fix(study): pinyin of names and sayings is typed and checked right

CC-CEDICT capitalises names, so their readings start with a capital vowel
with its tone mark (欧洲 Ōu zhōu, 澳门 Ào mén, 俄语 É yǔ), which the test
for pinyin didn't know: those words were never typed. It now reads the
reading in lower case.

CC-CEDICT also keeps a saying's comma, written between spaces (bù pà màn ,
jiù pà zhàn), and the comma split the reading into two readings, so typing
the whole saying was wrong and half of it right. A comma with a space
before it now stays part of the reading, and commas count for nothing when
readings are compared; a comma written straight after a reading (xíng,
háng) still separates readings.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
Signed-off-by: Tizian 「ティツィアーン」 <raisondetredev@pm.me>

halkyon_learn_flutter/lib/src/core/chinese/pinyin_input.dart

@@ -77,22 +77,33 @@ class const PinyinInputFormatter() extends TextInputFormatter {
 }
 
 /// [pinyin] as pinyin is compared: with tone marks ([withToneMarks]),
-/// lower case, without spaces, apostrophes or hyphens (`Xué xí`, `xue2xi2`
-/// → `xuéxí`).
+/// lower case, without spaces, apostrophes, hyphens or commas (`Xué xí`,
+/// `xue2xi2` → `xuéxí`; a saying's `màn , jiù` → `mànjiù`).
 String pinyinKey(String pinyin) => withToneMarks(
   pinyin.trim(),
-).toLowerCase().replaceAll(RegExp(r"[\s'’·-]+"), '');
+).toLowerCase().replaceAll(RegExp(r"[\s'’·,，-]+"), '');
 
 /// The readings a card's pinyin field gives, one or several (`xíng,
 /// háng`), as compared ([pinyinKey]).
 Set<String> givenPinyin(String field) => {
-  for (final reading in field.split(RegExp('[,，;；/／、]')))
+  for (final reading in field.split(_readingsApart))
     if (pinyinKey(reading).isNotEmpty) pinyinKey(reading),
 };
 
-/// Whether [reading] is pinyin: Latin letters, with or without tone marks,
-/// and no kana or Han characters.
-bool isPinyin(String reading) =>
-    reading.trim().isNotEmpty && _pinyin.hasMatch(reading.trim());
+/// What separates the readings of a pinyin field: a comma written straight
+/// after a reading (`xíng, háng`), a semicolon, a slash or an enumeration
+/// comma. A comma with a space before it is a saying's own, as CC-CEDICT
+/// writes its punctuation between spaces (`bù pà màn , jiù pà zhàn`): the
+/// whole saying is its one reading.
+final _readingsApart = RegExp(r'(?<!\s)[,，]|[;；/／、]');
 
-final _pinyin = RegExp(r"^[A-Za-zÜüāáǎàēéěèīíǐìōóǒòūúǔùǖǘǚǜ\s'’·,，:-]+$");
+/// Whether [reading] is pinyin: Latin letters, with or without tone marks,
+/// whatever their case (CC-CEDICT capitalises names: `Ōu zhōu`), and no
+/// kana or Han characters.
+bool isPinyin(String reading) {
+  final lower = reading.trim().toLowerCase();
+  return lower.isNotEmpty && _pinyin.hasMatch(lower);
+}
+
+/// Pinyin in lower case, as [isPinyin] reads it.
+final _pinyin = RegExp(r"^[a-züāáǎàēéěèīíǐìōóǒòūúǔùǖǘǚǜ\s'’·,，:-]+$");

halkyon_learn_flutter/test/core/chinese/pinyin_input_test.dart

@@ -27,6 +27,17 @@ void main() {
     expect(givenPinyin('xíng').contains(pinyinKey('xing1')), isFalse);
   });
 
+  test("a saying's comma is part of its reading, as CC-CEDICT writes it "
+      'between spaces; a comma after a reading separates readings', () {
+    final saying = givenPinyin('bù pà màn , jiù pà zhàn');
+
+    expect(saying.contains(pinyinKey('bu4 pa4 man4 jiu4 pa4 zhan4')), isTrue);
+    expect(saying.contains(pinyinKey('bu4pa4man4, jiu4pa4zhan4')), isTrue);
+    expect(saying.contains(pinyinKey('bu4 pa4 man4')), isFalse);
+    expect(saying.contains(pinyinKey('jiu4 pa4 zhan4')), isFalse);
+    expect(givenPinyin('xíng，háng'), {'xíng', 'háng'});
+  });
+
   test("pinyin is told from kana and from words' own letters", () {
     expect(isPinyin('xué xí'), isTrue);
     expect(isPinyin('shuō'), isTrue);
@@ -35,6 +46,14 @@ void main() {
     expect(isPinyin(''), isFalse);
   });
 
+  test('a proper noun, capitalised with its tone mark, is pinyin too', () {
+    expect(isPinyin('Ōu zhōu'), isTrue);
+    expect(isPinyin('Ào mén'), isTrue);
+    expect(isPinyin('É yǔ'), isTrue);
+    expect(isPinyin('Lǘ'), isTrue);
+    expect(givenPinyin('Ōu zhōu').contains(pinyinKey('ou1zhou1')), isTrue);
+  });
+
   test('typing turns a syllable and its tone into the marked syllable, the '
       'caret where it was typed', () {
     const formatter = PinyinInputFormatter();