scriptconv 0.0.1a5__tar.gz → 0.0.1a6__tar.gz

{scriptconv-0.0.1a5 → scriptconv-0.0.1a6}/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: scriptconv
-Version: 0.0.1a5
+Version: 0.0.1a6
 Summary: Shared script-conversion core — ISO-15924 detection, IPA↔ARPA/X-SAMPA/Lexique, Buckwalter, Hangul→jamo
 License: Apache-2.0
 Project-URL: Homepage, https://github.com/TigreGotico/scriptconv
@@ -75,7 +75,7 @@ from scriptconv import (
     xsampa_to_ipa, ipa_to_xsampa,
     buckwalter_to_arabic, arabic_to_buckwalter,
     lexique_to_ipa, ipa_to_lexique,
-    kirshenbaum_to_ipa, ipa_to_kirshenbaum,
+    kirshenbaum_to_ipa, ipa_to_kirshenbaum, looks_like_ipa,
     decompose_hangul, hira_to_kana, kana_to_hira,
     convert, can_convert, convert_batch, Notation, NOTATION_INFO,
 )
@@ -107,6 +107,9 @@ ipa_to_lexique("vɛ̃")               # "v5"  (vin)
 kirshenbaum_to_ipa("S")             # "ʃ"   (espeak-ng ASCII-IPA)
 ipa_to_kirshenbaum("ŋ")             # "N"
 
+looks_like_ipa("pʰɑtʃ")             # True  (heuristic: has IPA-distinctive symbols)
+looks_like_ipa("hello")             # False
+
 decompose_hangul("국민")             # "ㄱㅜㄱㅁㅣㄴ"  (orthographic jamo, no assimilation)
 hira_to_kana("ひらがな")             # "ヒラガナ"
 kana_to_hira("カタカナ")             # "かたかな"
@@ -136,7 +139,7 @@ python -m scriptconv lang ko
 | Module | Contents |
 |--------|----------|
 | `scriptconv.scripts` | `Script` dataclass (with `script_type`), `SCRIPT_REGISTRY` (34 scripts), `detect_script`, `char_script`, `script_distribution`, `script_runs`, `base_direction`, `lang_to_script`, `script_to_langs`, `normalize_script_tag` |
-| `scriptconv.notation` | `Notation` enum, `NotationInfo`/`NOTATION_INFO` fidelity registry, `convert` facade, `can_convert` predicate, `convert_batch` generator, pair-wise converters (ARPABET ↔ IPA, X-SAMPA ↔ IPA, Buckwalter ↔ Arabic, Lexique ↔ IPA, Kirshenbaum ↔ IPA) |
+| `scriptconv.notation` | `Notation` enum, `NotationInfo`/`NOTATION_INFO` fidelity registry, `convert` facade, `can_convert` predicate, `convert_batch` generator, pair-wise converters (ARPABET ↔ IPA, X-SAMPA ↔ IPA, Buckwalter ↔ Arabic, Lexique ↔ IPA, Kirshenbaum ↔ IPA), `looks_like_ipa` detector |
 | `scriptconv.translit` | `decompose_hangul` (Hangul blocks → jamo, compatibility or conjoining), `hira_to_kana`/`kana_to_hira` — all orthographic only |
 
 ## Documentation

{scriptconv-0.0.1a5 → scriptconv-0.0.1a6}/README.md

@@ -48,7 +48,7 @@ from scriptconv import (
     xsampa_to_ipa, ipa_to_xsampa,
     buckwalter_to_arabic, arabic_to_buckwalter,
     lexique_to_ipa, ipa_to_lexique,
-    kirshenbaum_to_ipa, ipa_to_kirshenbaum,
+    kirshenbaum_to_ipa, ipa_to_kirshenbaum, looks_like_ipa,
     decompose_hangul, hira_to_kana, kana_to_hira,
     convert, can_convert, convert_batch, Notation, NOTATION_INFO,
 )
@@ -80,6 +80,9 @@ ipa_to_lexique("vɛ̃")               # "v5"  (vin)
 kirshenbaum_to_ipa("S")             # "ʃ"   (espeak-ng ASCII-IPA)
 ipa_to_kirshenbaum("ŋ")             # "N"
 
