Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
|
@@ -42,23 +42,19 @@ def fix_punctuation_spacing(text):
|
|
| 42 |
# Split the text into words and punctuation
|
| 43 |
words = text.split(' ')
|
| 44 |
cleaned_words = []
|
| 45 |
-
|
| 46 |
-
# Define punctuation marks to check
|
| 47 |
punctuation_marks = {',', '.', "'", '!', '?', ':'}
|
| 48 |
|
| 49 |
for word in words:
|
| 50 |
-
# If the word ends with a punctuation mark, remove space before it
|
| 51 |
if cleaned_words and word and word[0] in punctuation_marks:
|
| 52 |
-
cleaned_words[-1] += word
|
| 53 |
else:
|
| 54 |
-
cleaned_words.append(word)
|
| 55 |
|
| 56 |
return ' '.join(cleaned_words).replace(' ,', ',').replace(' .', '.').replace(" '", "'") \
|
| 57 |
.replace(' !', '!').replace(' ?', '?').replace(' :', ':')
|
| 58 |
|
| 59 |
# Function to fix possessives like "Earth's"
|
| 60 |
def fix_possessives(text):
|
| 61 |
-
# Simple rule to catch possessives and correct spacing
|
| 62 |
text = re.sub(r'(\w)\s\'\s?s', r"\1's", text)
|
| 63 |
return text
|
| 64 |
|
|
@@ -70,9 +66,9 @@ def capitalize_sentences_and_nouns(text):
|
|
| 70 |
for sent in doc.sents:
|
| 71 |
sentence = []
|
| 72 |
for token in sent:
|
| 73 |
-
if token.i == sent.start:
|
| 74 |
sentence.append(token.text.capitalize())
|
| 75 |
-
elif token.pos_ == "PROPN":
|
| 76 |
sentence.append(token.text.capitalize())
|
| 77 |
else:
|
| 78 |
sentence.append(token.text)
|
|
@@ -82,14 +78,12 @@ def capitalize_sentences_and_nouns(text):
|
|
| 82 |
|
| 83 |
# Function to force capitalization of the first letter of every sentence and ensure full stops
|
| 84 |
def force_first_letter_capital(text):
|
| 85 |
-
sentences = re.split(r'(?<=\w[.!?])\s+', text)
|
| 86 |
capitalized_sentences = []
|
| 87 |
|
| 88 |
for sentence in sentences:
|
| 89 |
if sentence:
|
| 90 |
-
# Capitalize the first letter if not already capitalized
|
| 91 |
capitalized_sentence = sentence[0].capitalize() + sentence[1:]
|
| 92 |
-
# Ensure there's a full stop at the end if there's no punctuation
|
| 93 |
if not re.search(r'[.!?]$', capitalized_sentence):
|
| 94 |
capitalized_sentence += '.'
|
| 95 |
capitalized_sentences.append(capitalized_sentence)
|
|
@@ -131,9 +125,9 @@ def ensure_subject_verb_agreement(text):
|
|
| 131 |
corrected_text = []
|
| 132 |
for token in doc:
|
| 133 |
if token.dep_ == "nsubj" and token.head.pos_ == "VERB":
|
| 134 |
-
if token.tag_ == "NN" and token.head.tag_ != "VBZ":
|
| 135 |
corrected_text.append(token.head.lemma_ + "s")
|
| 136 |
-
elif token.tag_ == "NNS" and token.head.tag_ == "VBZ":
|
| 137 |
corrected_text.append(token.head.lemma_)
|
| 138 |
corrected_text.append(token.text)
|
| 139 |
return ' '.join(corrected_text)
|
|
@@ -147,31 +141,41 @@ def correct_spelling(text):
|
|
| 147 |
if corrected_word is not None:
|
| 148 |
corrected_words.append(corrected_word)
|
| 149 |
else:
|
| 150 |
-
corrected_words.append(word)
|
| 151 |
return ' '.join(corrected_words)
|
| 152 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 153 |
# Main function for paraphrasing and grammar correction
|
| 154 |
def paraphrase_and_correct(text):
|
| 155 |
-
# Remove meaningless or redundant words first
|
| 156 |
cleaned_text = remove_redundant_words(text)
|
| 157 |
-
|
| 158 |
-
# Capitalize sentences and nouns
|
| 159 |
paraphrased_text = capitalize_sentences_and_nouns(cleaned_text)
|
| 160 |
-
|
| 161 |
-
# Ensure first letter of each sentence is capitalized
|
| 162 |
paraphrased_text = force_first_letter_capital(paraphrased_text)
|
| 163 |
-
|
| 164 |
-
# Apply grammatical corrections
|
| 165 |
paraphrased_text = correct_article_errors(paraphrased_text)
|
| 166 |
paraphrased_text = correct_tense_errors(paraphrased_text)
|
| 167 |
paraphrased_text = ensure_subject_verb_agreement(paraphrased_text)
|
| 168 |
-
|
| 169 |
-
# Fix punctuation spacing and possessives
|
| 170 |
paraphrased_text = fix_possessives(paraphrased_text)
|
| 171 |
-
|
| 172 |
-
# Correct spelling errors
|
| 173 |
paraphrased_text = correct_spelling(paraphrased_text)
|
| 174 |
paraphrased_text = fix_punctuation_spacing(paraphrased_text)
