python-negar 1.4.3__tar.gz → 1.4.4__tar.gz

{python_negar-1.4.3 → python_negar-1.4.4}/Changelog.txt

@@ -1,3 +1,8 @@
+1.4.4 - 2026-10-31
+-- Fixed handling of Arabic words with diacritics (ایّام, بیمِ, بُستان, ناوِلْها): no wrong ZWNJ/separation
+-- Immutable words are now matched ignoring diacritics
+-- Added several additional immutable words.
+
 1.4.3 - 2025-01-28
 -- Save user defined immutable words under .config/python-negar
 

{python_negar-1.4.3 → python_negar-1.4.4}/PKG-INFO

@@ -1,6 +1,6 @@
-Metadata-Version: 2.1
+Metadata-Version: 2.4
 Name: python-negar
-Version: 1.4.3
+Version: 1.4.4
 Summary: Negar is a spell corrector and Persian text editor
 Home-page: http://github.com/shahinism/python-negar
 Author: Shahin Azad
@@ -18,6 +18,20 @@ Description-Content-Type: text/markdown
 License-File: LICENSE
 License-File: AUTHORS
 Requires-Dist: regex
+Dynamic: author
+Dynamic: author-email
+Dynamic: classifier
+Dynamic: description
+Dynamic: description-content-type
+Dynamic: home-page
+Dynamic: keywords
+Dynamic: license
+Dynamic: license-file
+Dynamic: maintainer
+Dynamic: maintainer-email
+Dynamic: requires-dist
+Dynamic: requires-python
+Dynamic: summary
 
 Negar
 ==========
@@ -31,6 +45,8 @@ Requires-Dist: regex
 
 Negar is an editor(=virastar in Persian) for Persian text. The project is initially inspired by [virastar](https://github.com/aziz/virastar/blob/master/lib/virastar.rb). Thank you [Aziz](https://github.com/aziz) for your great job.
 
+## [Demo View ![Streamlit App](https://static.streamlit.io/badges/streamlit_badge_black_white.svg)](https://negar-web.streamlit.app/)
+
 Installation
 ==============
 

{python_negar-1.4.3 → python_negar-1.4.4}/README.md

@@ -10,6 +10,8 @@
 
 Negar is an editor(=virastar in Persian) for Persian text. The project is initially inspired by [virastar](https://github.com/aziz/virastar/blob/master/lib/virastar.rb). Thank you [Aziz](https://github.com/aziz) for your great job.
 
+## [Demo View ![Streamlit App](https://static.streamlit.io/badges/streamlit_badge_black_white.svg)](https://negar-web.streamlit.app/)
+
 Installation
 ==============
 

{python_negar-1.4.3 → python_negar-1.4.4}/negar/constants.py

@@ -1,7 +1,7 @@
 import platform
 from pathlib import Path
 
-__version__ = "1.4.3"
+__version__ = "1.4.4"
 
 DATAFILE = Path(__file__).parent.absolute() / "data/immutable.words"
 APPDATA = "AppData/Roaming/" if platform.system() == "Windows" else ".config/"

{python_negar-1.4.3 → python_negar-1.4.4}/negar/data/immutable.words

@@ -1915,4 +1915,10 @@
 یادداشتتان
 یاریتان
 یاقوت‌نشان
-یاقوت‌نشان
+یاقوت‌نشان
+راستان
+بیارا
+بیائید
+بیایید
+عمان
+بیم

{python_negar-1.4.3 → python_negar-1.4.4}/negar/virastar.py

@@ -15,6 +15,9 @@ sys.path.append(
 from negar.constants import DATAFILE, INFO, URLREGX, USERFILE  # noqa: E402
 
