · 195b2b5 · 195b2b5
2 parents b4d96f5 + e49569e
commit 195b2b5
Show file tree

Hide file tree

Showing 6 changed files with 28 additions and 23 deletions.
diff --git a/py_asciimath/asciimath.py b/py_asciimath/asciimath.py
@@ -387,8 +387,8 @@ def get_symbols_for(symbol_group, lang_to):  # pragma: no cover
     "/": {"latex": "/", "mathml": "/"},
     "|": {"latex": ["|", "\\vert", "\\mid"], "mathml": "|"},
     ":": {"latex": ":", "mathml": ":"},
-    "int": {"latex": "\\int", "mathml": "&Integral;"},
-    "integral": {"latex": "\\int", "mathml": "&Integral;"},
+    "int": {"latex": "\\int", "mathml": "&int;"},
+    "integral": {"latex": "\\int", "mathml": "&int;"},
     "oint": {"latex": "\\oint", "mathml": "&conint;"},
     "del": {"latex": "\\partial", "mathml": "&part;"},
     "partial": {"latex": "\\partial", "mathml": "&part;"},

diff --git a/py_asciimath/latex.py b/py_asciimath/latex.py
@@ -308,7 +308,6 @@ def get_symbols_for(symbol_group, lang_to):  # pragma: no cover
     ".": {"asciimath": ".", "mathml": "."},
     "_": {"asciimath": "_", "mathml": "_"},
     "'": {"asciimath": "'", "mathml": "'"},
-    "/": {"asciimath": "/", "mathml": "/"},
     "|": {"asciimath": "|", "mathml": "|"},
     "\\vert": {"asciimath": "|", "mathml": "|"},
     "\\mid": {"asciimath": "|", "mathml": "|"},

diff --git a/py_asciimath/transformer/transformer.py b/py_asciimath/transformer/transformer.py
@@ -312,6 +312,10 @@ def exp_par(self, items):
     def exp_unary(self, items):
         unary = mathml_una[items[0]]
         items[1] = self.remove_parenthesis(items[1])
+        if items[0] == "text":
+            return encapsulate_mrow(
+                unary.format(re.sub(r"<.*?>", "", items[1]))
+            )
         return encapsulate_mrow(unary.format(encapsulate_mrow(items[1])))
 
     @MathTransformer.log
@@ -401,8 +405,8 @@ def exp_super(self, items):
     def exp_under_super(self, items):
         items[0] = self.remove_parenthesis(items[0])
         items[1] = self.remove_parenthesis(items[1])
-        items[1] = self.remove_parenthesis(items[2])
-        return "(" + items[0] + ")^(" + items[1] + ")_(" + items[2] + ")"
+        items[2] = self.remove_parenthesis(items[2])
+        return "(" + items[0] + ")_(" + items[1] + ")^(" + items[2] + ")"
 
     @log
     def exp_par(self, items):
@@ -424,13 +428,10 @@ def exp_par(self, items):
 
     @log
     def exp_unary(self, items):
-        items[1] = self.remove_parenthesis(items[1])
         return l2mml_una[items[0]] + "(" + items[1] + ")"
 
     @log
     def exp_binary(self, items):
-        items[1] = self.remove_parenthesis(items[1])
-        items[2] = self.remove_parenthesis(items[2])
         if items[0].startswith("\\sqrt"):
             return "root(" + "".join(items[1:-1]) + ")(" + items[-1] + ")"
         return l2mml_bin[items[0]] + "(" + items[1] + ")(" + items[2] + ")"

diff --git a/py_asciimath/translator/translator.py b/py_asciimath/translator/translator.py
@@ -272,7 +272,7 @@ def _translate(
             parsed = MathMLParser.parse(
                 parsed,
                 dtd=dtd,
-                dtd_validation=dtd_validation,
+                dtd_validation=True,
                 network=network,
                 **kwargs
             )
@@ -313,7 +313,12 @@ def translate(
                 against. It can be: `mathml1`, `mathml2` or `mathml3`.
                 Defaults to None.
             dtd_validation (bool, optional): If True validate output against
-                the DTD version specified by `dtd`.
+                the DTD version specified by `dtd`. By default, if one of
+                `dtd`, `dtd_validation`, `xml_declaration` or `xml_pprint` is
+                True or is not None, then `dtd_validation` will be set to True.
+                This is because if either one of them set to True, then
+                py_asciimath must parse the input XML, but in order to do that
+                it needs to know how to intepret the entities from the DTD.
                 Defaults to False.
             from_file (bool, optional): If True, load the string to translate
                 from the file specified by s. Defaults to False.

