From c9259c47238d2c1874468111c404c2c79ccf4c86 Mon Sep 17 00:00:00 2001
From: Cosimo Lupo <cosimo@anthrotype.com>
Date: Thu, 18 Jan 2018 20:20:17 +0000
Subject: [PATCH] unicodedata: add ot_tags_from_script function

Fixes https://github.com/fonttools/fonttools/issues/1112

This implements the same logic found in harbfuzz hb-ot-tag.cc to
convert between Unicode (or ISO 15924) script codes to OpenType script
tags as defined at:
https://www.microsoft.com/typography/otspec/scripttags.htm

https://github.com/harfbuzz/harfbuzz/blob/461a605fdec3361a038d3715adf615353c4f91fa/src/hb-ot-tag.cc#L127
---
 Lib/fontTools/unicodedata/OTTags.py   | 37 +++++++++++++++++++++++++++
 Lib/fontTools/unicodedata/__init__.py | 23 ++++++++++++++++-
 Tests/unicodedata_test.py             | 15 +++++++++++
 3 files changed, 74 insertions(+), 1 deletion(-)
 create mode 100644 Lib/fontTools/unicodedata/OTTags.py

diff --git a/Lib/fontTools/unicodedata/OTTags.py b/Lib/fontTools/unicodedata/OTTags.py
new file mode 100644
index 000000000..a40aec2ef
--- /dev/null
+++ b/Lib/fontTools/unicodedata/OTTags.py
@@ -0,0 +1,37 @@
+# Data updated to OpenType 1.8.2 as of January 2018.
+
+# Complete list of OpenType script tags at:
+# https://www.microsoft.com/typography/otspec/scripttags.htm
+
+# Most of the script tags are the same as the ISO 15924 tag but lowercased,
+# so we only have to handle the exceptional cases:
+# - KATAKANA and HIRAGANA both map to 'kana';
+# - spaces at the end are preserved, unlike ISO 15924;
+# - we map special script codes for Inherited, Common and Unknown to DFLT.
+
+DEFAULT_SCRIPT = "DFLT"
+
+SCRIPT_EXCEPTIONS = {
+    "Hira": "kana",
+    "Hrkt": "kana",
+    "Laoo": "lao ",
+    "Yiii": "yi  ",
+    "Nkoo": "nko ",
+    "Vaii": "vai ",
+    "Zinh": DEFAULT_SCRIPT,
+    "Zyyy": DEFAULT_SCRIPT,
+    "Zzzz": DEFAULT_SCRIPT,
+}
+
+NEW_SCRIPT_TAGS = {
+    "Beng": ("bng2",),
+    "Deva": ("dev2",),
+    "Gujr": ("gjr2",),
+    "Guru": ("gur2",),
+    "Knda": ("knd2",),
+    "Mlym": ("mlm2",),
+    "Orya": ("ory2",),
+    "Taml": ("tml2",),
+    "Telu": ("tel2",),
+    "Mymr": ("mym2",),
+}
diff --git a/Lib/fontTools/unicodedata/__init__.py b/Lib/fontTools/unicodedata/__init__.py
index c2bc29e15..bec2b4f0b 100644
--- a/Lib/fontTools/unicodedata/__init__.py
+++ b/Lib/fontTools/unicodedata/__init__.py
@@ -13,7 +13,7 @@ except ImportError:  # pragma: no cover
     # fall back to built-in unicodedata (possibly outdated)
     from unicodedata import *
 
-from . import Blocks, Scripts, ScriptExtensions
+from . import Blocks, Scripts, ScriptExtensions, OTTags
 
 
 __all__ = [tostr(s) for s in (
@@ -147,3 +147,24 @@ def block(char):
     code = byteord(char)
     i = bisect_right(Blocks.RANGES, code)
     return Blocks.VALUES[i-1]
+
+
+def ot_tags_from_script(script_code):
+    """ Return a list of OpenType script tags associated with a given
+    Unicode script code.
+    Return ['DFLT'] script tag for invalid/unknown script codes.
+    """
+    if script_code not in Scripts.NAMES:
+        return [OTTags.DEFAULT_SCRIPT]
+
+    script_tags = [
+        OTTags.SCRIPT_EXCEPTIONS.get(
+            script_code,
+            script_code[0].lower() + script_code[1:]
+        )
+    ]
+    if script_code in OTTags.NEW_SCRIPT_TAGS:
+        script_tags.extend(OTTags.NEW_SCRIPT_TAGS[script_code])
+        script_tags.reverse()  # last in, first out
+
+    return script_tags
diff --git a/Tests/unicodedata_test.py b/Tests/unicodedata_test.py
index dba02e7c9..a3ac66249 100644
--- a/Tests/unicodedata_test.py
+++ b/Tests/unicodedata_test.py
@@ -203,6 +203,21 @@ def test_block():
     assert unicodedata.block("\u1c90") == "No_Block"
 
 
+def test_ot_tags_from_script():
+    # simple
+    assert unicodedata.ot_tags_from_script("Latn") == ["latn"]
+    # script mapped to multiple new and old script tags
+    assert unicodedata.ot_tags_from_script("Deva") == ["dev2", "deva"]
+    # exceptions
+    assert unicodedata.ot_tags_from_script("Hira") == ["kana"]
+    # special script codes map to DFLT
+    assert unicodedata.ot_tags_from_script("Zinh") == ["DFLT"]
+    assert unicodedata.ot_tags_from_script("Zyyy") == ["DFLT"]
+    assert unicodedata.ot_tags_from_script("Zzzz") == ["DFLT"]
+    # this in invalid or unknown
+    assert unicodedata.ot_tags_from_script("Aaaa") == ["DFLT"]
+
+
 if __name__ == "__main__":
     import sys
     sys.exit(pytest.main(sys.argv))