ihub-aether-lang-pack 1.0.0__tar.gz

ihub_aether_lang_pack-1.0.0/PKG-INFO

@@ -0,0 +1,90 @@
+Metadata-Version: 2.4
+Name: ihub-aether-lang-pack
+Version: 1.0.0
+Summary: Complete ISO 639 language registry: all 7,927 registered languages with official English and French names, CLDR native names, status, scope, macrolanguage links, legacy aliases, and a defined fallback chain for every code.
+Author: iHub Aether
+License-Expression: MIT
+Project-URL: Homepage, https://4m1wp03rd34yk.aiforce.cloud/app/app_17e3evgsy6q
+Keywords: iso639,language,i18n,locale,langpack,iso-639-3,internationalization
+Classifier: Development Status :: 5 - Production/Stable
+Classifier: Intended Audience :: Developers
+Classifier: Programming Language :: Python :: 3
+Classifier: Topic :: Software Development :: Internationalization
+Requires-Python: >=3.9
+Description-Content-Type: text/markdown
+
+# iHub Aether Lang Pack
+
+The complete ISO 639 language registry as a zero-dependency Python package.
+
+Every registered language code (ISO 639-1 / 639-2 / 639-3) with:
+
+- official English and French names (ISO / IANA)
+- native names from Unicode CLDR where they exist
+- status: living / extinct / historical / constructed / special
+- scope: individual / macrolanguage / special
+- macrolanguage links (e.g. `cmn` -> `zh`)
+- legacy aliases (e.g. `iw` -> `he`, `in` -> `id`, `sh` -> `sr`)
+- a defined fallback chain on **every** code, so no lookup ever breaks
+
+## Install
+
+```bash
+pip install ihub-aether-lang-pack
+```
+
+## Usage
+
+```python
+from langpack import langpack
+
+len(langpack)                       # 7927 registered languages
+langpack.counts["living"]           # 7082 spoken (living) tongues
+
+langpack.get("zh")
+# {'iso6393': 'zho', 'iso6391': 'zh', 'name': 'Chinese', 'native': '中文', ...}
+
+langpack.resolve("iw")              # 'he'  (legacy alias -> canonical)
+langpack.resolve("zh-Hans")         # 'zh'  (BCP-47 tag -> base code)
+
+langpack.fallback_chain("yue")      # ['yue', 'zh', 'en']
+langpack.fallback("cmn")            # 'zh'
+
+langpack.search("cantonese")        # [{... 'yue' ...}]
+langpack.native_name("ja")          # '日本語'
+```
+
+Command line:
+
+```bash
+python -m langpack          # summary
+python -m langpack chinese  # search
+```
+
+## Fallback rule
+
+- non-living codes -> `en`
+- macrolanguages -> `en`
+- individual under a macrolanguage -> that macrolanguage -> `en`
+- everything else -> `en`
+
+The chain is always `[code, fallback, en]` (deduplicated).
+
+## Coverage and dialects
+
+The registry contains all **7,927 codes registered by ISO 639**, which includes
+the local varieties that have official codes: Chinese "dialects" (Mandarin,
+Cantonese, Wu, Hakka, Min Nan...), Arabic regional varieties, German dialects
+and so on. Sub-dialects that ISO does not register cannot be in any real pack;
+script variants (zh-Hans / zh-Hant) are handled on top via BCP-47 tags, which
+`resolve()` already accepts.
+
+## Data sources
+
+- ISO 639-3 official code tables (SIL International)
+- IANA Language Subtag Registry
+- Unicode CLDR language names
+
+## License
+
+MIT (code). Data compiled from the public ISO 639 / IANA / CLDR registries.

