mittens 0.2.0 → 0.3.0

data/vendor/snowball/pascal/SnowballProgram.pas

@@ -28,9 +28,13 @@ Type
 
     Protected
         Function InGrouping(s : array of char; min, max : Integer) : Boolean;
+        Function GoInGrouping(s : array of char; min, max : Integer) : Boolean;
         Function InGroupingBk(s : array of char; min, max : Integer) : Boolean;
+        Function GoInGroupingBk(s : array of char; min, max : Integer) : Boolean;
         Function OutGrouping(s : array of char; min, max : Integer) : Boolean;
+        Function GoOutGrouping(s : array of char; min, max : Integer) : Boolean;
         Function OutGroupingBk(s : array of char; min, max : Integer) : Boolean;
+        Function GoOutGroupingBk(s : array of char; min, max : Integer) : Boolean;
 
         Function EqS(s_size : Integer; s : AnsiString) : Boolean;
         Function EqSBk(s_size : Integer; s : AnsiString) : Boolean;
@@ -45,7 +49,7 @@ Type
         Procedure SliceCheck;
         Procedure SliceFrom(s : AnsiString);
 
-        Function  ReplaceS(bra, ket : Integer; s : AnsiString) : Integer;        
+        Function  ReplaceS(bra, ket : Integer; s : AnsiString) : Integer;
         Procedure Insert(bra, ket : Integer; s : AnsiString);
 
         Function SliceTo : AnsiString;
@@ -89,6 +93,23 @@ Begin
     Result := True;
 End;
 
+Function TSnowballProgram.GoInGrouping(s : array of char; min, max : Integer) : Boolean;
+Var ch : Integer;
+Begin
+    Result := True;
+    While (FCursor < FLimit) Do
+    Begin
+        ch := Ord(FCurrent[FCursor + 1]);
+        If (ch > max) Or (ch < min) Then Exit;
+
+        ch := ch - min;
+        If (Ord(s[ch Shr 3]) And Ord(1 Shl (ch And $7))) = 0 Then Exit;
+
+        Inc(FCursor);
+    End;
+    Result := False;
+End;
+
 Function TSnowballProgram.InGroupingBk(s : array of char; min, max : Integer) : Boolean;
 Var ch : Integer;
 Begin
@@ -105,6 +126,23 @@ Begin
     Result := True;
 End;
 
+Function TSnowballProgram.GoInGroupingBk(s : array of char; min, max : Integer) : Boolean;
+Var ch : Integer;
+Begin
+    Result := True;
+    While (FCursor > FBkLimit) Do
+    Begin
+        ch := Ord(FCurrent[FCursor]);
+        If (ch > max) Or (ch < min) Then Exit;
+
+        ch := ch - min;
+        If (Ord(s[ch Shr 3]) And Ord(1 Shl (ch And $7))) = 0 Then Exit;
+
+        Dec(FCursor);
+    End;
+    Result := False;
+End;
+
 Function TSnowballProgram.OutGrouping(s : array of char; min, max : Integer) : Boolean;
 Var ch : Integer;
 Begin
@@ -128,6 +166,29 @@ Begin
     End;
 End;
 
+Function TSnowballProgram.GoOutGrouping(s : array of char; min, max : Integer) : Boolean;
+Var ch : Integer;
+Begin
+    Result := True;
+
+    While (FCursor < FLimit) Do
+    Begin
+        ch := Ord(FCurrent[FCursor + 1]);
+
+        If (ch <= max) And (ch >= min) Then
+        Begin
+            ch := ch - min;
+            If (Ord(s[ch Shr 3]) And Ord(1 Shl (ch And $7))) <> 0 Then
+            Begin
+                Exit;
+            End;
+        End;
+
+        Inc(FCursor);
+    End;
+    Result := False;
+End;
+
 Function TSnowballProgram.OutGroupingBk(s : array of char; min, max : Integer) : Boolean;
 Var ch : Integer;
 Begin
@@ -151,6 +212,27 @@ Begin
     End;
 End;
 