+looks_like_ipa("pʰɑtʃ")             # True  (heuristic: has IPA-distinctive symbols)
+looks_like_ipa("hello")             # False
+
 decompose_hangul("국민")             # "ㄱㅜㄱㅁㅣㄴ"  (orthographic jamo, no assimilation)
 hira_to_kana("ひらがな")             # "ヒラガナ"
 kana_to_hira("カタカナ")             # "かたかな"
@@ -109,7 +112,7 @@ python -m scriptconv lang ko
 | Module | Contents |
 |--------|----------|
 | `scriptconv.scripts` | `Script` dataclass (with `script_type`), `SCRIPT_REGISTRY` (34 scripts), `detect_script`, `char_script`, `script_distribution`, `script_runs`, `base_direction`, `lang_to_script`, `script_to_langs`, `normalize_script_tag` |
-| `scriptconv.notation` | `Notation` enum, `NotationInfo`/`NOTATION_INFO` fidelity registry, `convert` facade, `can_convert` predicate, `convert_batch` generator, pair-wise converters (ARPABET ↔ IPA, X-SAMPA ↔ IPA, Buckwalter ↔ Arabic, Lexique ↔ IPA, Kirshenbaum ↔ IPA) |
+| `scriptconv.notation` | `Notation` enum, `NotationInfo`/`NOTATION_INFO` fidelity registry, `convert` facade, `can_convert` predicate, `convert_batch` generator, pair-wise converters (ARPABET ↔ IPA, X-SAMPA ↔ IPA, Buckwalter ↔ Arabic, Lexique ↔ IPA, Kirshenbaum ↔ IPA), `looks_like_ipa` detector |
 | `scriptconv.translit` | `decompose_hangul` (Hangul blocks → jamo, compatibility or conjoining), `hira_to_kana`/`kana_to_hira` — all orthographic only |
 
 ## Documentation

{scriptconv-0.0.1a5 → scriptconv-0.0.1a6}/scriptconv/__init__.py

@@ -50,6 +50,7 @@ from scriptconv.notation import (
     ipa_to_lexique,
     kirshenbaum_to_ipa,
     ipa_to_kirshenbaum,
+    looks_like_ipa,
 )
 from scriptconv.translit import (
     decompose_hangul,
@@ -86,6 +87,7 @@ __all__ = [
     "ipa_to_lexique",
     "kirshenbaum_to_ipa",
     "ipa_to_kirshenbaum",
+    "looks_like_ipa",
     # translit
     "decompose_hangul",
     "hira_to_kana",

{scriptconv-0.0.1a5 → scriptconv-0.0.1a6}/scriptconv/notation.py

@@ -37,6 +37,7 @@ __all__ = [
     "ipa_to_lexique",
     "kirshenbaum_to_ipa",
     "ipa_to_kirshenbaum",
+    "looks_like_ipa",
 ]
 
 
@@ -894,3 +895,56 @@ def convert_batch(
             yield stripped
         else:
             yield convert(stripped, _src, _dst)
+
+
+# ---------------------------------------------------------------------------
+# looks_like_ipa — heuristic notation detection
+# ---------------------------------------------------------------------------
+
+# Unicode blocks/characters that are distinctive to IPA transcription.  Basic
+# Latin letters (p, a, t, …) and the Greek letters IPA borrows (β, θ, χ) are
+# deliberately excluded: they are indistinguishable from ordinary Latin/Greek
+# text, so they carry no signal.
+_IPA_DISTINCTIVE_RANGES = (
+    (0x0250, 0x02AF),  # IPA Extensions
+    (0x02B0, 0x02FF),  # Spacing Modifier Letters (ˈ ˌ ː ˑ ʰ ʲ …)
+    (0x0363, 0x036F),  # Combining diacritics used in IPA
+    (0x1D00, 0x1D7F),  # Phonetic Extensions
+    (0x1D80, 0x1DBF),  # Phonetic Extensions Supplement
+)
+# Individual combining marks common in IPA (nasalisation, syllabicity, etc.).
+_IPA_DISTINCTIVE_CHARS = frozenset("̩̪̥̃̊͜͡")
+
+
+def _is_ipa_distinctive(ch: str) -> bool:
+    cp = ord(ch)
+    if ch in _IPA_DISTINCTIVE_CHARS:
+        return True
+    return any(lo <= cp <= hi for lo, hi in _IPA_DISTINCTIVE_RANGES)
+
+
+def looks_like_ipa(text: str) -> bool:
+    """Heuristically decide whether *text* is an IPA transcription.
+
+    Returns ``True`` when *text* contains at least one character that is
+    distinctive to IPA — from the IPA Extensions, Spacing Modifier Letters,
+    or Phonetic Extensions blocks, or an IPA combining diacritic (e.g.
+    ``ɑ``, ``ʃ``, ``ː``, ``ˈ``, nasalisation).
+
+    This is a positive-signal heuristic, not a classifier: a transcription
+    written only with characters IPA shares with the Latin alphabet
+    (``"pat"``, ``"bad"``) has no distinctive marker and returns ``False``.
+    It cannot be otherwise — those characters are the same codepoints as
+    ordinary text. Use it to *guard* against feeding IPA where orthography
+    is expected, not to prove a string is not IPA.
+
+    Examples
+    --------
+    >>> looks_like_ipa("pʰɑtʃ")
+    True
+    >>> looks_like_ipa("ˈhɛloʊ")
+    True
+    >>> looks_like_ipa("hello")
+    False
+    """
+    return any(_is_ipa_distinctive(ch) for ch in text)

{scriptconv-0.0.1a5 → scriptconv-0.0.1a6}/scriptconv/version.py

@@ -2,7 +2,7 @@
 VERSION_MAJOR = 0
 VERSION_MINOR = 0
 VERSION_BUILD = 1
-VERSION_ALPHA = 5
+VERSION_ALPHA = 6
 # END_VERSION_BLOCK
 
