add implementation of UAX#29 word bounds algorithm

This patch does the following:

1. Adds three new structs in libunicode/str.rs:

   a. UnicodeWords: a filter on the UWordBounds iterator that yields only
      the "words" of a string as defined in Section 4 of Unicode Standard
      Annex rust-lang#29 (UAX#29), http://unicode.org/reports/tr29/#Word_Boundaries

   b. UWordBounds: an iterator that segments a string on its word
      boundaries as defined in UAX#29. Note that this *only* segments
      the string, and does *not* drop whitespace and other non-word
      pieces of the text (that's what UnicodeWords does).

      Note that UWordBounds has both a forward and backward iterator
      that have total running time (that is, to segment the entire
      string) linear in the size of the string. It should be noted that
      with pathological inputs the reverse iterator could be about 2x less
      efficient than the forward iterator, but on reasonable inputs
      their costs are similar.

   c. UWordBoundIndices: the above iterator, but returning tuples of
      (offset, &str).

2. Adds three new functions in the `UnicodeStr` trait:

   a. words_unicode(): returns a UnicodeWords iterator.

   b. split_words_uax29(): returns a UWordBounds iterator.

   c. split_words_uax29_indices(): returns a UWordBoundIndices iterator.

3. Updates the `src/etc/unicode.py` script to generate tables necessary
   for running the UWordBounds iterators.

4. Adds a new script, `src/etc/unicode_gen_breaktests.py`,
   which processes the grapheme and word break tests published
   by the Unicode consortium into a format for inclusion in
   libcollectionstest.

5. Adds new impls in libcollections's `str` corresponding to the
   `UnicodeStr` functions of (2).

   Note that this new functionality is gated with `feature(unicode)`.

6. Adds tests in libcollectionstest to exercise this new functionality.

   In addition, updates the test data for the graphemes test to
   correspond to the output from the script of (4). (Note that at the
   moment this change is primarily cosmetic.)

This patch does not settle the question raised by @huonw in rust-lang#15628;
rather, it introduces a new function alongside `words()` that follows
UAX#29.

In addition, it does not address the concerns that @SimonSapin raises in
rust-lang/rfcs#1054 since it leaves `words()`
alone.

Author: kwantam

src/etc/unicode.py

@@ -15,6 +15,7 @@
 # - DerivedNormalizationProps.txt
 # - EastAsianWidth.txt
 # - auxiliary/GraphemeBreakProperty.txt
+# - auxiliary/WordBreakProperty.txt
 # - PropList.txt
 # - ReadMe.txt
 # - Scripts.txt
@@ -290,11 +291,13 @@ def emit_bsearch_range_table(f):
 """)
 
 def emit_table(f, name, t_data, t_type = "&'static [(char, char)]", is_pub=True,
-        pfun=lambda x: "(%s,%s)" % (escape_char(x[0]), escape_char(x[1]))):
-    pub_string = ""
+        pfun=lambda x: "(%s,%s)" % (escape_char(x[0]), escape_char(x[1])), is_const=True):
+    pub_string = "const"
+    if not is_const:
+        pub_string = "let"
     if is_pub:
-        pub_string = "pub "
-    f.write("    %sconst %s: %s = &[\n" % (pub_string, name, t_type))
+        pub_string = "pub " + pub_string
+    f.write("    %s %s: %s = &[\n" % (pub_string, name, t_type))
     data = ""
     first = True
     for dat in t_data:
@@ -375,21 +378,25 @@ def emit_conversions_module(f, lowerupper, upperlower):
         sorted(lowerupper.iteritems(), key=operator.itemgetter(0)), is_pub=False)
     f.write("}\n\n")
 
-def emit_grapheme_module(f, grapheme_table, grapheme_cats):
-    f.write("""pub mod grapheme {
+def emit_break_module(f, break_table, break_cats, name):
+    Name = name.capitalize()
+    f.write("""pub mod %s {
     use core::slice::SliceExt;
-    pub use self::GraphemeCat::*;
+    pub use self::%sCat::*;
     use core::result::Result::{Ok, Err};
 
     #[allow(non_camel_case_types)]
-    #[derive(Clone, Copy)]
-    pub enum GraphemeCat {
-""")
-    for cat in grapheme_cats + ["Any"]:
-        f.write("        GC_" + cat + ",\n")
+    #[derive(Clone, Copy, PartialEq, Eq)]
+    pub enum %sCat {
+""" % (name, Name, Name))
+
+    break_cats.append("Any")
+    break_cats.sort()
+    for cat in break_cats:
+        f.write(("        %sC_" % Name[0]) + cat + ",\n")
     f.write("""    }
 
