Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
|
@@ -14,8 +14,10 @@ def preprocess_capitalization(text: str) -> str:
|
|
| 14 |
processed_words = []
|
| 15 |
|
| 16 |
for word in words:
|
|
|
|
| 17 |
if re.match(r"^[A-Z]+$", word):
|
| 18 |
processed_words.append(word) # Leave acronyms unchanged
|
|
|
|
| 19 |
elif re.search(r"[A-Z]", word) and re.search(r"[a-z]", word):
|
| 20 |
processed_words.append(word[0].upper() + word[1:].lower()) # Correct capitalization
|
| 21 |
else:
|
|
@@ -23,7 +25,7 @@ def preprocess_capitalization(text: str) -> str:
|
|
| 23 |
|
| 24 |
return " ".join(processed_words)
|
| 25 |
|
| 26 |
-
def preprocess_text(text: str
|
| 27 |
"""Process text and return corrections with position information."""
|
| 28 |
result = {
|
| 29 |
"spell_suggestions": [],
|
|
@@ -42,22 +44,20 @@ def preprocess_text(text: str, is_spell_corrected: bool = False):
|
|
| 42 |
})
|
| 43 |
text = capitalized_text # Update text for further processing
|
| 44 |
|
| 45 |
-
# Transformer spell check - only for words that look misspelled
|
| 46 |
-
|
| 47 |
-
|
| 48 |
-
|
| 49 |
-
|
| 50 |
-
|
| 51 |
-
|
| 52 |
-
|
| 53 |
-
|
| 54 |
-
|
| 55 |
-
|
| 56 |
-
|
| 57 |
-
|
| 58 |
-
|
| 59 |
-
})
|
| 60 |
-
text = spell_checked # Update text after spell correction
|
| 61 |
|
| 62 |
# Add fluency/style suggestions (other suggestions)
|
| 63 |
# Only add if the word isn't already in spell suggestions
|
|
@@ -69,8 +69,8 @@ def preprocess_text(text: str, is_spell_corrected: bool = False):
|
|
| 69 |
if word in spell_originals or word.endswith(('.', '?', '!')):
|
| 70 |
continue
|
| 71 |
|
| 72 |
-
# Add style
|
| 73 |
-
if not word.endswith('!'):
|
| 74 |
result["other_suggestions"].append({
|
| 75 |
"original": word,
|
| 76 |
"corrected": word + "!",
|
|
@@ -89,6 +89,7 @@ def translate_text(text: str):
|
|
| 89 |
client = Client("Frenchizer/space_21")
|
| 90 |
try:
|
| 91 |
translation = client.predict(text)
|
|
|
|
| 92 |
result = {
|
| 93 |
"spell_suggestions": [],
|
| 94 |
"other_suggestions": [],
|
|
@@ -99,10 +100,11 @@ def translate_text(text: str):
|
|
| 99 |
except Exception as e:
|
| 100 |
return f"Error: {str(e)}", {}
|
| 101 |
|
| 102 |
-
def preprocess_and_forward(text: str
|
| 103 |
"""Process text and forward to translation service."""
|
| 104 |
-
original_text, preprocessing_result = preprocess_text(text
|
| 105 |
|
|
|
|
| 106 |
client = Client("Frenchizer/space_21")
|
| 107 |
try:
|
| 108 |
translation = client.predict(original_text)
|
|
@@ -119,10 +121,9 @@ def translate_only(text: str):
|
|
| 119 |
with gr.Blocks() as demo:
|
| 120 |
with gr.Tab("Main"):
|
| 121 |
input_text = gr.Textbox(label="Input Text")
|
| 122 |
-
is_spell_corrected = gr.Checkbox(label="Spell Corrected", value=False, visible=False) # Hidden flag
|
| 123 |
output_text = gr.Textbox(label="Output Text")
|
| 124 |
preprocess_button = gr.Button("Process and Translate")
|
| 125 |
-
preprocess_button.click(fn=preprocess_and_forward, inputs=
|
| 126 |
with gr.Tab("Translation Only"):
|
| 127 |
translate_input = gr.Textbox(label="Input Text")
|
| 128 |
translate_output = gr.Textbox(label="Output Text")
|
|
|
|
| 14 |
processed_words = []
|
| 15 |
|
| 16 |
for word in words:
|
| 17 |
+
# Check if the word is an acronym (all uppercase letters)
|
| 18 |
if re.match(r"^[A-Z]+$", word):
|
| 19 |
processed_words.append(word) # Leave acronyms unchanged
|
| 20 |
+
# Check if the word has mixed capitalization (e.g., "HEllo")
|
| 21 |
elif re.search(r"[A-Z]", word) and re.search(r"[a-z]", word):
|
| 22 |
processed_words.append(word[0].upper() + word[1:].lower()) # Correct capitalization
|
| 23 |
else:
|
|
|
|
| 25 |
|
| 26 |
return " ".join(processed_words)
|
| 27 |
|
| 28 |
+
def preprocess_text(text: str):
|
| 29 |
"""Process text and return corrections with position information."""
