scriptconv 0.0.3a3__tar.gz → 0.0.3a5__tar.gz

{scriptconv-0.0.3a3 → scriptconv-0.0.3a5}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: scriptconv
-Version: 0.0.3a3
+Version: 0.0.3a5
 Summary: Zero-dependency script & phoneme-notation core — ISO-15924 detection & metadata, IPA↔ARPABET/X-SAMPA/Lexique/Kirshenbaum/Cotovía/RFE, Buckwalter↔Arabic, Hangul→jamo, kana
 License: Apache-2.0
 Project-URL: Homepage, https://github.com/TigreGotico/scriptconv

{scriptconv-0.0.3a3 → scriptconv-0.0.3a5}/pyproject.toml

@@ -46,6 +46,7 @@ include = ["scriptconv*"]
 
 [tool.setuptools.package-data]
 scriptconv = ["py.typed"]
+"scriptconv.data" = ["*.tsv.gz"]
 
 [tool.pytest.ini_options]
 testpaths = ["tests"]

{scriptconv-0.0.3a3 → scriptconv-0.0.3a5}/scriptconv/__init__.py

@@ -20,6 +20,8 @@ translit:
 readings:
     Dictionary-backed respelling (Japanese kanji → kana via the ``ja``
     extra; Chinese hanzi → pinyin/bopomofo via the ``zh`` extra).
+cangjie:
+    Hanzi → Cangjie5 input codes (shape decomposition, vendored table).
 
 Zero required runtime dependencies (stdlib only); the readings module
 needs the optional ``ja``/``zh`` extras.
@@ -71,6 +73,10 @@ from scriptconv.readings import (
     to_pinyin,
     to_bopomofo,
 )
+from scriptconv.cangjie import (
+    cangjie_code,
+    to_cangjie,
+)
 
 __all__ = [
     # scripts
@@ -113,6 +119,9 @@ __all__ = [
     # readings (dictionary-backed, needs scriptconv[ja])
     "to_hiragana",
     "to_katakana",
+    # cangjie (vendored Cangjie5 table)
+    "cangjie_code",
+    "to_cangjie",
     "to_pinyin",
     "to_bopomofo",
 ]

scriptconv-0.0.3a5/scriptconv/cangjie.py

@@ -0,0 +1,63 @@
+"""Hanzi → Cangjie input codes (fifth generation, traditional table).
+
+Cangjie codes encode a glyph's SHAPE — its decomposition into geometric
+radicals — not its reading: 日 is ``a`` because of what it looks like, however
+it is pronounced.  That makes this a deterministic, table-driven orthographic
+transcode like the rest of scriptconv's core: no reading dictionary, no
+ambiguity, no phonology.
+
+The table is vendored (gzipped, loaded lazily on first use) from the Cangjie5
+project — https://github.com/Jackchows/Cangjie5, MIT License — so no runtime
+download or extra dependency is needed.  Where the upstream table lists
+alternative codes for a glyph, the primary (first) code is kept.
+"""
+from __future__ import annotations
+
+import gzip
+from importlib.resources import files
+from typing import Dict, Optional
+
+__all__ = ["cangjie_code", "to_cangjie"]
+
+_TABLE: Optional[Dict[str, str]] = None
+
+
+def _table() -> Dict[str, str]:
+    global _TABLE
+    if _TABLE is None:
+        _TABLE = {}
+        raw = files("scriptconv.data").joinpath("cangjie5_tc.tsv.gz").read_bytes()
+        for line in gzip.decompress(raw).decode("utf-8").splitlines():
+            if line.startswith("#"):
+                continue
+            ch, _, code = line.partition("\t")
+            _TABLE[ch] = code
+    return _TABLE
+
+
+def cangjie_code(char: str) -> Optional[str]:
+    """Return the Cangjie5 code for a single glyph, or None if unmapped."""
+    return _table().get(char)
+
+
+def to_cangjie(text: str, sep: str = " ") -> str:
+    """Transcode hanzi in *text* to their Cangjie5 codes.
+
+    Mapped glyphs become their code, joined by *sep*; runs of unmapped
+    characters (Latin, digits, punctuation, rare glyphs) pass through
+    unchanged as single tokens.
+    """
+    table = _table()
+    tokens = []
+    last_raw = False
+    for ch in text:
+        code = table.get(ch)
+        if code is not None:
+            tokens.append(code)
+            last_raw = False
+        elif last_raw:
+            tokens[-1] += ch
+        else:
+            tokens.append(ch)
+            last_raw = True
+    return sep.join(tokens)