-    fn bsearch_range_value_table(c: char, r: &'static [(char, char, GraphemeCat)]) -> GraphemeCat {
+    fn bsearch_range_value_table(c: char, r: &'static [(char, char, %sCat)]) -> %sCat {
         use core::cmp::Ordering::{Equal, Less, Greater};
         match r.binary_search_by(|&(lo, hi, _)| {
             if lo <= c && c <= hi { Equal }
@@ -400,19 +407,19 @@ def emit_grapheme_module(f, grapheme_table, grapheme_cats):
                 let (_, _, cat) = r[idx];
                 cat
             }
-            Err(_) => GC_Any
+            Err(_) => %sC_Any
         }
     }
 
-    pub fn grapheme_category(c: char) -> GraphemeCat {
-        bsearch_range_value_table(c, grapheme_cat_table)
+    pub fn %s_category(c: char) -> %sCat {
+        bsearch_range_value_table(c, %s_cat_table)
     }
 
-""")
+""" % (Name, Name, Name[0], name, Name, name))
 
-    emit_table(f, "grapheme_cat_table", grapheme_table, "&'static [(char, char, GraphemeCat)]",
-        pfun=lambda x: "(%s,%s,GC_%s)" % (escape_char(x[0]), escape_char(x[1]), x[2]),
-        is_pub=False)
+    emit_table(f, "%s_cat_table" % name, break_table, "&'static [(char, char, %sCat)]" % Name,
+        pfun=lambda x: "(%s,%s,%sC_%s)" % (escape_char(x[0]), escape_char(x[1]), Name[0], x[2]),
+        is_pub=False, is_const=True)
     f.write("}\n")
 
 def emit_charwidth_module(f, width_table):
@@ -690,4 +697,12 @@ def optimize_width_table(wtable):
         for cat in grapheme_cats:
             grapheme_table.extend([(x, y, cat) for (x, y) in grapheme_cats[cat]])
         grapheme_table.sort(key=lambda w: w[0])
-        emit_grapheme_module(rf, grapheme_table, grapheme_cats.keys())
+        emit_break_module(rf, grapheme_table, grapheme_cats.keys(), "grapheme")
+        rf.write("\n")
+
+        word_cats = load_properties("auxiliary/WordBreakProperty.txt", [])
+        word_table = []
+        for cat in word_cats:
+            word_table.extend([(x, y, cat) for (x, y) in word_cats[cat]])
+        word_table.sort(key=lambda w: w[0])
+        emit_break_module(rf, word_table, word_cats.keys(), "word")

