Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
|
@@ -1,4 +1,3 @@
|
|
| 1 |
-
import os
|
| 2 |
import gradio as gr
|
| 3 |
from transformers import pipeline
|
| 4 |
import spacy
|
|
@@ -6,121 +5,33 @@ import subprocess
|
|
| 6 |
import nltk
|
| 7 |
from nltk.corpus import wordnet
|
| 8 |
from spellchecker import SpellChecker
|
| 9 |
-
import language_tool_python
|
| 10 |
|
| 11 |
-
# Initialize
|
| 12 |
-
pipeline_en = pipeline(task="text-classification", model="Hello-SimpleAI/chatgpt-detector-roberta")
|
| 13 |
|
| 14 |
-
#
|
| 15 |
-
|
| 16 |
-
|
| 17 |
-
|
| 18 |
-
|
| 19 |
-
|
| 20 |
-
# Ensure necessary NLTK data is downloaded
|
| 21 |
-
nltk.download('wordnet')
|
| 22 |
-
nltk.download('omw-1.4')
|
| 23 |
-
|
| 24 |
-
# Ensure the SpaCy model is installed
|
| 25 |
-
try:
|
| 26 |
-
nlp = spacy.load("en_core_web_sm")
|
| 27 |
-
except OSError:
|
| 28 |
-
subprocess.run(["python", "-m", "spacy", "download", "en_core_web_sm"])
|
| 29 |
-
nlp = spacy.load("en_core_web_sm")
|
| 30 |
-
|
| 31 |
-
# Function to predict the label and score for English text (AI Detection)
|
| 32 |
-
def predict_en(text):
|
| 33 |
-
res = pipeline_en(text)[0]
|
| 34 |
-
return res['label'], res['score']
|
| 35 |
-
|
| 36 |
-
# Function to remove redundant and meaningless words
|
| 37 |
-
def remove_redundant_words(text):
|
| 38 |
-
doc = nlp(text)
|
| 39 |
-
meaningless_words = {"actually", "basically", "literally", "really", "very", "just"}
|
| 40 |
-
filtered_text = [token.text for token in doc if token.text.lower() not in meaningless_words]
|
| 41 |
-
return ' '.join(filtered_text)
|
| 42 |
-
|
| 43 |
-
# Function to apply grammatical corrections using LanguageTool
|
| 44 |
-
def correct_grammar(text):
|
| 45 |
-
corrected_text = tool.correct(text)
|
| 46 |
-
return corrected_text
|
| 47 |
-
|
| 48 |
-
# Function to correct spelling errors
|
| 49 |
-
def correct_spelling(text):
|
| 50 |
-
words = text.split()
|
| 51 |
-
corrected_words = []
|
| 52 |
-
for word in words:
|
| 53 |
-
corrected_word = spell.correction(word)
|
| 54 |
-
corrected_words.append(corrected_word if corrected_word else word) # Keep original word if no correction
|
| 55 |
-
return ' '.join(corrected_words)
|
| 56 |
-
|
| 57 |
-
# Function to capitalize the first letter of each sentence and proper nouns
|
| 58 |
-
def capitalize_sentences_and_nouns(text):
|
| 59 |
-
doc = nlp(text)
|
| 60 |
-
corrected_text = []
|
| 61 |
-
|
| 62 |
-
for sent in doc.sents:
|
| 63 |
-
sentence = []
|
| 64 |
-
for token in sent:
|
| 65 |
-
if token.i == sent.start: # First word of the sentence
|
| 66 |
-
sentence.append(token.text.capitalize())
|
| 67 |
-
elif token.pos_ == "PROPN": # Proper noun
|
| 68 |
-
sentence.append(token.text.capitalize())
|
| 69 |
-
else:
|
| 70 |
-
sentence.append(token.text)
|
| 71 |
-
corrected_text.append(' '.join(sentence))
|
| 72 |
-
|
| 73 |
-
return ' '.join(corrected_text)
|
| 74 |
-
|
| 75 |
-
# Function to rephrase with contextually appropriate synonyms
|
| 76 |
-
def rephrase_with_synonyms(text):
|
| 77 |
-
doc = nlp(text)
|
| 78 |
-
rephrased_text = []
|
| 79 |
-
|
| 80 |
-
for token in doc:
|
| 81 |
-
pos_tag = None
|
| 82 |
-
if token.pos_ == "NOUN":
|
| 83 |
-
pos_tag = wordnet.NOUN
|
| 84 |
-
elif token.pos_ == "VERB":
|
| 85 |
-
pos_tag = wordnet.VERB
|
| 86 |
-
elif token.pos_ == "ADJ":
|
| 87 |
-
pos_tag = wordnet.ADJ
|
| 88 |
-
elif token.pos_ == "ADV":
|
| 89 |
-
pos_tag = wordnet.ADV
|
| 90 |
-
|
| 91 |
-
if pos_tag:
|
| 92 |
-
synonyms = wordnet.synsets(token.text, pos=pos_tag)
|
| 93 |
-
if synonyms:
|
| 94 |
-
synonym = synonyms[0].lemmas()[0].name() # Choose the first synonym
|
| 95 |
-
rephrased_text.append(synonym)
|
| 96 |
-
else:
|
| 97 |
-
rephrased_text.append(token.text)
|
| 98 |
-
else:
|
| 99 |
-
rephrased_text.append(token.text)
|
| 100 |
-
|
| 101 |
-
return ' '.join(rephrased_text)
|
| 102 |
-