ihub_aether_lang_pack-1.0.0/README.md

@@ -0,0 +1,75 @@
+# iHub Aether Lang Pack
+
+The complete ISO 639 language registry as a zero-dependency Python package.
+
+Every registered language code (ISO 639-1 / 639-2 / 639-3) with:
+
+- official English and French names (ISO / IANA)
+- native names from Unicode CLDR where they exist
+- status: living / extinct / historical / constructed / special
+- scope: individual / macrolanguage / special
+- macrolanguage links (e.g. `cmn` -> `zh`)
+- legacy aliases (e.g. `iw` -> `he`, `in` -> `id`, `sh` -> `sr`)
+- a defined fallback chain on **every** code, so no lookup ever breaks
+
+## Install
+
+```bash
+pip install ihub-aether-lang-pack
+```
+
+## Usage
+
+```python
+from langpack import langpack
+
+len(langpack)                       # 7927 registered languages
+langpack.counts["living"]           # 7082 spoken (living) tongues
+
+langpack.get("zh")
+# {'iso6393': 'zho', 'iso6391': 'zh', 'name': 'Chinese', 'native': '中文', ...}
+
+langpack.resolve("iw")              # 'he'  (legacy alias -> canonical)
+langpack.resolve("zh-Hans")         # 'zh'  (BCP-47 tag -> base code)
+
+langpack.fallback_chain("yue")      # ['yue', 'zh', 'en']
+langpack.fallback("cmn")            # 'zh'
+
+langpack.search("cantonese")        # [{... 'yue' ...}]
+langpack.native_name("ja")          # '日本語'
+```
+
+Command line:
+
+```bash
+python -m langpack          # summary
+python -m langpack chinese  # search
+```
+
+## Fallback rule
+
+- non-living codes -> `en`
+- macrolanguages -> `en`
+- individual under a macrolanguage -> that macrolanguage -> `en`
+- everything else -> `en`
+
+The chain is always `[code, fallback, en]` (deduplicated).
+
+## Coverage and dialects
+
+The registry contains all **7,927 codes registered by ISO 639**, which includes
+the local varieties that have official codes: Chinese "dialects" (Mandarin,
+Cantonese, Wu, Hakka, Min Nan...), Arabic regional varieties, German dialects
+and so on. Sub-dialects that ISO does not register cannot be in any real pack;
+script variants (zh-Hans / zh-Hant) are handled on top via BCP-47 tags, which
+`resolve()` already accepts.
+
+## Data sources
+
+- ISO 639-3 official code tables (SIL International)
+- IANA Language Subtag Registry
+- Unicode CLDR language names
+
+## License
+
+MIT (code). Data compiled from the public ISO 639 / IANA / CLDR registries.

ihub_aether_lang_pack-1.0.0/pyproject.toml

@@ -0,0 +1,30 @@
+[build-system]
+requires = ["setuptools>=68"]
+build-backend = "setuptools.build_meta"
+
+[project]
+name = "ihub-aether-lang-pack"
+version = "1.0.0"
+description = "Complete ISO 639 language registry: all 7,927 registered languages with official English and French names, CLDR native names, status, scope, macrolanguage links, legacy aliases, and a defined fallback chain for every code."
+readme = "README.md"
+requires-python = ">=3.9"
+license = "MIT"
+authors = [{ name = "iHub Aether" }]
+keywords = ["iso639", "language", "i18n", "locale", "langpack", "iso-639-3", "internationalization"]
+classifiers = [
+    "Development Status :: 5 - Production/Stable",
+    "Intended Audience :: Developers",
+    "Programming Language :: Python :: 3",
+    "Topic :: Software Development :: Internationalization",
+]
+dependencies = []
+
+[project.urls]
+Homepage = "https://4m1wp03rd34yk.aiforce.cloud/app/app_17e3evgsy6q"
+
+[tool.setuptools]
+package-dir = { "" = "src" }
+packages = ["langpack", "langpack._data"]
+
+[tool.setuptools.package-data]
+langpack = ["_data/languages.json"]