+Function TSnowballProgram.GoOutGroupingBk(s : array of char; min, max : Integer) : Boolean;
+Var ch : Integer;
+Begin
+    Result := True;
+
+    While (FCursor > FBkLimit) Do
+    Begin
+        ch := Ord(FCurrent[FCursor]);
+        If (ch <= max) And (ch >= min) Then
+        Begin
+            ch := ch - min;
+            If (Ord(s[ch Shr 3]) And Ord(1 Shl (ch And $7))) <> 0 Then
+            Begin
+                Exit;
+            End;
+        End;
+        Dec(FCursor);
+    End;
+    Result := False;
+End;
+
 Function TSnowballProgram.EqS(s_size : Integer; s : AnsiString) : Boolean;
 Var I : Integer;
 Begin
@@ -262,7 +344,7 @@ Begin
             If Not Assigned(w.Method) Then
             Begin
                 Result := w.Result;
-                Exit;   
+                Exit;
             End;
 
             res := w.Method;

data/vendor/snowball/pascal/generate.pl

@@ -7,17 +7,17 @@ use warnings;
 my @sources = @ARGV;
 
 while (defined(my $line = <STDIN>)) {
-	if ($line =~ /\{\s*BEGIN TEMPLATE\s*\}/) {
-		my $template = '';
-		while (defined($line = <STDIN>) && $line !~ /\{\s*END TEMPLATE\s*\}/) {
-			$template .= $line;
-		}
-		foreach my $source(@sources) {
-			my $out = $template;
-			$out =~ s/%STEMMER%/$source/g;
-			print $out;
-		}
-		next;
-	}
-	print $line;
+    if ($line =~ /\{\s*BEGIN TEMPLATE\s*\}/) {
+        my $template = '';
+        while (defined($line = <STDIN>) && $line !~ /\{\s*END TEMPLATE\s*\}/) {
+            $template .= $line;
+        }
+        foreach my $source(@sources) {
+            my $out = $template;
+            $out =~ s/%STEMMER%/$source/g;
+            print $out;
+        }
+        next;
+    }
+    print $line;
 }

data/vendor/snowball/python/create_init.py

@@ -15,12 +15,15 @@ for pyscript in os.listdir(python_out_folder):
     match = filematch.match(pyscript)
     if (match):
         langname = match.group(1)
-        titlecase = langname.title()
+        titlecase = re.sub(r"_", "", langname.title())
         languages.append("    '%(lang)s': %(title)sStemmer," % {'lang': langname, 'title': titlecase})
         imports.append('from .%(lang)s_stemmer import %(title)sStemmer' % {'lang': langname, 'title': titlecase})
 imports.sort()
 languages.sort()
 
+if len(languages) == 0:
+    raise AssertionError('languages list is empty!')
+
 src = '''__all__ = ('language', 'stemmer')
 
 %(imports)s

data/vendor/snowball/python/setup.cfg

@@ -1,6 +1,3 @@
 [metadata]
 long_description = file: README.rst
 long_description_content_type = text/x-rst
-
-[bdist_wheel]
-universal=1

data/vendor/snowball/python/setup.py

@@ -3,7 +3,7 @@
 from setuptools import setup
 import re
 
-SNOWBALL_VERSION = '2.2.0'
+SNOWBALL_VERSION = '3.0.1'
 
 n_stemmers = 0
 
@@ -42,8 +42,8 @@ classifiers = [
 for lang in langs:
     lang_titlecase = lang.title()
     # Only classifiers listed in https://pypi.org/classifiers/ are allowed
-    if lang_titlecase not in ('Armenian', 'Yiddish'):
-        classifiers.append('Natural Language :: ' + lang_titlecase)
+    # Remove them here or submit them to https://github.com/pypa/trove-classifiers
+    classifiers.append('Natural Language :: ' + lang_titlecase)
 
