Spaces:
Running
Running
Update app.py
Browse files
app.py
CHANGED
|
@@ -5,6 +5,7 @@ import spacy
|
|
| 5 |
import subprocess
|
| 6 |
import nltk
|
| 7 |
from nltk.corpus import wordnet
|
|
|
|
| 8 |
|
| 9 |
# Initialize the English text classification pipeline for AI detection
|
| 10 |
pipeline_en = pipeline(task="text-classification", model="Hello-SimpleAI/chatgpt-detector-roberta")
|
|
@@ -25,6 +26,9 @@ except OSError:
|
|
| 25 |
subprocess.run(["python", "-m", "spacy", "download", "en_core_web_sm"])
|
| 26 |
nlp = spacy.load("en_core_web_sm")
|
| 27 |
|
|
|
|
|
|
|
|
|
|
| 28 |
# Function to get synonyms using NLTK WordNet (Humanifier)
|
| 29 |
def get_synonyms_nltk(word, pos):
|
| 30 |
synsets = wordnet.synsets(word, pos=pos)
|
|
@@ -51,20 +55,6 @@ def capitalize_sentences_and_nouns(text):
|
|
| 51 |
|
| 52 |
return ' '.join(corrected_text)
|
| 53 |
|
| 54 |
-
# Function to correct tense errors in a sentence (Tense Correction)
|
| 55 |
-
def correct_tense_errors(text):
|
| 56 |
-
doc = nlp(text)
|
| 57 |
-
corrected_text = []
|
| 58 |
-
for token in doc:
|
| 59 |
-
# Check for tense correction based on modal verbs
|
| 60 |
-
if token.pos_ == "VERB" and token.dep_ in {"aux", "auxpass"}:
|
| 61 |
-
# Replace with appropriate verb form
|
| 62 |
-
lemma = wordnet.morphy(token.text, wordnet.VERB) or token.text
|
| 63 |
-
corrected_text.append(lemma)
|
| 64 |
-
else:
|
| 65 |
-
corrected_text.append(token.text)
|
| 66 |
-
return ' '.join(corrected_text)
|
| 67 |
-
|
| 68 |
# Function to correct singular/plural errors (Singular/Plural Correction)
|
| 69 |
def correct_singular_plural_errors(text):
|
| 70 |
doc = nlp(text)
|
|
@@ -72,15 +62,12 @@ def correct_singular_plural_errors(text):
|
|
| 72 |
|
| 73 |
for token in doc:
|
| 74 |
if token.pos_ == "NOUN":
|
| 75 |
-
# Check if the noun is singular or plural
|
| 76 |
if token.tag_ == "NN": # Singular noun
|
| 77 |
-
# Look for determiners like "many", "several", "few" to correct to plural
|
| 78 |
if any(child.text.lower() in ['many', 'several', 'few'] for child in token.head.children):
|
| 79 |
corrected_text.append(token.lemma_ + 's')
|
| 80 |
else:
|
| 81 |
corrected_text.append(token.text)
|
| 82 |
elif token.tag_ == "NNS": # Plural noun
|
| 83 |
-
# Look for determiners like "a", "one" to correct to singular
|
| 84 |
if any(child.text.lower() in ['a', 'one'] for child in token.head.children):
|
| 85 |
corrected_text.append(token.lemma_)
|
| 86 |
else:
|
|
@@ -90,6 +77,18 @@ def correct_singular_plural_errors(text):
|
|
| 90 |
|
| 91 |
return ' '.join(corrected_text)
|
| 92 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 93 |
# Function to check and correct article errors
|
| 94 |
def correct_article_errors(text):
|
| 95 |
doc = nlp(text)
|
|
@@ -123,12 +122,11 @@ def replace_with_synonym(token):
|
|
| 123 |
|
| 124 |
if synonyms:
|
| 125 |
synonym = synonyms[0]
|
| 126 |
-
|
| 127 |
-
if token.tag_ == "VBG": # Present participle (e.g., running)
|
| 128 |
