Update Proficiency v0.5.10

custom_lemmas.py

@@ -215,7 +215,7 @@ def check_empty_kindle_gloss(self) -> None:
             if not package_name:
                 device_not_found_dialog(self)
                 return
-            custom_folder = custom_lemmas_folder(plugin_path)
+            custom_folder = custom_lemmas_folder(plugin_path, "en")
             if isinstance(package_name, str):
                 copy_klld_from_android(package_name, custom_folder)
             else:

data/deps.json

@@ -1,9 +1,9 @@
 {
   "cupy": "12.2.0",
   "lxml": "4.9.3",
-  "thinc-apple-ops": "0.1.3",
+  "thinc-apple-ops": "0.1.4",
   "torch": "2.0.1",
-  "rapidfuzz": "3.3.0",
+  "rapidfuzz": "3.3.1",
   "spacy_cpu_model": "3.6.0",
   "spacy_trf_model": "3.6.1"
 }

deps.py

@@ -1,9 +1,8 @@
 #!/usr/bin/env python3
 
+import bz2
 import platform
 import shutil
-import tarfile
-from io import BytesIO
 from pathlib import Path
 from typing import Any
 from urllib.request import urlopen
@@ -12,16 +11,13 @@
 
 from .utils import (
     PROFICIENCY_RELEASE_URL,
-    PROFICIENCY_VERSION,
     Prefs,
-    custom_lemmas_folder,
     get_plugin_path,
     get_wiktionary_klld_path,
     kindle_db_path,
     load_plugin_json,
     mac_bin_path,
     run_subprocess,
-    use_kindle_ww_db,
     wiktionary_db_path,
 )
 
@@ -173,25 +169,18 @@ def download_word_wise_file(
     else:
         db_path = wiktionary_db_path(plugin_path, lemma_lang, gloss_lang)
 
-    extract_folder = custom_lemmas_folder(get_plugin_path())
     if not db_path.exists():
-        filename = f"wiktionary_{lemma_lang}_{gloss_lang}_v{PROFICIENCY_VERSION}.bz2"
-        if is_kindle and use_kindle_ww_db(lemma_lang, prefs):
-            filename = f"kindle_en_en_v{PROFICIENCY_VERSION}.bz2"
-        url = f"{PROFICIENCY_RELEASE_URL}/{filename}"
-        download_and_extract(url, extract_folder)
+        bz2_filename = db_path.with_suffix(db_path.suffix + ".bz2").name
+        download_and_extract(f"{PROFICIENCY_RELEASE_URL}/{bz2_filename}", db_path)
 
     if is_kindle:
         klld_path = get_wiktionary_klld_path(plugin_path, lemma_lang, gloss_lang)
         if not klld_path.exists():
-            url = (
-                PROFICIENCY_RELEASE_URL
-                + f"/kll.{lemma_lang}.{gloss_lang}_v{PROFICIENCY_VERSION}.klld.bz2"
-            )
-            download_and_extract(url, extract_folder)
+            bz2_filename = klld_path.with_suffix(klld_path.suffix + ".bz2").name
+            download_and_extract(f"{PROFICIENCY_RELEASE_URL}/{bz2_filename}", klld_path)
 
 
-def download_and_extract(url: str, extract_folder: Path) -> None:
-    with urlopen(url) as r:
-        with tarfile.open(fileobj=BytesIO(r.read())) as tar:
-            tar.extractall(extract_folder)
+def download_and_extract(url: str, extract_path: Path) -> None:
+    extract_path.parent.mkdir(parents=True, exist_ok=True)
+    with urlopen(url) as r, bz2.open(r) as bz2_f, extract_path.open("wb") as f:
+        shutil.copyfileobj(bz2_f, f)

dump_lemmas.py

@@ -34,9 +34,8 @@ def spacy_doc_path(
     if is_kindle and not use_kindle_ww_db(lemma_lang, prefs):
         is_kindle = False
     py_version = ".".join(platform.python_version_tuple()[:2])
-    path = custom_lemmas_folder(plugin_path).joinpath(
-        f"{lemma_lang}/{spacy_model}_"
-        f"{'kindle' if is_kindle else 'wiktionary'}"
+    path = custom_lemmas_folder(plugin_path, lemma_lang).joinpath(
+        f"{spacy_model}_{'kindle' if is_kindle else 'wiktionary'}"
         f"_{gloss_lang}_{model_version}_{py_version}"
     )
     if prefs["use_pos"]:

utils.py

@@ -11,7 +11,7 @@
 from typing import Any, TypedDict
 
 CJK_LANGS = ["zh", "ja", "ko"]
-PROFICIENCY_VERSION = "0.5.8"
+PROFICIENCY_VERSION = "0.5.10"
 PROFICIENCY_RELEASE_URL = (
     f"https://github.com/xxyzz/Proficiency/releases/download/v{PROFICIENCY_VERSION}"
 )
@@ -98,8 +98,8 @@ def get_plugin_path() -> Path:
     return Path(config_dir) / "plugins/WordDumb.zip"
 
 
-def custom_lemmas_folder(plugin_path: Path) -> Path:
-    return plugin_path.parent / "worddumb-lemmas"
+def custom_lemmas_folder(plugin_path: Path, lemma_lang: str) -> Path:
+    return plugin_path.parent / "worddumb-lemmas" / lemma_lang
 
 
 def use_kindle_ww_db(lemma_lang: str, prefs: Prefs) -> bool:
@@ -113,8 +113,7 @@ def use_kindle_ww_db(lemma_lang: str, prefs: Prefs) -> bool:
 def kindle_db_path(plugin_path: Path, lemma_lang: str, prefs: Prefs) -> Path:
     if use_kindle_ww_db(lemma_lang, prefs):
         return (
-            custom_lemmas_folder(plugin_path)
-            / lemma_lang
+            custom_lemmas_folder(plugin_path, lemma_lang)
             / f"kindle_en_en_v{PROFICIENCY_MAJOR_VERSION}.db"
         )
     else:
@@ -123,14 +122,13 @@ def kindle_db_path(plugin_path: Path, lemma_lang: str, prefs: Prefs) -> Path:
 
 def wiktionary_db_path(plugin_path: Path, lemma_lang: str, gloss_lang: str) -> Path:
     return (
-        custom_lemmas_folder(plugin_path)
-        / lemma_lang
+        custom_lemmas_folder(plugin_path, lemma_lang)
         / f"wiktionary_{lemma_lang}_{gloss_lang}_v{PROFICIENCY_MAJOR_VERSION}.db"
     )
 
 
 def get_klld_path(plugin_path: Path) -> Path | None:
-    custom_folder = custom_lemmas_folder(plugin_path)
+    custom_folder = custom_lemmas_folder(plugin_path, "")
     for path in custom_folder.glob("*.en.klld"):
         return path
     for path in custom_folder.glob("*.en.db"):
@@ -142,8 +140,7 @@ def get_wiktionary_klld_path(
     plugin_path: Path, lemma_lang: str, gloss_lang: str
 ) -> Path:
     return (
-        custom_lemmas_folder(plugin_path)
-        / lemma_lang
+        custom_lemmas_folder(plugin_path, lemma_lang)
         / f"kll.{lemma_lang}.{gloss_lang}_v{PROFICIENCY_MAJOR_VERSION}.klld"
     )