|
| 103 |
-
# Comprehensive function for paraphrasing and grammar correction
|
| 104 |
-
def paraphrase_and_correct(text):
|
| 105 |
-
# Step 1: Remove meaningless or redundant words
|
| 106 |
-
cleaned_text = remove_redundant_words(text)
|
| 107 |
-
|
| 108 |
-
# Step 2: Capitalize sentences and proper nouns
|
| 109 |
-
paraphrased_text = capitalize_sentences_and_nouns(cleaned_text)
|
| 110 |
|
| 111 |
-
|
| 112 |
-
|
|
|
|
|
|
|
|
|
|
| 113 |
|
| 114 |
-
|
| 115 |
-
|
| 116 |
-
|
| 117 |
-
|
| 118 |
-
|
| 119 |
-
|
| 120 |
-
|
| 121 |
-
|
| 122 |
-
|
| 123 |
-
|
|
|
|
|
|
|
| 124 |
|
| 125 |
# Gradio app setup with two tabs
|
| 126 |
with gr.Blocks() as demo:
|
|
@@ -130,15 +41,15 @@ with gr.Blocks() as demo:
|
|
| 130 |
label1 = gr.Textbox(lines=1, label='Predicted Label π')
|
| 131 |
score1 = gr.Textbox(lines=1, label='Prob')
|
| 132 |
|
| 133 |
-
# Connect the prediction function to the button
|
| 134 |
button1.click(fn=predict_en, inputs=t1, outputs=[label1, score1])
|
| 135 |
|
| 136 |
with gr.Tab("Paraphrasing & Grammar Correction"):
|
| 137 |
t2 = gr.Textbox(lines=5, label='Enter text for paraphrasing and grammar correction')
|
| 138 |
button2 = gr.Button("π Paraphrase and Correct")
|
|
|
|
| 139 |
result2 = gr.Textbox(lines=5, label='Corrected Text')
|
| 140 |
|
| 141 |
-
# Connect the paraphrasing and correction function to the button
|
| 142 |
button2.click(fn=paraphrase_and_correct, inputs=t2, outputs=result2)
|
|
|
|
| 143 |
|
| 144 |
demo.launch(share=True)
|
|
|
|
|
|
|
| 1 |
import gradio as gr
|
| 2 |
from transformers import pipeline
|
| 3 |
import spacy
|
|
|
|
| 5 |
import nltk
|
| 6 |
from nltk.corpus import wordnet
|
| 7 |
from spellchecker import SpellChecker
|
|
|
|
| 8 |
|
| 9 |
+
# Initialize other components (AI detection, NLP, etc.) as before...
|
|
|
|
| 10 |
|
| 11 |
+
# Function to paraphrase and correct grammar using Ginger
|
| 12 |
+
def correct_with_ginger(text):
|
| 13 |
+
ginger_result = get_ginger_result(text)
|
| 14 |
+
if "error" in ginger_result:
|
| 15 |
+
return ginger_result["error"]
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 16 |
|
| 17 |
+
original_text = text
|
| 18 |
+
fixed_text = original_text
|
| 19 |
+
color_gap, fixed_gap = 0, 0
|
| 20 |
+
if not ginger_result["LightGingerTheTextResult"]:
|
| 21 |
+
return "No grammatical issues found!"
|
| 22 |
|
| 23 |
+
for result in ginger_result["LightGingerTheTextResult"]:
|
| 24 |
+
if result["Suggestions"]:
|
| 25 |
+
from_index = result["From"] + color_gap
|
| 26 |
+
to_index = result["To"] + 1 + color_gap
|
| 27 |
+
suggest = result["Suggestions"][0]["Text"]
|
| 28 |
+
|
| 29 |
+
original_text = original_text[:from_index] + original_text[from_index:to_index] + original_text[to_index:]
|
| 30 |
+
fixed_text = fixed_text[:from_index-fixed_gap] + suggest + fixed_text[to_index-fixed_gap:]
|
| 31 |
+
color_gap += len(suggest) - (to_index - from_index)
|
| 32 |
+
fixed_gap += to_index - from_index - len(suggest)
|
| 33 |
+
|
| 34 |
+
return fixed_text
|
| 35 |
|
| 36 |
# Gradio app setup with two tabs
|
| 37 |
with gr.Blocks() as demo:
|
|
|
|
| 41 |
label1 = gr.Textbox(lines=1, label='Predicted Label π')
|
| 42 |
score1 = gr.Textbox(lines=1, label='Prob')
|
| 43 |
|
|
|
|
| 44 |
button1.click(fn=predict_en, inputs=t1, outputs=[label1, score1])
|
| 45 |
|
| 46 |
with gr.Tab("Paraphrasing & Grammar Correction"):
|
| 47 |
t2 = gr.Textbox(lines=5, label='Enter text for paraphrasing and grammar correction')
|
| 48 |
button2 = gr.Button("π Paraphrase and Correct")
|
| 49 |
+
ginger_button = gr.Button("π§ Correct with Ginger")
|
| 50 |
result2 = gr.Textbox(lines=5, label='Corrected Text')
|
| 51 |
|
|
|
|
| 52 |
button2.click(fn=paraphrase_and_correct, inputs=t2, outputs=result2)
|
| 53 |
+
ginger_button.click(fn=correct_with_ginger, inputs=t2, outputs=result2)
|
| 54 |
|
| 55 |
demo.launch(share=True)
|