synonym = synonym + 'ing'
|
| 129 |
-
elif token.tag_ == "VBD" or token.tag_ == "VBN":
|
| 130 |
synonym = synonym + 'ed'
|
| 131 |
-
elif token.tag_ == "VBZ":
|
| 132 |
synonym = synonym + 's'
|
| 133 |
return synonym
|
| 134 |
return token.text
|
|
@@ -139,7 +137,6 @@ def correct_double_negatives(text):
|
|
| 139 |
corrected_text = []
|
| 140 |
for token in doc:
|
| 141 |
if token.text.lower() == "not" and any(child.text.lower() == "never" for child in token.head.children):
|
| 142 |
-
# Replace the double negative with a positive statement
|
| 143 |
corrected_text.append("always")
|
| 144 |
else:
|
| 145 |
corrected_text.append(token.text)
|
|
@@ -151,15 +148,20 @@ def ensure_subject_verb_agreement(text):
|
|
| 151 |
corrected_text = []
|
| 152 |
for token in doc:
|
| 153 |
if token.dep_ == "nsubj" and token.head.pos_ == "VERB":
|
| 154 |
-
|
| 155 |
-
if token.tag_ == "NN" and token.head.tag_ != "VBZ": # Singular noun, should use singular verb
|
| 156 |
corrected_text.append(token.head.lemma_ + "s")
|
| 157 |
-
elif token.tag_ == "NNS" and token.head.tag_ == "VBZ":
|
| 158 |
corrected_text.append(token.head.lemma_)
|
| 159 |
corrected_text.append(token.text)
|
| 160 |
return ' '.join(corrected_text)
|
| 161 |
|
| 162 |
-
# Function to
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 163 |
def paraphrase_and_correct(text):
|
| 164 |
# Capitalize first to ensure proper noun capitalization
|
| 165 |
paraphrased_text = capitalize_sentences_and_nouns(text)
|
|
@@ -180,7 +182,10 @@ def paraphrase_and_correct(text):
|
|
| 180 |
else:
|
| 181 |
final_text.append(token.text)
|
| 182 |
|
| 183 |
-
|
|
|
|
|
|
|
|
|
|
| 184 |
|
| 185 |
# Gradio app setup with two tabs
|
| 186 |
with gr.Blocks() as demo:
|
|
|
|
| 5 |
import subprocess
|
| 6 |
import nltk
|
| 7 |
from nltk.corpus import wordnet
|
| 8 |
+
from spellchecker import SpellChecker # Import SpellChecker for spelling correction
|
| 9 |
|
| 10 |
# Initialize the English text classification pipeline for AI detection
|
| 11 |
pipeline_en = pipeline(task="text-classification", model="Hello-SimpleAI/chatgpt-detector-roberta")
|
|
|
|
| 26 |
subprocess.run(["python", "-m", "spacy", "download", "en_core_web_sm"])
|
| 27 |
nlp = spacy.load("en_core_web_sm")
|
| 28 |
|
| 29 |
+
# Initialize SpellChecker
|
| 30 |
+
spell = SpellChecker()
|
| 31 |
+
|
| 32 |
# Function to get synonyms using NLTK WordNet (Humanifier)
|
| 33 |
def get_synonyms_nltk(word, pos):
|
| 34 |
synsets = wordnet.synsets(word, pos=pos)
|
|
|
|
| 55 |
|
| 56 |
return ' '.join(corrected_text)
|
| 57 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 58 |
# Function to correct singular/plural errors (Singular/Plural Correction)
|
| 59 |
def correct_singular_plural_errors(text):
|
| 60 |
doc = nlp(text)
|
|
|
|
| 62 |
|
| 63 |
for token in doc:
|
| 64 |
if token.pos_ == "NOUN":
|
|
|
|
| 65 |
if token.tag_ == "NN": # Singular noun
|
|
|
|
| 66 |
if any(child.text.lower() in ['many', 'several', 'few'] for child in token.head.children):