|
| 30 |
result = {
|
| 31 |
"spell_suggestions": [],
|
|
|
|
| 44 |
})
|
| 45 |
text = capitalized_text # Update text for further processing
|
| 46 |
|
| 47 |
+
# Transformer spell check - only for words that look misspelled
|
| 48 |
+
spell_checked = spell_checker(text, max_length=512)[0]['generated_text']
|
| 49 |
+
if spell_checked != text:
|
| 50 |
+
spell_words = spell_checked.split()
|
| 51 |
+
text_words = text.split()
|
| 52 |
+
|
| 53 |
+
# Only suggest spell corrections for words that are actually different
|
| 54 |
+
for orig, corrected in zip(text_words, spell_words):
|
| 55 |
+
if orig.lower() != corrected.lower(): # Compare case-insensitive
|
| 56 |
+
result["spell_suggestions"].append({
|
| 57 |
+
"original": orig,
|
| 58 |
+
"corrected": corrected,
|
| 59 |
+
"type": "spell"
|
| 60 |
+
})
|
|
|
|
|
|
|
| 61 |
|
| 62 |
# Add fluency/style suggestions (other suggestions)
|
| 63 |
# Only add if the word isn't already in spell suggestions
|
|
|
|
| 69 |
if word in spell_originals or word.endswith(('.', '?', '!')):
|
| 70 |
continue
|
| 71 |
|
| 72 |
+
# Add style suggestions (like adding punctuation)
|
| 73 |
+
if not word.endswith('!'): # Only suggest adding exclamation if it doesn't already have one
|
| 74 |
result["other_suggestions"].append({
|
| 75 |
"original": word,
|
| 76 |
"corrected": word + "!",
|
|
|
|
| 89 |
client = Client("Frenchizer/space_21")
|
| 90 |
try:
|
| 91 |
translation = client.predict(text)
|
| 92 |
+
# Return an empty suggestions object since we're just translating
|
| 93 |
result = {
|
| 94 |
"spell_suggestions": [],
|
| 95 |
"other_suggestions": [],
|
|
|
|
| 100 |
except Exception as e:
|
| 101 |
return f"Error: {str(e)}", {}
|
| 102 |
|
| 103 |
+
def preprocess_and_forward(text: str):
|
| 104 |
"""Process text and forward to translation service."""
|
| 105 |
+
original_text, preprocessing_result = preprocess_text(text)
|
| 106 |
|
| 107 |
+
# Forward original text to translation service
|
| 108 |
client = Client("Frenchizer/space_21")
|
| 109 |
try:
|
| 110 |
translation = client.predict(original_text)
|
|
|
|
| 121 |
with gr.Blocks() as demo:
|
| 122 |
with gr.Tab("Main"):
|
| 123 |
input_text = gr.Textbox(label="Input Text")
|
|
|
|
| 124 |
output_text = gr.Textbox(label="Output Text")
|
| 125 |
preprocess_button = gr.Button("Process and Translate")
|
| 126 |
+
preprocess_button.click(fn=preprocess_and_forward, inputs=input_text, outputs=output_text)
|
| 127 |
with gr.Tab("Translation Only"):
|
| 128 |
translate_input = gr.Textbox(label="Input Text")
|
| 129 |
translate_output = gr.Textbox(label="Output Text")
|