 
+DIACRITICS = "\u064B-\u0652\u0670"  # to be used inside a character class
+
+
 class State(enum.Enum):
     save = 1
     restore = 2
@@ -131,11 +134,18 @@ class PersianEditor:
         """Remove immutable words temporarily and reapply them at the end of the process."""
         if state == State.save:
             self.immutable_words = {}
-            words = re.findall(r"\b[\w\u200c]+\b", self.text, re.UNICODE)
-            for i, word in enumerate(words):
-                if (word := word.strip()) in ImmutableWords.get():
-                    self.immutable_words[i] = word
-                    self.text = regex.sub(rf"\b{word}\b", rf"__IMM__#{i}__", self.text)
+
+            def protect(m):
+                # look up without diacritics (بُستان -> بستان) but keep them in the text
+                word = m.group()
+                if re.sub(f"[{DIACRITICS}]", "", word) not in ImmutableWords.get():
+                    return word
+                i = len(self.immutable_words)
+                self.immutable_words[i] = word
+                return f"__IMM__#{i}__"
+
+            # diacritics are part of the word, otherwise `re` cuts it at them
+            self.text = re.sub(rf"\b[\w\u200c{DIACRITICS}]+", protect, self.text)
         if state == State.restore:
             for imm_key, imm_word in self.immutable_words.items():
                 self.text = re.sub(f"__IMM__#{imm_key}__", imm_word, self.text)
@@ -228,7 +238,7 @@ class PersianEditor:
     def fix_suffix_spacing(self):
         """Put ZWNJ between a word and its suffix (*ha[ye] *tar[in])."""
         regx = re.compile(
-            r"""\s+
+            r"""(?<![\u064B-\u0652\u0670])\s+  # a diacritic before the suffix means an Arabic word
             (تر(ی(ن)?)?
             |[تمش]ان
             |ا[متش]
@@ -237,7 +247,7 @@ class PersianEditor:
             |شناس(ی)?
             |گذار(ی)?|گزار(ی)?
             |مند|ور|پور
-            )\b""",
+            )\b(?![\u064B-\u0652\u0670])""",  # a trailing diacritic (e.g. ایّام) means the word goes on
             re.VERBOSE,
         )
         self.text = re.sub(regx, r"‌\1", self.text)
@@ -261,7 +271,7 @@ class PersianEditor:
             else ""
         )
         regx = regex.compile(
-            rf"""\b(\S+?) # not-greedy fetch to handle some case like هایشان instead شان
+            rf"""\b(\S+?(?<![\u064B-\u0652\u0670])) # not-greedy fetch to handle some case like هایشان instead شان; no diacritic before the suffix
             ({exag}
             # تر(ی(ن)?)?
             # [تمش]ان|

{python_negar-1.4.3 → python_negar-1.4.4}/python_negar.egg-info/PKG-INFO

@@ -1,6 +1,6 @@
-Metadata-Version: 2.1
+Metadata-Version: 2.4
 Name: python-negar
-Version: 1.4.3
+Version: 1.4.4
 Summary: Negar is a spell corrector and Persian text editor
 Home-page: http://github.com/shahinism/python-negar
 Author: Shahin Azad
@@ -18,6 +18,20 @@ Description-Content-Type: text/markdown
 License-File: LICENSE
 License-File: AUTHORS
 Requires-Dist: regex
+Dynamic: author
+Dynamic: author-email
+Dynamic: classifier
+Dynamic: description
+Dynamic: description-content-type
+Dynamic: home-page
+Dynamic: keywords
+Dynamic: license
+Dynamic: license-file
+Dynamic: maintainer
+Dynamic: maintainer-email
+Dynamic: requires-dist
+Dynamic: requires-python
+Dynamic: summary
 
 Negar
 ==========
@@ -31,6 +45,8 @@ Requires-Dist: regex
 
 Negar is an editor(=virastar in Persian) for Persian text. The project is initially inspired by [virastar](https://github.com/aziz/virastar/blob/master/lib/virastar.rb). Thank you [Aziz](https://github.com/aziz) for your great job.
 
+## [Demo View ![Streamlit App](https://static.streamlit.io/badges/streamlit_badge_black_white.svg)](https://negar-web.streamlit.app/)
+
 Installation
 ==============
 

{python_negar-1.4.3 → python_negar-1.4.4}/tests/test_unittest.py

@@ -134,6 +134,26 @@ class Test(unittest.TestCase):
         self.input_ = "تنها ترین خدمتگزار (تنها ترین خدمتگزار)"
         self.output_ = "تنهاترین خدمتگزار (تنهاترین خدمتگزار)"
 
+    @_assertEqual("Suffix With Diacritics")
+    def test_suffix_with_diacritics(self):
+        self.input_ = "غم ایّام"
+        self.output_ = "غم ایّام"
+
+    @_assertEqual("Suffix With Diacritics")
+    def test_suffix_with_diacritics_ii(self):
+        self.input_ = "غمِ ایّام"
+        self.output_ = "غمِ ایّام"
+
+    @_assertEqual("Immutable Word With Diacritics")
+    def test_immutable_word_with_diacritics(self):
+        self.input_ = "بیمِ جان بُستان"
+        self.output_ = "بیمِ جان بُستان"
+
+    @_assertEqual("Arabic Word Before Plural Suffix")
+    def test_arabic_word_before_plural_suffix(self):
+        self.input_ = "ناوِلْها اَهْمِلْها بُستان"
+        self.output_ = "ناوِلْها اَهْمِلْها بُستان"
+
     @_assertEqual("Fix Suffix ZWNJ II")
     def test_fix_suffix_ZWNJ_II(self):
         self.input_ = "بیستم ماه میلادی"