Fixed issue #19; *hayeshan$

shahinism · Mar 31, 2022 · 5d9d54c · 5d9d54c
1 parent 800dc45
commit 5d9d54c
Show file tree

Hide file tree

Showing 4 changed files with 10 additions and 6 deletions.
diff --git a/Changelog.txt b/Changelog.txt
@@ -1,3 +1,6 @@
+0.9.3 - 2022-03-31
+-- Fixed issue #19; some cases than ended with *hayeshan (کتابهایشان)
+
 0.9.2 - 2022-03-30
 -- Fixed some issues with ^bi*
 -- Fixed the issue of extra ZWNJ beside brackets

diff --git a/negar/constants.py b/negar/constants.py
@@ -1,6 +1,6 @@
 from pathlib import Path
 
-__version__ = "0.9.2"
+__version__ = "0.9.3"
 
 LOGO = (Path(__file__).parent.absolute()/"logo.png").as_posix()
 DATAFILE = Path(__file__).parent.absolute()/"data/untouchable.dat"

diff --git a/negar/data/untouchable.dat b/negar/data/untouchable.dat
@@ -944,6 +944,7 @@
 کم‌بها
 کمتر
 کنتر
+کهتر
 کهکشان
 کوپنها
 کودکستان

diff --git a/negar/virastar.py b/negar/virastar.py
@@ -150,8 +150,8 @@ def fix_suffix_spacing(self):
         regex = re.compile(
             r"""\s+
             (تر(ی(ن)?)?
-            |ها(ی(ی)?)?
-            |[تمش]ان)
+            |[تمش]ان
+            |ها(ی(ی|ت|م|ش|تان|شان)?)?)
             \b""",
             re.VERBOSE
         )
@@ -164,10 +164,10 @@ def fix_suffix_spacing(self):
     def fix_suffix_separate(self):
         """Puts ZWNJ between a word with its suffix (haye, ...)"""
         regex = re.compile(
-            r"""(\S+)
+            r"""(\S+?) # not-greedy fetch to handle some case like هایشان instead شان
             (تر(ی(ن)?)?
-            |ها(ی(ی)?)?|
-            [تمش]ان)\b""",
+            |[تمش]ان
+            |ها(ی(ی|ت|م|ش|تان|شان)?)?)\b""",
             re.VERBOSE
         )
         wlist = self.text.split(" ")