{scriptconv-0.0.3a3 → scriptconv-0.0.3a5}/scriptconv/version.py

@@ -2,7 +2,7 @@
 VERSION_MAJOR = 0
 VERSION_MINOR = 0
 VERSION_BUILD = 3
-VERSION_ALPHA = 3
+VERSION_ALPHA = 5
 # END_VERSION_BLOCK
 
 VERSION_STR = f"{VERSION_MAJOR}.{VERSION_MINOR}.{VERSION_BUILD}"

{scriptconv-0.0.3a3 → scriptconv-0.0.3a5}/scriptconv.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: scriptconv
-Version: 0.0.3a3
+Version: 0.0.3a5
 Summary: Zero-dependency script & phoneme-notation core — ISO-15924 detection & metadata, IPA↔ARPABET/X-SAMPA/Lexique/Kirshenbaum/Cotovía/RFE, Buckwalter↔Arabic, Hangul→jamo, kana
 License: Apache-2.0
 Project-URL: Homepage, https://github.com/TigreGotico/scriptconv

{scriptconv-0.0.3a3 → scriptconv-0.0.3a5}/scriptconv.egg-info/SOURCES.txt

@@ -4,6 +4,7 @@ pyproject.toml
 requirements.txt
 scriptconv/__init__.py
 scriptconv/__main__.py
+scriptconv/cangjie.py
 scriptconv/notation.py
 scriptconv/py.typed
 scriptconv/readings.py
@@ -15,6 +16,9 @@ scriptconv.egg-info/SOURCES.txt
 scriptconv.egg-info/dependency_links.txt
 scriptconv.egg-info/requires.txt
 scriptconv.egg-info/top_level.txt
+scriptconv/data/__init__.py
+scriptconv/data/cangjie5_tc.tsv.gz
+tests/test_cangjie.py
 tests/test_cli.py
 tests/test_examples.py
 tests/test_notation.py

scriptconv-0.0.3a5/tests/test_cangjie.py

@@ -0,0 +1,58 @@
+import unittest
+
+from scriptconv import cangjie_code, to_cangjie
+from scriptconv import cangjie
+
+
+class TestCangjieCode(unittest.TestCase):
+    def test_official_single_radical_glyphs(self):
+        self.assertEqual(cangjie_code("日"), "a")
+        self.assertEqual(cangjie_code("木"), "d")
+
+    def test_multi_radical_glyph(self):
+        self.assertEqual(cangjie_code("昌"), "aa")
+
+    def test_unmapped_returns_none(self):
+        self.assertIsNone(cangjie_code("X"))
+        self.assertIsNone(cangjie_code("1"))
+        self.assertIsNone(cangjie_code(" "))
+
+    def test_codes_are_lowercase_alpha(self):
+        table = cangjie._table()
+        self.assertGreater(len(table), 100000)
+        sample = list(table.values())[:5000]
+        self.assertTrue(all(c.isalpha() and c.islower() for c in sample))
+
+
+class TestTableIsPackaged(unittest.TestCase):
+    def test_data_file_resolvable_via_importlib_resources(self):
+        from importlib.resources import files
+        raw = files("scriptconv.data").joinpath("cangjie5_tc.tsv.gz").read_bytes()
+        self.assertGreater(len(raw), 100000)
+
+
+class TestToCangjie(unittest.TestCase):
+    def test_all_hanzi_sentence(self):
+        self.assertEqual(to_cangjie("倉頡"), "oiar grmbc")
+
+    def test_custom_separator(self):
+        self.assertEqual(to_cangjie("倉頡", sep="|"), "oiar|grmbc")
+
+    def test_unmapped_runs_stay_contiguous(self):
+        self.assertEqual(to_cangjie("我有ABC 123!"), "hqi kb ABC 123!")
+
+    def test_non_chinese_only_passes_through_verbatim(self):
+        self.assertEqual(to_cangjie("OVOS 123!"), "OVOS 123!")
+
+    def test_empty_string(self):
+        self.assertEqual(to_cangjie(""), "")
+
+    def test_table_loaded_once(self):
+        to_cangjie("日")
+        first = cangjie._TABLE
+        cangjie_code("木")
+        self.assertIs(cangjie._TABLE, first)
+
+
+if __name__ == "__main__":
+    unittest.main()