|
| 67 |
corrected_text.append(token.lemma_ + 's')
|
| 68 |
else:
|
| 69 |
corrected_text.append(token.text)
|
| 70 |
elif token.tag_ == "NNS": # Plural noun
|
|
|
|
| 71 |
if any(child.text.lower() in ['a', 'one'] for child in token.head.children):
|
| 72 |
corrected_text.append(token.lemma_)
|
| 73 |
else:
|
|
|
|
| 77 |
|
| 78 |
return ' '.join(corrected_text)
|
| 79 |
|
| 80 |
+
# Function to correct tense errors in a sentence (Tense Correction)
|
| 81 |
+
def correct_tense_errors(text):
|
| 82 |
+
doc = nlp(text)
|
| 83 |
+
corrected_text = []
|
| 84 |
+
for token in doc:
|
| 85 |
+
if token.pos_ == "VERB" and token.dep_ in {"aux", "auxpass"}:
|
| 86 |
+
lemma = wordnet.morphy(token.text, wordnet.VERB) or token.text
|
| 87 |
+
corrected_text.append(lemma)
|
| 88 |
+
else:
|
| 89 |
+
corrected_text.append(token.text)
|
| 90 |
+
return ' '.join(corrected_text)
|
| 91 |
+
|
| 92 |
# Function to check and correct article errors
|
| 93 |
def correct_article_errors(text):
|
| 94 |
doc = nlp(text)
|
|
|
|
| 122 |
|
| 123 |
if synonyms:
|
| 124 |
synonym = synonyms[0]
|
| 125 |
+
if token.tag_ == "VBG":
|
|
|
|
| 126 |
synonym = synonym + 'ing'
|
| 127 |
+
elif token.tag_ == "VBD" or token.tag_ == "VBN":
|
| 128 |
synonym = synonym + 'ed'
|
| 129 |
+
elif token.tag_ == "VBZ":
|
| 130 |
synonym = synonym + 's'
|
| 131 |
return synonym
|
| 132 |
return token.text
|
|
|
|
| 137 |
corrected_text = []
|
| 138 |
for token in doc:
|
| 139 |
if token.text.lower() == "not" and any(child.text.lower() == "never" for child in token.head.children):
|
|
|
|
| 140 |
corrected_text.append("always")
|
| 141 |
else:
|
| 142 |
corrected_text.append(token.text)
|
|
|
|
| 148 |
corrected_text = []
|
| 149 |
for token in doc:
|
| 150 |
if token.dep_ == "nsubj" and token.head.pos_ == "VERB":
|
| 151 |
+
if token.tag_ == "NN" and token.head.tag_ != "VBZ":
|
|
|
|
| 152 |
corrected_text.append(token.head.lemma_ + "s")
|
| 153 |
+
elif token.tag_ == "NNS" and token.head.tag_ == "VBZ":
|
| 154 |
corrected_text.append(token.head.lemma_)
|
| 155 |
corrected_text.append(token.text)
|
| 156 |
return ' '.join(corrected_text)
|
| 157 |
|
| 158 |
+
# Function to correct spelling errors
|
| 159 |
+
def correct_spelling(text):
|
| 160 |
+
words = text.split()
|
| 161 |
+
corrected_words = [spell.candidates(word) or word for word in words]
|
| 162 |
+
return ' '.join(corrected_words)
|
| 163 |
+
|
| 164 |
+
# Function to paraphrase, correct grammar, and fix spelling errors
|
| 165 |
def paraphrase_and_correct(text):
|
| 166 |
# Capitalize first to ensure proper noun capitalization
|
| 167 |
paraphrased_text = capitalize_sentences_and_nouns(text)
|
|
|
|
| 182 |
else:
|
| 183 |
final_text.append(token.text)
|
| 184 |
|
| 185 |
+
# Correct spelling errors
|
| 186 |
+
final_text = correct_spelling(' '.join(final_text))
|
| 187 |
+
|
| 188 |
+
return final_text
|
| 189 |
|
| 190 |
# Gradio app setup with two tabs
|
| 191 |
with gr.Blocks() as demo:
|