diff --git a/tests/test_ASCIIMath2MathML.py b/tests/test_ASCIIMath2MathML.py
@@ -80,7 +80,7 @@ def test_asciimath2mathml_ok_4(self):
             s,
             '<!DOCTYPE math SYSTEM "'
             + PROJECT_ROOT
-            + '/dtd/mathml1/mathml1.dtd">\n<math><mstyle displaystyle="true"><mrow><mo>&langle;</mo><mtable><mtr><mtd><mn>1</mn></mtd><mtd><mn>2</mn></mtd></mtr><mtr><mtd><mn>2</mn></mtd><mtd><mo>&Integral;</mo><mrow><mo>[</mo><mrow><mn>3</mn><mrow><mo>(</mo><mrow><mi>x</mi><mo>+</mo><mn>1</mn></mrow><mo>)</mo></mrow></mrow><mo>]</mo></mrow><mi>dx</mi></mtd></mtr></mtable><mo/></mrow></mstyle></math>',
+            + '/dtd/mathml1/mathml1.dtd">\n<math><mstyle displaystyle="true"><mrow><mo>&langle;</mo><mtable><mtr><mtd><mn>1</mn></mtd><mtd><mn>2</mn></mtd></mtr><mtr><mtd><mn>2</mn></mtd><mtd><mo>&int;</mo><mrow><mo>[</mo><mrow><mn>3</mn><mrow><mo>(</mo><mrow><mi>x</mi><mo>+</mo><mn>1</mn></mrow><mo>)</mo></mrow></mrow><mo>]</mo></mrow><mi>dx</mi></mtd></mtr></mtable><mo/></mrow></mstyle></math>',
         )
 
     def test_asciimath2mathml_ok_5(self):
@@ -113,7 +113,7 @@ def test_asciimath2mathml_xml_fields_ok_1(self):
         )
         self.assertEqual(
             s,
-            "<?xml version='1.0' encoding='UTF-8'?>\n<math><mn>1</mn></math>",
+            "<?xml version='1.0' encoding='UTF-8'?>\n<!DOCTYPE math SYSTEM \"http://www.w3.org/Math/DTD/mathml1/mathml.dtd\">\n<math><mn>1</mn></math>",
         )
 
     def test_asciimath2mathml_xml_fields_ok_2(self):

diff --git a/tests/test_Tex2ASCIIMath.py b/tests/test_Tex2ASCIIMath.py
@@ -17,17 +17,17 @@ def test_tex2asciimath_ok_1(self):
         )
         self.assertEqual(s, "[{:[int x text(d x)],[log (x + 1)]:}]")
 
-    # def test_tex2asciimath_ok_2(self):
-    #     s = Tex2ASCIIMath(
-    #         inplace=True, parser="lalr", lexer="contextual",
-    #     ).translate(
-    #         "((1,2))int sin{x^2}/4pidxroot(5)(x_1^2+x_2^2)", displaystyle=True
-    #     )
-    #     self.assertEqual(
-    #         s,
-    #         r"\[\left(\left(1 , 2\right)\right) \int \sin "
-    #         r"\frac{{x}^{2}}{4} \pi \mathrm{dx} \sqrt[5]{{x}_{1}^{2} + {x}_{2}^{2}}\]",
-    #     )
+    def test_tex2asciimath_ok_2(self):
+        s = Tex2ASCIIMath(
+            inplace=True, parser="lalr", lexer="contextual",
+        ).translate(
+            r"\[\left(\left(1 , 2\right)\right) \int \sin "
+            r"\frac{{x}^{2}}{4} \pi \mathrm{dx} \sqrt[5]{{x}_{1}^{2} + {x}_{2}^{2}}\]",
+        )
+        self.assertEqual(
+            s,
+            "((1 , 2)) int sin frac((x)^(2))(4) pi text(d x) root(5)((x)_(1)^(2) + (x)_(2)^(2))",
+        )
 
     # def test_tex2asciimath_ok_3(self):
     #     s = Tex2ASCIIMath(