add str.casefold() (closes #13752)

Doc/library/stdtypes.rst

@@ -1002,6 +1002,14 @@ functions based on regular expressions.
    rest lowercased.
 
 
+.. method:: str.casefold()
+
+   Return a casefolded copy of the string. Casefolded strings may be used for
+   caseless matching. For example, ``"MASSE".casefold() == "maße".casefold()``.
+
+   .. versionadded:: 3.3
+
+
 .. method:: str.center(width[, fillchar])
 
    Return centered in a string of length *width*. Padding is done using the

Include/unicodeobject.h

@@ -2023,6 +2023,11 @@ PyAPI_FUNC(int) _PyUnicode_ToUpperFull(
     Py_UCS4 *res
     );
 
+PyAPI_FUNC(int) _PyUnicode_ToFoldedFull(
+    Py_UCS4 ch,       /* Unicode character */
+    Py_UCS4 *res
+    );
+
 PyAPI_FUNC(int) _PyUnicode_IsCaseIgnorable(
     Py_UCS4 ch         /* Unicode character */
     );

Lib/test/test_unicode.py

@@ -565,6 +565,14 @@ class UnicodeTest(string_tests.CommonTest,
         self.assertEqual('\U0008fffe'.lower(), '\U0008fffe')
         self.assertEqual('\u2177'.lower(), '\u2177')
 
+    def test_casefold(self):
+        self.assertEqual('hello'.casefold(), 'hello')
+        self.assertEqual('hELlo'.casefold(), 'hello')
+        self.assertEqual('ß'.casefold(), 'ss')
+        self.assertEqual('fi'.casefold(), 'fi')
+        self.assertEqual('\u03a3'.casefold(), '\u03c3')
+        self.assertEqual('A\u0345\u03a3'.casefold(), 'a\u03b9\u03c3')
+
     def test_upper(self):
         string_tests.CommonTest.test_upper(self)
         self.assertEqual('\U0001044F'.upper(), '\U00010427')

Misc/NEWS

@@ -10,6 +10,8 @@ What's New in Python 3.3 Alpha 1?
 Core and Builtins
 -----------------
 
+- Issue #13752: Add a casefold() method to str.
+
 - Issue #13761: Add a "flush" keyword argument to the print() function,
   used to ensure flushing the output stream.
 

Objects/unicodectype.c

@@ -185,7 +185,7 @@ Py_UCS4 _PyUnicode_ToUppercase(Py_UCS4 ch)
     const _PyUnicode_TypeRecord *ctype = gettyperecord(ch);
 
     if (ctype->flags & EXTENDED_CASE_MASK)
-        return _PyUnicode_ExtendedCase[ctype->upper & 0xFFFFFF];
+        return _PyUnicode_ExtendedCase[ctype->upper & 0xFFFF];
     return ctype->upper ? ctype->upper : ch;
 }
 
@@ -197,7 +197,7 @@ Py_UCS4 _PyUnicode_ToLowercase(Py_UCS4 ch)
     const _PyUnicode_TypeRecord *ctype = gettyperecord(ch);
 
     if (ctype->flags & EXTENDED_CASE_MASK)
-        return _PyUnicode_ExtendedCase[ctype->lower & 0xFFFFFF];
+        return _PyUnicode_ExtendedCase[ctype->lower & 0xFFFF];
     return ctype->lower ? ctype->lower : ch;
 }
 
@@ -206,7 +206,7 @@ int _PyUnicode_ToLowerFull(Py_UCS4 ch, Py_UCS4 *res)
     const _PyUnicode_TypeRecord *ctype = gettyperecord(ch);
 
     if (ctype->flags & EXTENDED_CASE_MASK) {
-        int index = ctype->lower & 0xFFFFFF;
+        int index = ctype->lower & 0xFFFF;
         int n = ctype->lower >> 24;
         int i;
         for (i = 0; i < n; i++)
@@ -222,7 +222,7 @@ int _PyUnicode_ToTitleFull(Py_UCS4 ch, Py_UCS4 *res)
     const _PyUnicode_TypeRecord *ctype = gettyperecord(ch);
 
     if (ctype->flags & EXTENDED_CASE_MASK) {
-        int index = ctype->title & 0xFFFFFF;
+        int index = ctype->title & 0xFFFF;
         int n = ctype->title >> 24;
         int i;
         for (i = 0; i < n; i++)
@@ -238,7 +238,7 @@ int _PyUnicode_ToUpperFull(Py_UCS4 ch, Py_UCS4 *res)
     const _PyUnicode_TypeRecord *ctype = gettyperecord(ch);
 
     if (ctype->flags & EXTENDED_CASE_MASK) {
-        int index = ctype->upper & 0xFFFFFF;
+        int index = ctype->upper & 0xFFFF;
         int n = ctype->upper >> 24;
         int i;
         for (i = 0; i < n; i++)
@@ -249,6 +249,21 @@ int _PyUnicode_ToUpperFull(Py_UCS4 ch, Py_UCS4 *res)
     return 1;
 }
 
+int _PyUnicode_ToFoldedFull(Py_UCS4 ch, Py_UCS4 *res)
+{
+    const _PyUnicode_TypeRecord *ctype = gettyperecord(ch);
+
+    if (ctype->flags & EXTENDED_CASE_MASK && (ctype->lower >> 20) & 7) {
+        int index = (ctype->lower & 0xFFFF) + (ctype->lower >> 24);
+        int n = (ctype->lower >> 20) & 7;
+        int i;
+        for (i = 0; i < n; i++)
+            res[i] = _PyUnicode_ExtendedCase[index + i];
+        return n;
+    }
+    return _PyUnicode_ToLowerFull(ch, res);
+}
+
 int _PyUnicode_IsCased(Py_UCS4 ch)
 {
     const _PyUnicode_TypeRecord *ctype = gettyperecord(ch);