src/etc/unicode_gen_breaktests.py

@@ -0,0 +1,196 @@
+#!/usr/bin/env python
+# -*- coding: utf-8
+#
+# Copyright 2015 The Rust Project Developers. See the COPYRIGHT
+# file at the top-level directory of this distribution and at
+# http://rust-lang.org/COPYRIGHT.
+#
+# Licensed under the Apache License, Version 2.0 <LICENSE-APACHE or
+# http://www.apache.org/licenses/LICENSE-2.0> or the MIT license
+# <LICENSE-MIT or http://opensource.org/licenses/MIT>, at your
+# option. This file may not be copied, modified, or distributed
+# except according to those terms.
+
+# This script uses the following Unicode tables:
+# - GraphemeBreakTest.txt
+# - WordBreakTest.txt
+#
+# Since this should not require frequent updates, we just store this
+# out-of-line and check the unicode.rs file into git.
+
+import unicode, re, os, fileinput
+
+def load_test_data(f, optsplit=[]):
+    outls = []
+    testRe1 = re.compile("^÷\s+([^\s].*[^\s])\s+÷\s+#\s+÷\s+\[0.2\].*?([÷×].*)\s+÷\s+\[0.3\]\s*$")
+
+    unicode.fetch(f)
+    data = []
+    for line in fileinput.input(os.path.basename(f)):
+        # lines that include a test start with the ÷ character
+        if len(line) < 2 or line[0:2] != '÷':
+            continue
+
+        m = testRe1.match(line)
+        if not m:
+            print "error: no match on line where test was expected: %s" % line
+            continue
+
+        # process the characters in this test case
+        chars = process_split_string(m.group(1))
+        # skip test case if it contains invalid characters (viz., surrogates)
+        if not chars:
+            continue
+
+        # now process test cases
+        (chars, info) = process_split_info(m.group(2), chars, optsplit)
+
+        # make sure that we have break info for each break!
+        assert len(chars) - 1 == len(info)
+
+        outls.append((chars, info))
+
+    return outls
+
+def process_split_info(s, c, o):
+    outcs = []
+    outis = []
+    workcs = c.pop(0)
+
+    # are we on a × or a ÷?
+    isX = False
+    if s[0:2] == '×':
+        isX = True
+
+    # find each instance of '(÷|×) [x.y] '
+    while s:
+        # find the currently considered rule number
+        sInd = s.index('[') + 1
+        eInd = s.index(']')
+
+        # if it's '× [a.b]' where 'a.b' is in o, then
+        # we consider it a split even though it's not
+        # marked as one
+        # if it's ÷ then it's always a split
+        if not isX or s[sInd:eInd] in o:
+            outis.append(s[sInd:eInd])
+            outcs.append(workcs)
+            workcs = c.pop(0)
+        else:
+            workcs.extend(c.pop(0))
+
+        idx = 1
+        while idx < len(s):
+            if s[idx:idx+2] == '×':
+                isX = True
+                break
+            if s[idx:idx+2] == '÷':
+                isX = False
+                break
+            idx += 1
+        s = s[idx:]
+    
+    outcs.append(workcs)
+    return (outcs, outis)
+
+def process_split_string(s):
+    outls = []
+    workls = []
+
+    inls = s.split()
+
+    for i in inls:
+        if i == '÷' or i == '×':
+            outls.append(workls)
+            workls = []
+            continue
+
+        ival = int(i,16)
+
+        if unicode.is_surrogate(ival):
+            return []
+
+        workls.append(ival)
+
+    if workls:
+        outls.append(workls)
+
+    return outls
+
+def showfun(x):
+    outstr = '("'
+    for c in x[0]:
+        outstr += "\\u{%x}" % c
+    outstr += '",&['
+    xfirst = True
+    for xx in x[1:]:
+        if not xfirst:
+            outstr += '],&['
+        xfirst = False
+        sfirst = True
+        for sp in xx:
+            if not sfirst:
+                outstr += ','
+            sfirst = False
+            outstr += '"'
+            for c in sp:
+                outstr += "\\u{%x}" % c
+            outstr += '"'
+    outstr += '])'
+    return outstr
+
+def create_grapheme_data():
+    # rules 9.1 and 9.2 are for extended graphemes only
+    optsplits = ['9.1','9.2']
+    d = load_test_data("auxiliary/GraphemeBreakTest.txt", optsplits)
+
+    test_same = []
+    test_diff = []
+
+    for (c, i) in d:
+        allchars = [cn for s in c for cn in s]
+        extgraphs = []
+        extwork = []
+
+        extwork.extend(c[0])
+        for n in range(0,len(i)):
+            if i[n] in optsplits:
+                extwork.extend(c[n+1])
+            else:
+                extgraphs.append(extwork)
+                extwork = []
+                extwork.extend(c[n+1])
+
+        # these are the extended grapheme clusters
+        extgraphs.append(extwork)
+
+        if extgraphs == c:
+            test_same.append((allchars, c))
+        else:
+            test_diff.append((allchars, extgraphs, c))
+
+    stype = "&[(&str, &[&str])]"
+    dtype = "&[(&str, &[&str], &[&str])]"
+    with open("graph_tests.rs", "w") as rf:
+        rf.write("    // official Unicode test data\n")
+        rf.write("    // http://www.unicode.org/Public/UNIDATA/auxiliary/GraphemeBreakTest.txt\n")
+        unicode.emit_table(rf, "test_same", test_same, stype, False, showfun, False)
+        unicode.emit_table(rf, "test_diff", test_diff, dtype, False, showfun, False)
+    
+def create_words_data():
+    d = load_test_data("auxiliary/WordBreakTest.txt")
+
+    test = []
+
+    for (c, i) in d:
+        allchars = [cn for s in c for cn in s]
+        test.append((allchars, c))
+
+    wtype = "&[(&str, &[&str])]"
+    with open("word_tests.rs", "w") as rf:
+        rf.write("    // official Unicode test data\n")
+        rf.write("    // http://www.unicode.org/Public/UNIDATA/auxiliary/WordBreakTest.txt\n")
+        unicode.emit_table(rf, "test_word", test, wtype, False, showfun, False)
+
+create_grapheme_data()
+create_words_data()