 classifiers.extend([
     'Operating System :: OS Independent',
@@ -52,6 +52,7 @@ classifiers.extend([
     'Programming Language :: Python :: 2.6',
     'Programming Language :: Python :: 2.7',
     'Programming Language :: Python :: 3',
+    'Programming Language :: Python :: 3.3',
     'Programming Language :: Python :: 3.4',
     'Programming Language :: Python :: 3.5',
     'Programming Language :: Python :: 3.6',
@@ -59,6 +60,9 @@ classifiers.extend([
     'Programming Language :: Python :: 3.8',
     'Programming Language :: Python :: 3.9',
     'Programming Language :: Python :: 3.10',
+    'Programming Language :: Python :: 3.11',
+    'Programming Language :: Python :: 3.12',
+    'Programming Language :: Python :: 3.13',
     'Programming Language :: Python :: Implementation :: CPython',
     'Programming Language :: Python :: Implementation :: PyPy',
     'Topic :: Database',
@@ -77,5 +81,6 @@ setup(name='snowballstemmer',
       license="BSD-3-Clause",
       packages=['snowballstemmer'],
       package_dir={"snowballstemmer": "src/snowballstemmer"},
+      python_requires='!=3.0.*, !=3.1.*, !=3.2.*',
       classifiers = classifiers
 )

data/vendor/snowball/python/snowballstemmer/basestemmer.py

@@ -27,95 +27,63 @@ class BaseStemmer(object):
         self.bra              = other.bra
         self.ket              = other.ket
 
-    def in_grouping(self, s, min, max):
+    def in_grouping(self, s):
         if self.cursor >= self.limit:
             return False
-        ch = ord(self.current[self.cursor])
-        if ch > max or ch < min:
-            return False
-        ch -= min
-        if (s[ch >> 3] & (0x1 << (ch & 0x7))) == 0:
+        if self.current[self.cursor] not in s:
             return False
         self.cursor += 1
         return True
 
-    def go_in_grouping(self, s, min, max):
+    def go_in_grouping(self, s):
         while self.cursor < self.limit:
-            ch = ord(self.current[self.cursor])
-            if ch > max or ch < min:
-                return True
-            ch -= min
-            if (s[ch >> 3] & (0x1 << (ch & 0x7))) == 0:
+            if self.current[self.cursor] not in s:
                 return True
             self.cursor += 1
         return False
 
-    def in_grouping_b(self, s, min, max):
+    def in_grouping_b(self, s):
         if self.cursor <= self.limit_backward:
             return False
-        ch = ord(self.current[self.cursor - 1])
-        if ch > max or ch < min:
-            return False
-        ch -= min
-        if (s[ch >> 3] & (0x1 << (ch & 0x7))) == 0:
+        if self.current[self.cursor - 1] not in s:
             return False
         self.cursor -= 1
         return True
 
-    def go_in_grouping_b(self, s, min, max):
+    def go_in_grouping_b(self, s):
         while self.cursor > self.limit_backward:
-            ch = ord(self.current[self.cursor - 1])
-            if ch > max or ch < min:
-                return True
-            ch -= min
-            if (s[ch >> 3] & (0x1 << (ch & 0x7))) == 0:
+            if self.current[self.cursor - 1] not in s:
                 return True
             self.cursor -= 1
         return False
 
-    def out_grouping(self, s, min, max):
+    def out_grouping(self, s):
         if self.cursor >= self.limit:
             return False
-        ch = ord(self.current[self.cursor])
-        if ch > max or ch < min:
-            self.cursor += 1
-            return True
-        ch -= min
-        if (s[ch >> 3] & (0X1 << (ch & 0x7))) == 0:
+        if self.current[self.cursor] not in s:
             self.cursor += 1
             return True
         return False
 
-    def go_out_grouping(self, s, min, max):
+    def go_out_grouping(self, s):
         while self.cursor < self.limit:
-            ch = ord(self.current[self.cursor])
-            if ch <= max and ch >= min:
-                ch -= min
-                if (s[ch >> 3] & (0X1 << (ch & 0x7))):
-                    return True
+            if self.current[self.cursor] in s:
+                return True
             self.cursor += 1
         return False
 
