feature: replace dictionary file with bz2-archived version

Author: unsektor

changelog.md

@@ -5,6 +5,11 @@ All notable changes to this project will be documented in this file.
 The format is based on [Keep a Changelog](https://keepachangelog.com/en/1.1.0/),
 and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0.html).
 
+## 0.1.7 — 2025-02-15
+### Changed
+
+- [x] replace dictionary file with bz2-archived version (decrease file size in 8.1 times)
+
 ## 0.1.6 — 2025-02-15
 ### Changed
 

setup.py

@@ -1,25 +1,41 @@
 import setuptools
+import os
+import bz2
 
 
 with open('readme.md') as fh:
     long_description = fh.read()
 
-setuptools.setup(
-    name='yoficator',
-    version='0.1.6',
-    description='A Russian text yoficator (ёфикатор)',
-    long_description=long_description,
-    long_description_content_type='text/markdown',
-    license='License :: OSI Approved :: MIT License',
-    packages=['yoficator'],
-    package_data={
-        'yoficator': ['_data/dictionary.ru_RU.txt'],
-    },
-    include_package_data=True,
-    classifiers=[
-        'Programming Language :: Python :: 3',
-        'License :: OSI Approved :: MIT License',
-        'Operating System :: OS Independent',
-    ],
-    python_requires='>=3',
-)
+
+def create_zip():
+    with open('yoficator/_data/dictionary.ru_RU.txt', 'rb') as source:
+        with bz2.BZ2File('yoficator/_data/dictionary.ru_RU.txt.bz2', 'w') as stream:
+            stream.write(source.read())
+
+
+try:
+    create_zip()
+    setuptools.setup(
+        name='yoficator',
+        version='0.1.7',
+        description='A Russian text yoficator (ёфикатор)',
+        long_description=long_description,
+        long_description_content_type='text/markdown',
+        license='License :: OSI Approved :: MIT License',
+        packages=['yoficator'],
+        package_data={
+            'yoficator': ['_data/dictionary.ru_RU.txt.bz2'],
+        },
+        include_package_data=True,
+        classifiers=[
+            'Programming Language :: Python :: 3',
+            'License :: OSI Approved :: MIT License',
+            'Operating System :: OS Independent',
+        ],
+        python_requires='>=3',
+    )
+finally:
+    try:
+        os.remove('yoficator/_data/dictionary.ru_RU.txt.bz2')
+    except FileNotFoundError:
+        pass

yoficator/__main__.py

@@ -2,6 +2,7 @@
 import os
 import sys
 import re
+import bz2
 
 #-------------------------------------------------------------------------#
 #
@@ -64,7 +65,7 @@
 
 if __name__ == '__main__':
     # TODO Better handle lowercase, uppercase
-    dictionary_file_path = os.path.abspath(os.path.dirname(__file__)) + '/_data/dictionary.ru_RU.txt'
+    dictionary_file_path = os.path.abspath(os.path.dirname(__file__)) + '/_data/dictionary.ru_RU.txt.bz2'
 
     if len(sys.argv) > 1:
         # Is the input a filename?
@@ -82,11 +83,10 @@
     # Splitter / tokenizer
     splitter = re.compile(r'(?P<word>[а-я]*е[а-я]*)|(?P<unknown>[^е]+\b)', re.IGNORECASE)
 
-    with open(dictionary_file_path) as stream:
+    with bz2.open(dictionary_file_path) as stream:
         for line in iter(stream):
-            if ':' in line:
-                key, value = line.split(':')
-                dictionary[key] = value.rstrip('\n')
+            key, value = line.decode('utf-8').split(':')
+            dictionary[key] = value.rstrip('\n')
 
     for token in splitter.finditer(text):
         word = token.group(0)