src/libcollections/str.rs

@@ -79,6 +79,7 @@ pub use core::str::{MatchIndices, RMatchIndices};
 pub use core::str::{from_utf8, Chars, CharIndices, Bytes};
 pub use core::str::{from_utf8_unchecked, ParseBoolError};
 pub use unicode::str::{Words, Graphemes, GraphemeIndices};
+pub use unicode::str::{UnicodeWords, UWordBounds, UWordBoundIndices};
 pub use core::str::pattern;
 
 /*
@@ -1736,6 +1737,30 @@ impl str {
         UnicodeStr::words(&self[..])
     }
 
+    /// An iterator over the words of `self`, separated on
+    /// [UAX#29 word boundaries](http://www.unicode.org/reports/tr29/#Word_Boundaries).
+    ///
+    /// In this function, "words" are just those substrings which, after splitting on
+    /// UAX#29 word boundaries, contain any alphanumeric characters. That is, the
+    /// substring must contain at least one character with the
+    /// [Alphabetic](http://unicode.org/reports/tr44/#Alphabetic)
+    /// property, or with
+    /// [General_Category=Number](http://unicode.org/reports/tr44/#General_Category_Values).
+    ///
+    /// # Example
+    /// # #![feature(unicode, core)]
+    /// let uws = "The quick (\"brown\") fox can't jump 32.3 feet, right?";
+    /// let uw1 = uws.words_unicode().collect::<Vec<&str>>();
+    /// let b: &[_] = &["The", "quick", "brown", "fox", "can't", "jump", "32.3", "feet", "right"];
+    ///
+    /// assert_eq!(&uw1[..], b);
+    /// ```
+    #[unstable(feature = "unicode",
+               reason = "questions remain regarding the naming of words() and words_unicode()")]
+    pub fn words_unicode(&self) -> UnicodeWords {
+        UnicodeStr::words_unicode(&self[..])
+    }
+
     /// Returns a string's displayed width in columns.
     ///
     /// Control characters have zero width.
@@ -1819,4 +1844,43 @@ impl str {
         s.extend(self[..].chars().flat_map(|c| c.to_uppercase()));
         return s;
     }
+
+    /// Returns an iterator over substrings of `self` separated on
+    /// [UAX#29 word boundaries](http://www.unicode.org/reports/tr29/#Word_Boundaries).
+    ///
+    /// The concatenation of the substrings returned by this function is just the original string.
+    ///
+    /// # Example
+    ///
+    /// ```
+    /// # #![feature(unicode, core)]
+    /// let swu1 = "The quick (\"brown\")  fox".split_words_uax29().collect::<Vec<&str>>();
+    /// let b: &[_] = &["The", " ", "quick", " ", "(", "\"", "brown", "\"", ")", " ", " ", "fox"];
+    ///
+    /// assert_eq!(&swu1[..], b);
+    /// ```
+    #[unstable(feature = "unicode",
+               reason = "this functionality may only be provided by libunicode")]
+    pub fn split_words_uax29(&self) -> UWordBounds {
+        UnicodeStr::split_words_uax29(&self[..])
+    }
+
+    /// Returns an iterator over substrings of `self`, split on UAX#29 word boundaries,
+    /// and their offsets. See `split_words_uax29()` for more information.
+    ///
+    /// # Example
+    ///
+    /// ```
+    /// # #![feature(unicode, core)]
+    /// let swi1 = "Brr, it's 29.3°F!".split_words_uax29_indices().collect::<Vec<(usize, &str)>>();
+    /// let b: &[_] = &[(0, "Brr"), (3, ","), (4, " "), (5, "it's"), (9, " "), (10, "29.3"),
+    ///                 (14, "°"), (16, "F"), (17, "!")];
+    ///
+    /// assert_eq!(&swi1[..], b);
+    /// ```
+    #[unstable(feature = "unicode",
+               reason = "this functionality may only be provided by libunicode")]
+    pub fn split_words_uax29_indices(&self) -> UWordBoundIndices {
+        UnicodeStr::split_words_uax29_indices(&self[..])
+    }
 }