ihub_aether_lang_pack-1.0.0/setup.cfg

@@ -0,0 +1,4 @@
+[egg_info]
+tag_build = 
+tag_date = 0
+

ihub_aether_lang_pack-1.0.0/src/ihub_aether_lang_pack.egg-info/PKG-INFO

@@ -0,0 +1,90 @@
+Metadata-Version: 2.4
+Name: ihub-aether-lang-pack
+Version: 1.0.0
+Summary: Complete ISO 639 language registry: all 7,927 registered languages with official English and French names, CLDR native names, status, scope, macrolanguage links, legacy aliases, and a defined fallback chain for every code.
+Author: iHub Aether
+License-Expression: MIT
+Project-URL: Homepage, https://4m1wp03rd34yk.aiforce.cloud/app/app_17e3evgsy6q
+Keywords: iso639,language,i18n,locale,langpack,iso-639-3,internationalization
+Classifier: Development Status :: 5 - Production/Stable
+Classifier: Intended Audience :: Developers
+Classifier: Programming Language :: Python :: 3
+Classifier: Topic :: Software Development :: Internationalization
+Requires-Python: >=3.9
+Description-Content-Type: text/markdown
+
+# iHub Aether Lang Pack
+
+The complete ISO 639 language registry as a zero-dependency Python package.
+
+Every registered language code (ISO 639-1 / 639-2 / 639-3) with:
+
+- official English and French names (ISO / IANA)
+- native names from Unicode CLDR where they exist
+- status: living / extinct / historical / constructed / special
+- scope: individual / macrolanguage / special
+- macrolanguage links (e.g. `cmn` -> `zh`)
+- legacy aliases (e.g. `iw` -> `he`, `in` -> `id`, `sh` -> `sr`)
+- a defined fallback chain on **every** code, so no lookup ever breaks
+
+## Install
+
+```bash
+pip install ihub-aether-lang-pack
+```
+
+## Usage
+
+```python
+from langpack import langpack
+
+len(langpack)                       # 7927 registered languages
+langpack.counts["living"]           # 7082 spoken (living) tongues
+
+langpack.get("zh")
+# {'iso6393': 'zho', 'iso6391': 'zh', 'name': 'Chinese', 'native': '中文', ...}
+
+langpack.resolve("iw")              # 'he'  (legacy alias -> canonical)
+langpack.resolve("zh-Hans")         # 'zh'  (BCP-47 tag -> base code)
+
+langpack.fallback_chain("yue")      # ['yue', 'zh', 'en']
+langpack.fallback("cmn")            # 'zh'
+
+langpack.search("cantonese")        # [{... 'yue' ...}]
+langpack.native_name("ja")          # '日本語'
+```
+
+Command line:
+
+```bash
+python -m langpack          # summary
+python -m langpack chinese  # search
+```
+
+## Fallback rule
+
+- non-living codes -> `en`
+- macrolanguages -> `en`
+- individual under a macrolanguage -> that macrolanguage -> `en`
+- everything else -> `en`
+
+The chain is always `[code, fallback, en]` (deduplicated).
+
+## Coverage and dialects
+
+The registry contains all **7,927 codes registered by ISO 639**, which includes
+the local varieties that have official codes: Chinese "dialects" (Mandarin,
+Cantonese, Wu, Hakka, Min Nan...), Arabic regional varieties, German dialects
+and so on. Sub-dialects that ISO does not register cannot be in any real pack;
+script variants (zh-Hans / zh-Hant) are handled on top via BCP-47 tags, which
+`resolve()` already accepts.
+
+## Data sources
+
+- ISO 639-3 official code tables (SIL International)
+- IANA Language Subtag Registry
+- Unicode CLDR language names
+
+## License
+
+MIT (code). Data compiled from the public ISO 639 / IANA / CLDR registries.

ihub_aether_lang_pack-1.0.0/src/ihub_aether_lang_pack.egg-info/SOURCES.txt

@@ -0,0 +1,9 @@
+README.md
+pyproject.toml
+src/ihub_aether_lang_pack.egg-info/PKG-INFO
+src/ihub_aether_lang_pack.egg-info/SOURCES.txt
+src/ihub_aether_lang_pack.egg-info/dependency_links.txt
+src/ihub_aether_lang_pack.egg-info/top_level.txt
+src/langpack/__init__.py
+src/langpack/__main__.py
+src/langpack/_data/languages.json

ihub_aether_lang_pack-1.0.0/src/ihub_aether_lang_pack.egg-info/dependency_links.txt