-    def out_grouping_b(self, s, min, max):
+    def out_grouping_b(self, s):
         if self.cursor <= self.limit_backward:
             return False
-        ch = ord(self.current[self.cursor - 1])
-        if ch > max or ch < min:
-            self.cursor -= 1
-            return True
-        ch -= min
-        if (s[ch >> 3] & (0X1 << (ch & 0x7))) == 0:
+        if self.current[self.cursor - 1] not in s:
             self.cursor -= 1
             return True
         return False
 
-    def go_out_grouping_b(self, s, min, max):
+    def go_out_grouping_b(self, s):
         while self.cursor > self.limit_backward:
-            ch = ord(self.current[self.cursor - 1])
-            if ch <= max and ch >= min:
-                ch -= min
-                if (s[ch >> 3] & (0X1 << (ch & 0x7))):
-                    return True
+            if self.current[self.cursor - 1] in s:
+                return True
             self.cursor -= 1
         return False
 
@@ -183,8 +151,7 @@ class BaseStemmer(object):
                 self.cursor = c + len(w.s)
                 if w.method is None:
                     return w.result
-                method = getattr(self, w.method)
-                res = method()
+                res = w.method(self)
                 self.cursor = c + len(w.s)
                 if res:
                     return w.result
@@ -241,8 +208,7 @@ class BaseStemmer(object):
                 self.cursor = c - len(w.s)
                 if w.method is None:
                     return w.result
-                method = getattr(self, w.method)
-                res = method()
+                res = w.method(self)
                 self.cursor = c - len(w.s)
                 if res:
                     return w.result

data/vendor/snowball/python/stemwords.py

@@ -16,7 +16,7 @@ and output files.  If it is omitted, the UTF-8 encoding is used.
 If -p is given the output file consists of each word of the input
 file followed by \"->\" followed by its stemmed equivalent.
 If -p2 is given the output file is a two column layout containing
-the input words in the first column and the stemmed eqivalents in
+the input words in the first column and the stemmed equivalents in
 the second column.
 
 Otherwise, the output file consists of the stemmed words, one per
@@ -36,8 +36,7 @@ def main():
         language = 'English'
         show_help = False
         while len(argv):
-            arg = argv[0]
-            argv = argv[1:]
+            arg = argv.pop(0)
             if arg == '-h':
                 show_help = True
                 break
@@ -49,25 +48,22 @@ def main():
                 if len(argv) == 0:
                     show_help = True
                     break
-                language = argv[0]
-                argv = argv[1:]
+                language = argv.pop(0)
             elif arg == "-i":
                 if len(argv) == 0:
                     show_help = True
                     break
-                input = argv[0]
-                argv = argv[1:]
+                input = argv.pop(0)
             elif arg == "-o":
                 if len(argv) == 0:
                     show_help = True
                     break
-                output = argv[0]
-                argv = argv[1:]
+                output = argv.pop(0)
             elif arg == "-c":
                 if len(argv) == 0:
                     show_help = True
                     break
-                encoding = argv[0]
+                encoding = argv.pop(0)
         if show_help or input == '' or output == '':
             usage()
         else:
@@ -76,8 +72,8 @@ def main():
 
 def stemming(lang, input, output, encoding, pretty):
     stemmer = snowballstemmer.stemmer(lang)
-    with codecs.open(output, "w", encoding) as outfile:
-        with codecs.open(input, "r", encoding) as infile:
+    with codecs.open(input, "r", encoding) as infile:
+        with codecs.open(output, "w", encoding) as outfile:
             for original in infile.readlines():
                 original = original.strip()
                 # Convert only ASCII-letters to lowercase, to match C behavior

data/vendor/snowball/runtime/api.c

@@ -1,23 +1,28 @@
 
-#include <stdlib.h> /* for calloc, free */
+#include <stdlib.h> /* for malloc, calloc, free */
 #include "header.h"
 
