From b16b5535bfbac7ed7af3e676e367eab0f4c6611b Mon Sep 17 00:00:00 2001
From: Kovid Goyal <kovid@kovidgoyal.net>
Date: Sun, 1 Dec 2019 10:50:32 +0530
Subject: [PATCH] Implement mapping of locales to dictionary names

---
 setup/test.py                                 |  2 +
 src/calibre/utils/hyphenation/dictionaries.py | 39 +++++++++++++++++++
 .../utils/hyphenation/test_hyphenation.py     | 26 +++++++++++++
 3 files changed, 67 insertions(+)
 create mode 100644 src/calibre/utils/hyphenation/test_hyphenation.py

diff --git a/setup/test.py b/setup/test.py
index 4825407a48..41d4d2c32a 100644
--- a/setup/test.py
+++ b/setup/test.py
@@ -133,6 +133,8 @@ def find_tests(which_tests=None):
         a(find_tests())
         from calibre.gui2.viewer.convert_book import find_tests
         a(find_tests())
+        from calibre.utils.hyphenation.test_hyphenation import find_tests
+        a(find_tests())
         if iswindows:
             from calibre.utils.windows.wintest import find_tests
             a(find_tests())
diff --git a/src/calibre/utils/hyphenation/dictionaries.py b/src/calibre/utils/hyphenation/dictionaries.py
index 0407e63cd6..0e683740e0 100644
--- a/src/calibre/utils/hyphenation/dictionaries.py
+++ b/src/calibre/utils/hyphenation/dictionaries.py
@@ -3,3 +3,42 @@
 # License: GPL v3 Copyright: 2019, Kovid Goyal <kovid at kovidgoyal.net>
 
 from __future__ import absolute_import, division, print_function, unicode_literals
+
+import json
+
+from calibre.utils.localization import lang_as_iso639_1
+from polyglot.builtins import iteritems
+from polyglot.functools import lru_cache
+
+
+def locale_map():
+    ans = getattr(locale_map, 'ans', None)
+    if ans is None:
+        ans = locale_map.ans = {k.lower(): v for k, v in iteritems(json.loads(P('hyphenation/locales.json', data=True)))}
+    return ans
+
+
+@lru_cache()
+def dictionary_name_for_locale(loc):
+    loc = loc.lower().replace('-', '_')
+    lmap = locale_map()
+    if loc in lmap:
+        return lmap[loc]
+    parts = loc.split('_')
+    if len(parts) > 2:
+        loc = '_'.join(parts[:2])
+        if loc in lmap:
+            return lmap[loc]
+    loc = lang_as_iso639_1(parts[0])
+    if not loc:
+        return
+    if loc in lmap:
+        return lmap[loc]
+    if loc == 'en':
+        return lmap['en_us']
+    if loc == 'de':
+        return lmap['de_de']
+    q = loc + '_'
+    for k, v in iteritems(lmap):
+        if k.startswith(q):
+            return lmap[k]
diff --git a/src/calibre/utils/hyphenation/test_hyphenation.py b/src/calibre/utils/hyphenation/test_hyphenation.py
new file mode 100644
index 0000000000..b65c78b89a
--- /dev/null
+++ b/src/calibre/utils/hyphenation/test_hyphenation.py
@@ -0,0 +1,26 @@
+#!/usr/bin/env python2
+# vim:fileencoding=utf-8
+# License: GPL v3 Copyright: 2019, Kovid Goyal <kovid at kovidgoyal.net>
+
+from __future__ import absolute_import, division, print_function, unicode_literals
+
+import unittest
+from calibre.utils.hyphenation.dictionaries import dictionary_name_for_locale
+
+
+class TestHyphenation(unittest.TestCase):
+
+    ae = unittest.TestCase.assertEqual
+
+    def test_locale_to_hyphen_dictionary(self):
+        def t(x, expected=None):
+            self.ae(dictionary_name_for_locale(x), 'hyph_{}.dic'.format(expected) if expected else None)
+        t('en', 'en_US')
+        t('en_IN', 'en_GB')
+        t('de', 'de_DE')
+        t('fr', 'fr')
+        t('XXX')
+
+
+def find_tests():
+    return unittest.defaultTestLoader.loadTestsFromTestCase(TestHyphenation)