mirror of
https://github.com/Paillat-dev/Botator.git
synced 2026-01-02 09:16:19 +00:00
Added some moderation features
This commit is contained in:
Paillat
@@ -1,7 +1,21 @@
|
||||
from googleapiclient import discovery
|
||||
from config import perspective_api_key
|
||||
import json
|
||||
import re
|
||||
toxicity_names = ["toxicity", "severe_toxicity", "identity_attack", "insult", "profanity", "threat", "sexually_explicit", "flirtation", "obscene", "spam"]
|
||||
toxicity_definitions = [
|
||||
"A rude, disrespectful, or unreasonable message that is likely to make people leave a discussion.",
|
||||
"A very hateful, aggressive, disrespectful message or otherwise very likely to make a user leave a discussion or give up on sharing their perspective. This attribute is much less sensitive to more mild forms of toxicity, such as messages that include positive uses of curse words.",
|
||||
"Negative or hateful messages targeting someone because of their identity.",
|
||||
"Insulting, inflammatory, or negative messages towards a person or a group of people.",
|
||||
"Swear words, curse words, or other obscene or profane language.",
|
||||
"Describes an intention to inflict pain, injury, or violence against an individual or group.",
|
||||
"Contains references to sexual acts, body parts, or other lewd content. \n **English only**",
|
||||
"Pickup lines, complimenting appearance, subtle sexual innuendos, etc. \n **English only**",
|
||||
"Obscene or vulgar language such as cursing. \n **English only**",
|
||||
"Irrelevant and unsolicited commercial content. \n **English only**"
|
||||
]
|
||||
|
||||
|
||||
|
||||
client = discovery.build("commentanalyzer",
|
||||
"v1alpha1",
|
||||
@@ -12,11 +26,20 @@ client = discovery.build("commentanalyzer",
|
||||
|
||||
analyze_request = {
|
||||
'comment': {'text': ''}, # The text to analyze
|
||||
'requestedAttributes': {'TOXICITY': {}}, # Requested attributes
|
||||
#we will analyze the text in english, french & italian
|
||||
'languages': ['en', 'fr', 'it'],
|
||||
#we will ask the following attributes to google: TOXICITY, SEVERE_TOXICITY, IDENTITY_ATTACK, INSULT, PRPFANITY, THREAT, SEXUALLY_EXPLICIT, FLIRTATION, OBSCENE, SPAM
|
||||
'requestedAttributes': {'TOXICITY': {}, 'SEVERE_TOXICITY': {}, 'IDENTITY_ATTACK': {}, 'INSULT': {}, 'PROFANITY': {}, 'THREAT': {}, 'SEXUALLY_EXPLICIT': {}, 'FLIRTATION': {}, 'OBSCENE': {}, 'SPAM': {}},
|
||||
#we will analyze the text in any language automatically detected by google
|
||||
'languages': [],
|
||||
'doNotStore': 'true' # We don't want google to store the data because of privacy reasons & the GDPR (General Data Protection Regulation, an EU law that protects the privacy of EU citizens and residents for data privacy and security purposes https://gdpr-info.eu/)
|
||||
}
|
||||
}
|
||||
analyze_request_not_en = {
|
||||
'comment': {'text': ''}, # The text to analyze
|
||||
#we will ask the following attributes to google: TOXICITY, SEVERE_TOXICITY, IDENTITY_ATTACK, INSULT, PRPFANITY, THREAT, SEXUALLY_EXPLICIT, FLIRTATION, OBSCENE, SPAM
|
||||
'requestedAttributes': {'TOXICITY': {}, 'SEVERE_TOXICITY': {}, 'IDENTITY_ATTACK': {}, 'INSULT': {}, 'PROFANITY': {}, 'THREAT': {}},
|
||||
#we will analyze the text in any language automatically detected by google
|
||||
'languages': [],
|
||||
'doNotStore': 'true' # We don't want google to store the data because of privacy reasons & the GDPR (General Data Protection Regulation, an EU law that protects the privacy of EU citizens and residents for data privacy and security purposes https://gdpr-info.eu/)
|
||||
}
|
||||
def get_toxicity(message: str):
|
||||
#we first remove all kind of markdown from the message to avoid exploits
|
||||
message = re.sub(r'\*([^*]+)\*', r'\1', message)
|
||||
@@ -28,22 +51,27 @@ def get_toxicity(message: str):
|
||||
message = re.sub(r'\~\~([^~]+)\~\~', r'\1', message)
|
||||