@@ -0,0 +1 @@
+

ihub_aether_lang_pack-1.0.0/src/ihub_aether_lang_pack.egg-info/top_level.txt

@@ -0,0 +1 @@
+langpack

ihub_aether_lang_pack-1.0.0/src/langpack/__init__.py

@@ -0,0 +1,148 @@
+"""iHub Aether Lang Pack - complete ISO 639 language registry.
+
+Every ISO 639-1/2/3 registered code, with official English and French names,
+Unicode CLDR native names, status, scope, macrolanguage links, legacy aliases,
+and a defined fallback chain so no language lookup ever breaks.
+"""
+
+from __future__ import annotations
+
+import json
+from importlib import resources
+from typing import Dict, Iterator, List, Optional
+
+FORMAT = "ihub-aether-lang-pack"
+FORMAT_VERSION = 1
+VERSION = "1.0.0"
+DEFAULT_FALLBACK = "en"
+
+
+def _load_pack() -> Dict:
+    raw = (resources.files("langpack") / "_data" / "languages.json").read_bytes()
+    return json.loads(raw.decode("utf-8"))
+
+
+class LanguagePack:
+    """Read-only view over the full ISO 639 registry with alias resolution
+    and per-code fallback chains."""
+
+    def __init__(self) -> None:
+        self._pack = _load_pack()
+        self._langs = self._pack["languages"]
+        self._aliases = self._pack["aliases"]
+
+    # -- metadata -----------------------------------------------------
+    @property
+    def counts(self) -> Dict:
+        return dict(self._pack["counts"])
+
+    @property
+    def aliases(self) -> Dict:
+        return dict(self._aliases)
+
+    @property
+    def sources(self) -> List[str]:
+        return list(self._pack["sources"])
+
+    # -- lookup -------------------------------------------------------
+    def resolve(self, code: str) -> Optional[str]:
+        """Return the canonical registry code for any input.
+
+        Accepts exact codes, legacy aliases (iw -> he, sh -> sr, ...),
+        and BCP-47 tags (zh-Hans -> zh, en-US -> en). Returns None when
+        the code is not registered.
+        """
+        c = str(code).strip().lower()
+        base = c.split("-")[0]
+        base = self._aliases.get(base, base)
+        if base in self._langs:
+            return base
+        if c in self._langs:
+            return c
+        return None
+
+    def get(self, code: str) -> Optional[Dict]:
+        """Return the full entry for a code, or None when not registered."""
+        canonical = self.resolve(code)
+        if canonical is None:
+            return None
+        return dict(self._langs[canonical])
+
+    def __contains__(self, code: str) -> bool:
+        return self.resolve(code) is not None
+
+    def __getitem__(self, code: str) -> Dict:
+        entry = self.get(code)
+        if entry is None:
+            raise KeyError(code)
+        return entry
+
+    def __iter__(self) -> Iterator[str]:
+        return iter(self._langs)
+
+    def __len__(self) -> int:
+        return len(self._langs)
+
+    # -- names ---------------------------------------------------------
+    def name(self, code: str, field: str = "name") -> Optional[str]:
+        """English name by default; pass 'native' or 'nameFr' for those."""
+        entry = self.get(code)
+        if entry is None:
+            return None
+        return entry.get(field) or entry.get("name")
+
+    def native_name(self, code: str) -> Optional[str]:
+        return self.name(code, "native")
+
+    def french_name(self, code: str) -> Optional[str]:
+        return self.name(code, "nameFr")
+
+    # -- fallback ------------------------------------------------------
+    def fallback_chain(self, code: str) -> List[str]:
+        """Defined fallback chain, always ending in 'en'."""
+        canonical = self.resolve(code)
+        if canonical is None:
+            return [DEFAULT_FALLBACK]
+        return list(
+            self._langs[canonical].get("fallbackChain")
+            or [canonical, DEFAULT_FALLBACK]
+        )
+
+    def fallback(self, code: str) -> str:
+        chain = self.fallback_chain(code)
+        return chain[1] if len(chain) > 1 else chain[0]
+
+    # -- iteration / search -------------------------------------------
+    def all(self) -> List[Dict]:
+        return [dict(e) for e in self._langs.values()]
+
+    def living(self) -> List[Dict]:
+        return [dict(e) for e in self._langs.values() if e.get("spoken")]
+
+    def search(self, query: str, limit: int = 50) -> List[Dict]:
+        q = str(query).strip().lower()
+        if not q:
+            return []
+        out = []
+        for code, e in self._langs.items():
+            hay = " ".join(
+                str(e.get(k) or "")
+                for k in ("iso6393", "iso6391", "name", "nameFr", "native", "macro", "fallback")
+            ).lower()
+            if q in code or q in hay:
+                out.append(dict(e))
+                if len(out) >= limit:
+                    break
+        return out
+
+
+langpack = LanguagePack()
+
+__all__ = [
+    "LanguagePack",
+    "langpack",
+    "FORMAT",
+    "FORMAT_VERSION",
+    "VERSION",
+    "DEFAULT_FALLBACK",
+]

ihub_aether_lang_pack-1.0.0/src/langpack/__main__.py

@@ -0,0 +1,31 @@
+"""Command line interface: python -m langpack [query]"""
+
+import sys
+
+from langpack import langpack
+
+
+def main(argv):
+    if not argv:
+        c = langpack.counts
+        print("iHub Aether Lang Pack - %d registered languages" % c["total"])
+        print(
+            "spoken: %d | native names: %d | aliases: %d"
+            % (c["living"], c["withNativeName"], c["aliases"])
+        )
+        return 0
+    q = argv[0]
+    res = langpack.search(q, limit=20)
+    if not res:
+        print("no match for %r" % q)
+        return 1
+    for e in res:
+        print(
+            "%-6s %-4s %-14s %s"
+            % (e.get("iso6393", ""), e.get("iso6391", "") or "", e.get("name", ""), e.get("native", "") or "")
+        )
+    return 0
+
+
+if __name__ == "__main__":
+    sys.exit(main(sys.argv[1:]))