 VERSION_STR = f"{VERSION_MAJOR}.{VERSION_MINOR}.{VERSION_BUILD}"

{scriptconv-0.0.1a5 → scriptconv-0.0.1a6}/scriptconv.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
 Metadata-Version: 2.4
 Name: scriptconv
-Version: 0.0.1a5
+Version: 0.0.1a6
 Summary: Shared script-conversion core — ISO-15924 detection, IPA↔ARPA/X-SAMPA/Lexique, Buckwalter, Hangul→jamo
 License: Apache-2.0
 Project-URL: Homepage, https://github.com/TigreGotico/scriptconv
@@ -75,7 +75,7 @@ from scriptconv import (
     xsampa_to_ipa, ipa_to_xsampa,
     buckwalter_to_arabic, arabic_to_buckwalter,
     lexique_to_ipa, ipa_to_lexique,
-    kirshenbaum_to_ipa, ipa_to_kirshenbaum,
+    kirshenbaum_to_ipa, ipa_to_kirshenbaum, looks_like_ipa,
     decompose_hangul, hira_to_kana, kana_to_hira,
     convert, can_convert, convert_batch, Notation, NOTATION_INFO,
 )
@@ -107,6 +107,9 @@ ipa_to_lexique("vɛ̃")               # "v5"  (vin)
 kirshenbaum_to_ipa("S")             # "ʃ"   (espeak-ng ASCII-IPA)
 ipa_to_kirshenbaum("ŋ")             # "N"
 
+looks_like_ipa("pʰɑtʃ")             # True  (heuristic: has IPA-distinctive symbols)
+looks_like_ipa("hello")             # False
+
 decompose_hangul("국민")             # "ㄱㅜㄱㅁㅣㄴ"  (orthographic jamo, no assimilation)
 hira_to_kana("ひらがな")             # "ヒラガナ"
 kana_to_hira("カタカナ")             # "かたかな"
@@ -136,7 +139,7 @@ python -m scriptconv lang ko
 | Module | Contents |
 |--------|----------|
 | `scriptconv.scripts` | `Script` dataclass (with `script_type`), `SCRIPT_REGISTRY` (34 scripts), `detect_script`, `char_script`, `script_distribution`, `script_runs`, `base_direction`, `lang_to_script`, `script_to_langs`, `normalize_script_tag` |
-| `scriptconv.notation` | `Notation` enum, `NotationInfo`/`NOTATION_INFO` fidelity registry, `convert` facade, `can_convert` predicate, `convert_batch` generator, pair-wise converters (ARPABET ↔ IPA, X-SAMPA ↔ IPA, Buckwalter ↔ Arabic, Lexique ↔ IPA, Kirshenbaum ↔ IPA) |
+| `scriptconv.notation` | `Notation` enum, `NotationInfo`/`NOTATION_INFO` fidelity registry, `convert` facade, `can_convert` predicate, `convert_batch` generator, pair-wise converters (ARPABET ↔ IPA, X-SAMPA ↔ IPA, Buckwalter ↔ Arabic, Lexique ↔ IPA, Kirshenbaum ↔ IPA), `looks_like_ipa` detector |
 | `scriptconv.translit` | `decompose_hangul` (Hangul blocks → jamo, compatibility or conjoining), `hira_to_kana`/`kana_to_hira` — all orthographic only |
 
 ## Documentation

{scriptconv-0.0.1a5 → scriptconv-0.0.1a6}/tests/test_notation.py

@@ -1027,3 +1027,33 @@ def test_lossless_to_ipa_flag_backed_by_round_trip():
         if info.lossless_to_ipa and notation in samples:
             for ch in samples[notation]:
                 assert from_ipa[notation](to_ipa[notation](ch)) == ch, notation
+
+
+# ---------------------------------------------------------------------------
+# looks_like_ipa — heuristic notation detection
+# ---------------------------------------------------------------------------
+
+from scriptconv.notation import looks_like_ipa  # noqa: E402
+
+
+@pytest.mark.parametrize("text", [
+    "pʰɑtʃ",       # IPA Extensions + modifier
+    "ˈhɛloʊ",      # stress mark + IPA Extensions
+    "ɑ",           # single IPA Extensions char
+    "næ̃",          # combining nasalisation
+    "kat̪",         # combining dental diacritic
+])
+def test_looks_like_ipa_true(text):
+    assert looks_like_ipa(text) is True
+
+
+@pytest.mark.parametrize("text", [
+    "",            # empty
+    "hello",       # plain Latin — ambiguous, no distinctive marker
+    "att",         # overlaps Latin
+    "café",        # accented Latin is not IPA-distinctive
+    "θaβ",         # bare Greek letters carry no IPA signal
+    "مرحبا",       # Arabic
+])
+def test_looks_like_ipa_false(text):
+    assert looks_like_ipa(text) is False