Update app.py

app.py

@@ -1,6 +1,4 @@
 import gradio as gr
-from datasets import load_dataset
-
 import random
 import time
 import requests
@@ -11,8 +9,7 @@ import traceback
 from base64 import b64decode,b64encode
 from io import BytesIO
 
-word_list_dataset = load_dataset("stabilityai/word-list", data_files="list.txt")
-word_list = word_list_dataset["train"]['text']
+from profanity_filter import ProfanityFilter
 
 
     
@@ -21,6 +18,27 @@ with gr.Blocks(theme="darkdefault") as demo:
     def welcome(name):
         return f"Welcome to AIXRPL.com Minter, {name}!"
 
+
+    def profanityCheck(prompt):
+            prompt = prompt.replace('+',' ').replace('|',' ')
+            nlp = spacy.load('en')
+            profanity_filter = ProfanityFilter(nlps={'en': nlp})  # reuse spacy Language (optional)
+            nlp.add_pipe(profanity_filter.spacy_component, last=True)
+            pf = ProfanityFilter()
+            pf.censor_whole_words = False
+    
+            doc = nlp(prompt)
+            if doc._.is_profane:
+                print(doc)
+                return True
+            for token in doc:
+                print(token)
+                if token._.is_profane:
+                    return True
+            
+            return False
+
+
     def inference(_prompt,_token):
         try:
             from PIL import Image
@@ -28,8 +46,7 @@ with gr.Blocks(theme="darkdefault") as demo:
             import os
             print(_prompt,_token)
             
-            for filter in word_list:
-                if re.search(rf"\b{filter}\b", _prompt):
+            if profanityCheck(_prompt):
                     return '','unsafe','','',''        
             
             r = requests.post(url='https://xcyborgart-app.herokuapp.com/create',data={"prompt":_prompt,"token":_token})