Add support for combining unicode_sets using multiple inheritance

3 files changed, 24 insertions, 10 deletions

diff --git a/CHANGES b/CHANGES
index b9ae519..585166b 100644
--- a/CHANGES
+++ b/CHANGES
@@ -6,6 +6,14 @@ Version 2.3.1 -
 ---------------

 - Added unicode sets to pyparsing_unicode for Latin-A and Latin-B ranges.

 

+- Added ability to define custom unicode sets as combinations of other sets

+  using multiple inheritance.

+

+    class Turkish_set(pp.pyparsing_unicode.Latin1, pp.pyparsing_unicode.LatinA):

+        pass

+

+    turkish_word = pp.Word(Turkish_set.alphas)

+

 

 Version 2.3.0 - October, 2018

 -----------------------------

diff --git a/pyparsing.py b/pyparsing.py
index 19b0141..d7be7b4 100644
--- a/pyparsing.py
+++ b/pyparsing.py
@@ -91,6 +91,7 @@ import pprint
 import traceback

 import types

 from datetime import datetime

+from itertools import takewhile

 try:

     # Python 3

     from itertools import filterfalse

@@ -5841,20 +5842,23 @@ class _lazyclassproperty(object):
         return ret

 

 

-class unicode_set:

+class unicode_set(object):

     _ranges = []

 

     @_lazyclassproperty

     def printables(cls):

-        return ''.join(filterfalse(unicode.isspace, (unichr(c) for r in cls._ranges for c in range(r[0], r[-1] + 1))))

+        ranges = set(sum((cc._ranges for cc in takewhile(lambda x: x is not unicode_set, cls.__mro__)), []))

+        return ''.join(filterfalse(unicode.isspace, (unichr(c) for r in ranges for c in range(r[0], r[-1] + 1))))

 

     @_lazyclassproperty

     def alphas(cls):

-        return ''.join(filter(unicode.isalpha, (unichr(c) for r in cls._ranges for c in range(r[0], r[-1] + 1))))

+        ranges = set(sum((cc._ranges for cc in takewhile(lambda x: x is not unicode_set, cls.__mro__)), []))

+        return ''.join(filter(unicode.isalpha, (unichr(c) for r in ranges for c in range(r[0], r[-1] + 1))))

 

     @_lazyclassproperty

     def nums(cls):

-        return ''.join(filter(unicode.isdigit, (unichr(c) for r in cls._ranges for c in range(r[0], r[-1] + 1))))

+        ranges = set(sum((cc._ranges for cc in takewhile(lambda x: x is not unicode_set, cls.__mro__)), []))

+        return ''.join(filter(unicode.isdigit, (unichr(c) for r in ranges for c in range(r[0], r[-1] + 1))))

 

     @_lazyclassproperty

     def alphanums(cls):

@@ -5901,9 +5905,8 @@ class pyparsing_unicode(unicode_set):
     class Korean(unicode_set):

         _ranges = [(0xac00, 0xd7af), (0x1100, 0x11ff), (0x3130, 0x318f), (0xa960, 0xa97f), (0xd7b0, 0xd7ff), ]

 

-    class CJK(unicode_set):

-        _ranges = [  # sum of Chinese, Japanese, and Korean ranges

-        ]

+    class CJK(Chinese, Japanese, Korean):

+        pass

 

     class Thai(unicode_set):

         _ranges = [(0x0e01, 0x0e3a), (0x0e3f, 0x0e5b), ]

@@ -5918,7 +5921,6 @@ class pyparsing_unicode(unicode_set):
         _ranges = [(0x0900, 0x097f), (0xa8e0, 0xa8ff)]

 

 pyparsing_unicode.Japanese._ranges = pyparsing_unicode.Japanese.Kanji._ranges + pyparsing_unicode.Japanese.Hiragana._ranges + pyparsing_unicode.Japanese.Katakana._ranges

-pyparsing_unicode.CJK._ranges = pyparsing_unicode.Chinese._ranges + pyparsing_unicode.Japanese._ranges + pyparsing_unicode.Korean._ranges

 

 # define ranges in language character sets

 if PY_3:

diff --git a/unitTests.py b/unitTests.py
index 5cb293d..2aff419 100644
--- a/unitTests.py
+++ b/unitTests.py
@@ -3650,8 +3650,9 @@ class UnicodeTests(ParseTestCase):
         self.assertTrue(result.asList() == [u'Καλημέρα', ',', u'κόσμε', '!'],

                         "Failed to parse Greek 'Hello, World!' using pyparsing_unicode.Greek.alphas")

 

-        class Turkish_set(pp.unicode_set):

-            _ranges = pp.pyparsing_unicode.Latin1._ranges + pp.pyparsing_unicode.LatinA._ranges

+        # define a custom unicode range using multiple inheritance

+        class Turkish_set(pp.pyparsing_unicode.Latin1, pp.pyparsing_unicode.LatinA):

+            pass

 

         key = pp.Word(Turkish_set.alphas)

         value = pp.pyparsing_common.integer | pp.Word(Turkish_set.alphas, Turkish_set.alphanums)

@@ -3667,6 +3668,9 @@ class UnicodeTests(ParseTestCase):
         print(result.asDict())

         self.assertEqual(result.asDict(), {'şehir': 'İzmir', 'ülke': 'Türkiye', 'nüfus': 4279677},

                          "Failed to parse Turkish key-value pairs")

+        self.assertEqual(len(pp.pyparsing_unicode.CJK.printables), 53760,

+                         "failed to construct ranges by merging Chinese, Japanese and Korean")

+        self.assertEqual(len(Turkish_set.printables), 317, "failed to construct ranges by merging Latin1 and LatinA")

 

 class IndentedBlockTest(ParseTestCase):

     # parse pseudo-yaml indented text