Merge pull request #21 from AdolfMacro/master

Added function for Remove punctuation characters from text .

PersianSwear.py

@@ -4,18 +4,27 @@
 created date : 7 October, 2021
 updated date : 11 October, 2021
 """
+import json
+from string import punctuation
 class PersianSwear(object):
 	def __init__(self):
 		self.data = json.load(open('data.json'))
 
+	#Rmove punctuation characters from text
 	# return string
-	def filter_words(self, text, symbol="*"):
+	def ignoreSY(self, text):
+		for i in punctuation:
+				if i in text:
+					text=text.replace(i,'')
+		return text
+	# return string
+	def filter_words(self, text , symbol="*" , ignoreOT=False):
 		if(self.is_empty()):
 			return text
 
 		text = text.split()
 		for i in range(len(text)):
-			if text[i] in self.data['word']:
+			if text[i] in self.data['word'] or (ignoreOT and self.ignoreSY(text[i]) in self.data['word']):
 				text[i] = symbol
 
 		return " ".join(text)
@@ -35,12 +44,16 @@ def remove_word(self, text):
 		self.data['word'].remove(text)	
 
 	# return boolean
-	def is_bad(self, text):
+	def is_bad(self, text , ignoreOT=False):
+		if ignoreOT:
+			text=self.ignoreSY(text)
 		text=text.replace("\u200c","")
 		return text in self.data['word']
 
 	# return boolean
-	def has_swear(self, text):
+	def has_swear(self, text , ignoreOT=False):
+		if ignoreOT:
+			text=self.ignoreSY(text)
 		text=text.replace("\u200c","")
 		if(self.is_empty()):
 			return text
@@ -54,4 +67,4 @@ def has_swear(self, text):
 
 	# return string
 	def tostring(self):
-		return ' - '.join(self.data['word'])
+		return ' - '.join(self.data['word'])

README.md

@@ -115,24 +115,39 @@ echo $persianswear->tostring(); // show all swear words
 ```
 persianswear = PersianSwear()
 
-print(persianswear.is_bad('خر')) # True
+print(persianswear.is_bad(,'خر',ignoreOT=False )) # True
 
-print(persianswear.is_bad('امروز')) # False
+print(persianswear.is_bad('امروز',ignoreOT=False )) # False
 
-print(persianswear.is_bad('چرت و پرت')) # False
+print(persianswear.is_bad('چرت و پرت',ignoreOT=False )) # False
 
 persianswear.add_word('چرت و پرت')
-print(persianswear.is_bad('چرت و پرت')) # True
+print(persianswear.is_bad('چرت و پرت' , ignoreOT=False )) # True
 
-print(persianswear.has_swear('تو دوست من هستی')) # False
+print(persianswear.has_swear('تو دوست من هستی' , ignoreOT=False )) # False
 
-print(persianswear.has_swear('تو هیز هستی')) # True
+print(persianswear.has_swear('تو هیز هستی' , ignoreOT=False )) # True
 
-print(persianswear.filter_words('تو دوست من هستی')) # تو دوست من هستی
+print(persianswear.filter_words('تو دوست من هستی' , ignoreOT=False )) # تو دوست من هستی
 
-print(persianswear.filter_words('تو هیز هستی')) # تو * هستی
+print(persianswear.filter_words('تو هیز هستی' , ignoreOT=False )) # تو * هستی
 
-print(persianswear.filter_words('تو هیز هستی', '&')) # تو & هستی
+print(persianswear.filter_words('تو هیز هستی', '&' , ignoreOT=False )) # تو & هستی
+
+
+print(persianswear.is_bad('خ.ر' , ignoreOT=True )) # True
+
+print(persianswear.is_bad( 'ام.روز' , ignoreOT=True )) # False
+
+print(persianswear.has_swear('تو دو.ست من هستی' , ignoreOT=True )) # False
+
+print(persianswear.has_swear('تو اسک.ل هستی' , ignoreOT=True )) # True
+
+print(persianswear.filter_words('تو دو.ست من هستی',ignoreOT=True )) # تو دو.ست من هستی
+
+print(persianswear.filter_words('تو هی.ز هستی',ignoreOT=True )) # تو * هستی
+
+print(persianswear.filter_words('تو هی.ز هس.تی' , ignoreOT=True )) # تو * هس.تی
 
 print(persianswear.tostring()) # show all swear words
 ```