 extern struct SN_env * SN_create_env(int S_size, int I_size)
 {
-    struct SN_env * z = (struct SN_env *) calloc(1, sizeof(struct SN_env));
+    static const struct SN_env default_SN_env = {};
+    struct SN_env * z = (struct SN_env *) malloc(sizeof(struct SN_env));
     if (z == NULL) return NULL;
+    *z = default_SN_env;
     z->p = create_s();
     if (z->p == NULL) goto error;
     if (S_size)
     {
         int i;
-        z->S = (symbol * *) calloc(S_size, sizeof(symbol *));
+        z->S = (symbol * *) malloc(S_size * sizeof(symbol *));
         if (z->S == NULL) goto error;
 
         for (i = 0; i < S_size; i++)
         {
             z->S[i] = create_s();
-            if (z->S[i] == NULL) goto error;
+            if (z->S[i] == NULL) {
+                S_size = i;
+                goto error;
+            }
         }
     }
 
@@ -36,7 +41,7 @@ error:
 extern void SN_close_env(struct SN_env * z, int S_size)
 {
     if (z == NULL) return;
-    if (S_size)
+    if (z->S)
     {
         int i;
         for (i = 0; i < S_size; i++)

data/vendor/snowball/runtime/header.h

@@ -1,11 +1,6 @@
 
-#include <limits.h>
-
 #include "api.h"
 
-#define MAXINT INT_MAX
-#define MININT INT_MIN
-
 #define HEAD 2*sizeof(int)
 
 #define SIZE(p)        ((int *)(p))[-1]
@@ -13,10 +8,16 @@
 #define CAPACITY(p)    ((int *)(p))[-2]
 
 struct among
-{   int s_size;     /* number of chars in string */
-    const symbol * s;       /* search string */
-    int substring_i;/* index to longest matching substring */
-    int result;     /* result of the lookup */
+{
+    /* Number of symbols in s. */
+    int s_size;
+    /* Search string. */
+    const symbol * s;
+    /* Delta of index to longest matching substring, or 0 if none. */
+    int substring_i;
+    /* Result of the lookup. */
+    int result;
+    /* Optional condition routine, or NULL if none. */
     int (* function)(struct SN_env *);
 };
 

data/vendor/snowball/runtime/utilities.c

@@ -282,19 +282,19 @@ extern int find_among(struct SN_env * z, const struct among * v, int v_size) {
             first_key_inspected = 1;
         }
     }
+    w = v + i;
     while (1) {
-        w = v + i;
         if (common_i >= w->s_size) {
             z->c = c + w->s_size;
-            if (w->function == 0) return w->result;
+            if (w->function == NULL) return w->result;
             {
                 int res = w->function(z);
                 z->c = c + w->s_size;
                 if (res) return w->result;
             }
         }
-        i = w->substring_i;
-        if (i < 0) return 0;
+        if (!w->substring_i) return 0;
+        w += w->substring_i;
     }
 }
 
@@ -337,19 +337,19 @@ extern int find_among_b(struct SN_env * z, const struct among * v, int v_size) {
             first_key_inspected = 1;
         }
     }
+    w = v + i;
     while (1) {
-        w = v + i;
         if (common_i >= w->s_size) {
             z->c = c - w->s_size;
-            if (w->function == 0) return w->result;
+            if (w->function == NULL) return w->result;
             {
                 int res = w->function(z);
                 z->c = c - w->s_size;
                 if (res) return w->result;
             }
         }
-        i = w->substring_i;
-        if (i < 0) return 0;
+        if (!w->substring_i) return 0;
+        w += w->substring_i;
     }
 }
 
@@ -434,7 +434,7 @@ extern int slice_from_v(struct SN_env * z, const symbol * p) {
 }
 
 extern int slice_del(struct SN_env * z) {
-    return slice_from_s(z, 0, 0);
+    return slice_from_s(z, 0, NULL);
 }
 
 extern int insert_s(struct SN_env * z, int bra, int ket, int s_size, const symbol * s) {

data/vendor/snowball/rust/build.rs

@@ -41,7 +41,7 @@ fn main() {
                 .unwrap();
 
             writeln!(&mut lang_include_file, "pub mod {}_stemmer;", langname).unwrap();
-                
+
         }
     }