message = re.sub(r'\`([^`]+)\`', r'\1', message)
|
||||
message = re.sub(r'\`\`\`([^`]+)\`\`\`', r'\1', message)
|
||||
analyze_request['comment']['text'] = message
|
||||
response = client.comments().analyze(body=analyze_request).execute()
|
||||
return float(response['attributeScores']['TOXICITY']['summaryScore']['value'])
|
||||
|
||||
#we try doing the request in english, but if we get 'errorType': 'LANGUAGE_NOT_SUPPORTED_BY_ATTRIBUTE' we try again with the analyze_request_not_en
|
||||
try:
|
||||
analyze_request['comment']['text'] = message
|
||||
response = client.comments().analyze(body=analyze_request).execute()
|
||||
except:
|
||||
analyze_request_not_en['comment']['text'] = message
|
||||
response = client.comments().analyze(body=analyze_request_not_en).execute()
|
||||
try: return [float(response['attributeScores']['TOXICITY']['summaryScore']['value']), float(response['attributeScores']['SEVERE_TOXICITY']['summaryScore']['value']), float(response['attributeScores']['IDENTITY_ATTACK']['summaryScore']['value']), float(response['attributeScores']['INSULT']['summaryScore']['value']), float(response['attributeScores']['PROFANITY']['summaryScore']['value']), float(response['attributeScores']['THREAT']['summaryScore']['value']), float(response['attributeScores']['SEXUALLY_EXPLICIT']['summaryScore']['value']), float(response['attributeScores']['FLIRTATION']['summaryScore']['value']), float(response['attributeScores']['OBSCENE']['summaryScore']['value']), float(response['attributeScores']['SPAM']['summaryScore']['value'])]
|
||||
except: return [float(response['attributeScores']['TOXICITY']['summaryScore']['value']), float(response['attributeScores']['SEVERE_TOXICITY']['summaryScore']['value']), float(response['attributeScores']['IDENTITY_ATTACK']['summaryScore']['value']), float(response['attributeScores']['INSULT']['summaryScore']['value']), float(response['attributeScores']['PROFANITY']['summaryScore']['value']), float(response['attributeScores']['THREAT']['summaryScore']['value'])]
|
||||
|
||||
#test part
|
||||
def test():
|
||||
print(get_toxicity("Hello world"))
|
||||
print(get_toxicity("You are a stupid bot I hate you!!!"))
|
||||
print(get_toxicity("Je suis un bot stupide, je vous déteste !!!"))
|
||||
print(get_toxicity("Ciao, come state?"))
|
||||
print(get_toxicity("Siete tutti degli scemi"))
|
||||
print(get_toxicity("Siete tutti degli stupidi"))
|
||||
print(get_toxicity("Je n'aime pas les gens stupides"))
|
||||
#markdown removal test
|
||||
print(get_toxicity("You are all stupid"))
|
||||
print(get_toxicity("You are all *s*t*u*p*i*d"))
|
||||
print(print("*** you"))
|
||||
print("Testing toxicity.py...")
|
||||
print("Hello world:")
|
||||
result = get_toxicity('Hello world')
|
||||
try: print(f"TOXICITY: {result[0]}; SEVERE_TOXICITY: {result[1]}; IDENTITY ATTACK: {result[2]}; INSULT: {result[3]}; PROFANITY: {result[4]}; THREAT: {result[5]}; SEXUALLY EXPLICIT: {result[6]}; FLIRTATION: {result[7]}; OBSCENE: {result[8]}; SPAM: {result[9]}")
|
||||
except: print(f"TOXICITY: {result[0]}; SEVERE_TOXICITY: {result[1]}; IDENTITY ATTACK: {result[2]}; INSULT: {result[3]}; PROFANITY: {result[4]}; THREAT: {result[5]}")
|
||||
print("HELLO WORLD GET ABSOLUTELY BUY MY NEW MERCH OMGGGGGGG:")
|
||||
result = get_toxicity('HELLO WORLD GET ABSOLUTELY BUY MY NEW MERCH OMGGGGGGG')
|
||||
try: print(f"TOXICITY: {result[0]}; SEVERE_TOXICITY: {result[1]}; IDENTITY ATTACK: {result[2]}; INSULT: {result[3]}; PROFANITY: {result[4]}; THREAT: {result[5]}; SEXUALLY EXPLICIT: {result[6]}; FLIRTATION: {result[7]}; OBSCENE: {result[8]}; SPAM: {result[9]}")
|
||||
except: print(f"TOXICITY: {result[0]}; SEVERE_TOXICITY: {result[1]}; IDENTITY ATTACK: {result[2]}; INSULT: {result[3]}; PROFANITY: {result[4]}; THREAT: {result[5]}")
|
||||
#uncomment the following line to test the code
|
||||
#test()
|
||||
Reference in New Issue
Block a user