refresh

nlp/replying.py

@@ -1,59 +0,0 @@
-#  "👍", "👎", "❤", "🔥", "🥰", "👏", "😁",
-# "🤔", "🤯", "😱", "🤬", "😢", "🎉", "🤩",
-# "🤮", "💩", "🙏", "👌", "🕊", "🤡", "🥱",
-# "🥴", "😍", "🐳", "❤‍🔥", "🌚", "🌭", "💯",
-# "🤣", "⚡", "🍌", "🏆", "💔", "🤨", "😐",
-# "🍓", "🍾", "💋", "🖕", "😈", "😴", "😭",
-# "🤓", "👻", "👨‍💻", "👀", "🎃", "🙈", "😇",
-# "😨", "🤝", "✍", "🤗", "🫡", "🎅", "🎄",
-# "☃", "💅", "🤪", "🗿", "🆒", "💘", "🙉",
-# "🦄", "😘", "💊", "🙊", "😎", "👾", "🤷‍♂",
-# "🤷", "🤷‍♀", "😡"
-
-toxic_reactions = {
-    "071": "🕊",
-    "073": "👀",
-    "075": "🙈",
-    "077": "🙊",
-    "079": "🙏",
-    "081": "🤔",
-    "083": "😐",
-    "085": "🤨",
-    "087": "🥴",
-    "089": "🤯",
-    "091": "😢",
-    "093": "😭",
-    "095": "😨",
-    "097": "😱",
-    "099": "🤬"
-}
-
-grads = list(toxic_reactions.keys())
-grads.sort()
-grads.reverse()
-
-abusive_reactions = {
-    "085": "🫡",
-    "088": "💅",
-    "091": "🤷‍♀",
-    "094": "👾",
-    "097": "👻",
-    "099": "😈"
-}
-
-abusive_grads = list(abusive_reactions.keys())
-abusive_grads.sort()
-abusive_grads.reverse()
-
-def get_toxic_reply(tx):
-    percentage = tx * 100
-    for key in grads:
-        if percentage > int(key):
-            return toxic_reactions[key]
-
-
-def get_abusive_reply(tx):
-    percentage = tx * 100
-    for key in abusive_grads:
-        if percentage > int(key):
-            return abusive_reactions[key]

nlp/toxicity.py

@@ -1,31 +0,0 @@
-import torch
-from transformers import AutoTokenizer, \
-    AutoModelForSequenceClassification
-
-tiny_tox_model_path = 'cointegrated/rubert-tiny-toxicity'
-tiny_tox_tokenizer = AutoTokenizer.from_pretrained(tiny_tox_model_path)
-tiny_tox_model = AutoModelForSequenceClassification.from_pretrained(
-    tiny_tox_model_path)
-
-
-# if torch.cuda.is_available():
-#    model.cuda()
-
-
-def text2toxicity(text, aggregate=True) -> float:
-    """ Calculate toxicity of a text (if aggregate=True)
-    or a vector of toxicity aspects (if aggregate=False)"""
-    proba = 0.0
-    with torch.no_grad():
-        inputs = tiny_tox_tokenizer(
-            text.lower(),
-            return_tensors='pt',
-            truncation=True,
-            padding=True
-        ).to(tiny_tox_model.device)
-        proba = torch.sigmoid(tiny_tox_model(**inputs).logits).cpu().numpy()
-    if isinstance(text, str):
-        proba = proba[0]
-    if aggregate:
-        return 1 - proba.T[0] * (1 - proba.T[-1])
-    return float(proba)