Objects/unicodeobject.c

@@ -9576,6 +9576,24 @@ do_lower(int kind, void *data, Py_ssize_t length, Py_UCS4 *res, Py_UCS4 *maxchar
     return do_upper_or_lower(kind, data, length, res, maxchar, 1);
 }
 
+static Py_ssize_t
+do_casefold(int kind, void *data, Py_ssize_t length, Py_UCS4 *res, Py_UCS4 *maxchar)
+{
+    Py_ssize_t i, k = 0;
+
+    for (i = 0; i < length; i++) {
+        Py_UCS4 c = PyUnicode_READ(kind, data, i);
+        Py_UCS4 mapped[3];
+        int j, n_res = _PyUnicode_ToFoldedFull(c, mapped);
+        for (j = 0; j < n_res; j++) {
+            if (mapped[j] > *maxchar)
+                *maxchar = mapped[j];
+            res[k++] = mapped[j];
+        }
+    }
+    return k;
+}
+
 static Py_ssize_t
 do_title(int kind, void *data, Py_ssize_t length, Py_UCS4 *res, Py_UCS4 *maxchar)
 {
@@ -10501,6 +10519,22 @@ unicode_capitalize(PyObject *self)
     return case_operation(self, do_capitalize);
 }
 
+PyDoc_STRVAR(casefold__doc__,
+             "S.casefold() -> str\n\
+\n\
+Return a version of S suitable for caseless comparisons.");
+
+static PyObject *
+unicode_casefold(PyObject *self)
+{
+    if (PyUnicode_READY(self) == -1)
+        return NULL;
+    if (PyUnicode_IS_ASCII(self))
+        return ascii_upper_or_lower(self, 1);
+    return case_operation(self, do_casefold);
+}
+
+
 /* Argument converter.  Coerces to a single unicode character */
 
 static int
@@ -12998,6 +13032,7 @@ static PyMethodDef unicode_methods[] = {
     {"rsplit", (PyCFunction) unicode_rsplit, METH_VARARGS, rsplit__doc__},
     {"join", (PyCFunction) unicode_join, METH_O, join__doc__},
     {"capitalize", (PyCFunction) unicode_capitalize, METH_NOARGS, capitalize__doc__},
+    {"casefold", (PyCFunction) unicode_casefold, METH_NOARGS, casefold__doc__},
     {"title", (PyCFunction) unicode_title, METH_NOARGS, title__doc__},
     {"center", (PyCFunction) unicode_center, METH_VARARGS, center__doc__},
     {"count", (PyCFunction) unicode_count, METH_VARARGS, count__doc__},

Tools/unicode/makeunicodedata.py

@@ -49,6 +49,7 @@ LINE_BREAK = "LineBreak%s.txt"
 NAME_ALIASES = "NameAliases%s.txt"
 NAMED_SEQUENCES = "NamedSequences%s.txt"
 SPECIAL_CASING = "SpecialCasing%s.txt"
+CASE_FOLDING = "CaseFolding%s.txt"
 
 # Private Use Areas -- in planes 1, 15, 16
 PUA_1 = range(0xE000, 0xF900)
@@ -424,28 +425,36 @@ def makeunicodetype(unicode, trace):
             if "Case_Ignorable" in properties:
                 flags |= CASE_IGNORABLE_MASK
             sc = unicode.special_casing.get(char)
+            cf = unicode.case_folding.get(char, [char])
+            if record[12]:
+                upper = int(record[12], 16)
+            else:
+                upper = char
+            if record[13]:
+                lower = int(record[13], 16)
+            else:
+                lower = char
+            if record[14]:
+                title = int(record[14], 16)
+            else:
+                title = upper
+            if sc is None and cf != [lower]:
+                sc = ([lower], [title], [upper])
             if sc is None:
-                if record[12]:
-                    upper = int(record[12], 16)
-                else:
-                    upper = char
-                if record[13]:
-                    lower = int(record[13], 16)
-                else:
-                    lower = char
-                if record[14]:
-                    title = int(record[14], 16)
-                else:
-                    title = upper
                 if upper == lower == title:
                     upper = lower = title = 0
             else:
-                # This happens when some character maps to more than one
-                # character in uppercase, lowercase, or titlecase. The extra
-                # characters are stored in a different array.
+                # This happens either when some character maps to more than one
+                # character in uppercase, lowercase, or titlecase or the
+                # casefolded version of the character is different from the
+                # lowercase. The extra characters are stored in a different
+                # array.
                 flags |= EXTENDED_CASE_MASK
                 lower = len(extra_casing) | (len(sc[0]) << 24)
                 extra_casing.extend(sc[0])
+                if cf != sc[0]:
+                    lower |= len(cf) << 20
+                    extra_casing.extend(cf)
                 upper = len(extra_casing) | (len(sc[2]) << 24)
                 extra_casing.extend(sc[2])
                 # Title is probably equal to upper.
@@ -1107,6 +1116,17 @@ class UnicodeData:
                 title = [int(char, 16) for char in data[2].split()]
                 upper = [int(char, 16) for char in data[3].split()]
                 sc[c] = (lower, title, upper)
+        cf = self.case_folding = {}
+        if version != '3.2.0':
+            with open_data(CASE_FOLDING, version) as file:
+                for s in file:
+                    s = s[:-1].split('#', 1)[0]
+                    if not s:
+                        continue
+                    data = s.split("; ")
+                    if data[1] in "CF":
+                        c = int(data[0], 16)
+                        cf[c] = [int(char, 16) for char in data[2].split()]
 
     def uselatin1(self):
         # restrict character range to ISO Latin 1