|
|
|
|
| 175 |
|
| 176 |
return paraphrased_text
|
| 177 |
|
|
|
|
| 42 |
# Split the text into words and punctuation
|
| 43 |
words = text.split(' ')
|
| 44 |
cleaned_words = []
|
|
|
|
|
|
|
| 45 |
punctuation_marks = {',', '.', "'", '!', '?', ':'}
|
| 46 |
|
| 47 |
for word in words:
|
|
|
|
| 48 |
if cleaned_words and word and word[0] in punctuation_marks:
|
| 49 |
+
cleaned_words[-1] += word
|
| 50 |
else:
|
| 51 |
+
cleaned_words.append(word)
|
| 52 |
|
| 53 |
return ' '.join(cleaned_words).replace(' ,', ',').replace(' .', '.').replace(" '", "'") \
|
| 54 |
.replace(' !', '!').replace(' ?', '?').replace(' :', ':')
|
| 55 |
|
| 56 |
# Function to fix possessives like "Earth's"
|
| 57 |
def fix_possessives(text):
|
|
|
|
| 58 |
text = re.sub(r'(\w)\s\'\s?s', r"\1's", text)
|
| 59 |
return text
|
| 60 |
|
|
|
|
| 66 |
for sent in doc.sents:
|
| 67 |
sentence = []
|
| 68 |
for token in sent:
|
| 69 |
+
if token.i == sent.start:
|
| 70 |
sentence.append(token.text.capitalize())
|
| 71 |
+
elif token.pos_ == "PROPN":
|
| 72 |
sentence.append(token.text.capitalize())
|
| 73 |
else:
|
| 74 |
sentence.append(token.text)
|
|
|
|
| 78 |
|
| 79 |
# Function to force capitalization of the first letter of every sentence and ensure full stops
|
| 80 |
def force_first_letter_capital(text):
|
| 81 |
+
sentences = re.split(r'(?<=\w[.!?])\s+', text)
|
| 82 |
capitalized_sentences = []
|
| 83 |
|
| 84 |
for sentence in sentences:
|
| 85 |
if sentence:
|
|
|
|
| 86 |
capitalized_sentence = sentence[0].capitalize() + sentence[1:]
|
|
|
|
| 87 |
if not re.search(r'[.!?]$', capitalized_sentence):
|
| 88 |
capitalized_sentence += '.'
|
| 89 |
capitalized_sentences.append(capitalized_sentence)
|
|
|
|
| 125 |
corrected_text = []
|
| 126 |
for token in doc:
|
| 127 |
if token.dep_ == "nsubj" and token.head.pos_ == "VERB":
|
| 128 |
+
if token.tag_ == "NN" and token.head.tag_ != "VBZ":
|
| 129 |
corrected_text.append(token.head.lemma_ + "s")
|
| 130 |
+
elif token.tag_ == "NNS" and token.head.tag_ == "VBZ":
|
| 131 |
corrected_text.append(token.head.lemma_)
|
| 132 |
corrected_text.append(token.text)
|
| 133 |
return ' '.join(corrected_text)
|
|
|
|
| 141 |
if corrected_word is not None:
|
| 142 |
corrected_words.append(corrected_word)
|
| 143 |
else:
|
| 144 |
+
corrected_words.append(word)
|
| 145 |
return ' '.join(corrected_words)
|
| 146 |
|
| 147 |
+
# Function to replace a word with its synonym
|
| 148 |
+
def replace_with_synonyms(text):
|
| 149 |
+
words = text.split()
|
| 150 |
+
replaced_words = []
|
| 151 |
+
|
| 152 |
+
for word in words:
|
| 153 |
+
synonyms = wordnet.synsets(word)
|
| 154 |
+
if synonyms:
|
| 155 |
+
# Take the first synonym if available
|
| 156 |
+
synonym = synonyms[0].lemmas()[0].name()
|
| 157 |
+
# Replace the word with its synonym if it's different
|
| 158 |
+
if synonym.lower() != word.lower():
|
| 159 |
+
replaced_words.append(synonym.replace('_', ' '))
|
| 160 |
+
else:
|
| 161 |
+
replaced_words.append(word)
|
| 162 |
+
else:
|
| 163 |
+
replaced_words.append(word)
|
| 164 |
+
|
| 165 |
+
return ' '.join(replaced_words)
|
| 166 |
+
|
| 167 |
# Main function for paraphrasing and grammar correction
|
| 168 |
def paraphrase_and_correct(text):
|
|
|
|
| 169 |
cleaned_text = remove_redundant_words(text)
|
|
|
|
|
|
|
| 170 |
paraphrased_text = capitalize_sentences_and_nouns(cleaned_text)
|
|
|
|
|
|
|
| 171 |
paraphrased_text = force_first_letter_capital(paraphrased_text)
|
|
|
|
|
|
|
| 172 |
paraphrased_text = correct_article_errors(paraphrased_text)
|
| 173 |
paraphrased_text = correct_tense_errors(paraphrased_text)
|
| 174 |
paraphrased_text = ensure_subject_verb_agreement(paraphrased_text)
|
|
|
|
|
|
|
| 175 |
paraphrased_text = fix_possessives(paraphrased_text)
|
|
|
|
|
|
|
| 176 |
paraphrased_text = correct_spelling(paraphrased_text)
|
| 177 |
paraphrased_text = fix_punctuation_spacing(paraphrased_text)
|
| 178 |
+
paraphrased_text = replace_with_synonyms(paraphrased_text) # Add synonym replacement here
|
| 179 |
|
| 180 |
return paraphrased_text
|
| 181 |
|