Spaces:
Runtime error
Runtime error
Update app.py
Browse files
app.py
CHANGED
|
@@ -7,7 +7,6 @@ from gtts import gTTS
|
|
| 7 |
import io
|
| 8 |
from pydub import AudioSegment
|
| 9 |
import time
|
| 10 |
-
import pronouncing
|
| 11 |
import epitran
|
| 12 |
|
| 13 |
# Create audio directory if it doesn't exist
|
|
@@ -20,6 +19,27 @@ try:
|
|
| 20 |
except Exception as e:
|
| 21 |
print(f"Error initializing Epitran: {e}")
|
| 22 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 23 |
# Step 1: Transcribe the audio file
|
| 24 |
def transcribe_audio(audio):
|
| 25 |
if audio is None:
|
|
@@ -54,6 +74,13 @@ def transcribe_audio(audio):
|
|
| 54 |
except sr.RequestError as e:
|
| 55 |
return f"Error with Google Speech Recognition service: {e}"
|
| 56 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 57 |
# Step 2: Create pronunciation audio for incorrect words (locally)
|
| 58 |
def create_pronunciation_audio(word):
|
| 59 |
try:
|
|
@@ -64,38 +91,6 @@ def create_pronunciation_audio(word):
|
|
| 64 |
except Exception as e:
|
| 65 |
return f"Failed to create pronunciation audio: {e}"
|
| 66 |
|
| 67 |
-
# Function for phonetic respelling
|
| 68 |
-
def phonetic_respelling(sentence):
|
| 69 |
-
words = sentence.split()
|
| 70 |
-
respelled = []
|
| 71 |
-
|
| 72 |
-
for word in words:
|
| 73 |
-
# Find close matches for each word
|
| 74 |
-
close_matches = pronouncing.search(word)
|
| 75 |
-
if close_matches:
|
| 76 |
-
# Get the first close match
|
| 77 |
-
closest_word = close_matches[0]
|
| 78 |
-
respelled.append(pronouncing.phones_for_word(closest_word)[0]) # Use phonemes for the closest match
|
| 79 |
-
else:
|
| 80 |
-
respelled.append(word)
|
| 81 |
-
|
| 82 |
-
# Convert phonemes to respelling
|
| 83 |
-
respelling = ' '.join(respelled)
|
| 84 |
-
|
| 85 |
-
# Replace phonemes with common respellings
|
| 86 |
-
respelling = respelling.replace('ˈ', '').replace('ˌ', '').replace('ː', '') # Clean up phoneme symbols
|
| 87 |
-
respelling = respelling.replace('ɑ', 'a').replace('ə', 'uh').replace('ɪ', 'i').replace('ʊ', 'u') # Sample conversions
|
| 88 |
-
|
| 89 |
-
return respelling
|
| 90 |
-
|
| 91 |
-
# Function for IPA transcription
|
| 92 |
-
def ipa_transcription(sentence):
|
| 93 |
-
try:
|
| 94 |
-
return epi.transliterate(sentence)
|
| 95 |
-
except Exception as e:
|
| 96 |
-
print(f"Error during IPA transcription: {e}")
|
| 97 |
-
return "IPA transcription failed."
|
| 98 |
-
|
| 99 |
# Step 3: Compare the transcribed text with the input paragraph
|
| 100 |
def compare_texts(reference_text, transcribed_text):
|
| 101 |
reference_words = reference_text.split()
|
|
@@ -118,9 +113,7 @@ def compare_texts(reference_text, transcribed_text):
|
|
| 118 |
|
| 119 |
html_output += f"<strong>Quality Score:</strong> {similarity_score}%<br>"
|
| 120 |
html_output += f"<strong>Transcribed Text:</strong> {transcribed_text}<br>"
|
| 121 |
-
html_output += f"<strong>
|
| 122 |
-
html_output += f"<strong>Phonetic Respelling:</strong> {phonetic_respelling(reference_text)}<br>"
|
| 123 |
-
html_output += f"<strong>IPA Transcription:</strong> {ipa_transcription(reference_text)}<br>"
|
| 124 |
html_output += "<strong>Word Score List:</strong><br>"
|
| 125 |
|
| 126 |
# Generate colored word score list
|
|
|
|
| 7 |
import io
|
| 8 |
from pydub import AudioSegment
|
| 9 |
import time
|
|
|
|
| 10 |
import epitran
|
| 11 |
|
| 12 |
# Create audio directory if it doesn't exist
|
|
|
|
| 19 |
except Exception as e:
|
| 20 |
print(f"Error initializing Epitran: {e}")
|
| 21 |
|
| 22 |
+
# Step 2: Create pronunciation audio for incorrect words
|
| 23 |
+
def upfilepath(local_filename):
|
| 24 |
+
ts = time.time()
|
| 25 |
+
upload_url = f"https://mr2along-speech-recognize.hf.space/gradio_api/upload?upload_id={ts}"
|
| 26 |
+
files = {'files': open(local_filename, 'rb')}
|
| 27 |
+
|
| 28 |
+
try:
|
| 29 |
+
response = requests.post(upload_url, files=files, timeout=30) # Set timeout (e.g., 30 seconds)
|
| 30 |
+
|
| 31 |
+
if response.status_code == 200:
|
| 32 |
+
result = response.json()
|
| 33 |
+
extracted_path = result[0]
|
| 34 |
+
return extracted_path
|
| 35 |
+
else:
|
| 36 |
+
return None
|
| 37 |
+
|
| 38 |
+
except requests.exceptions.Timeout:
|
| 39 |
+
return "Request timed out. Please try again."
|
| 40 |
+
except Exception as e:
|
| 41 |
+
return f"An error occurred: {e}"
|
| 42 |
+
|
| 43 |
# Step 1: Transcribe the audio file
|
| 44 |
def transcribe_audio(audio):
|
| 45 |
if audio is None:
|
|
|
|
| 74 |
except sr.RequestError as e:
|
| 75 |
return f"Error with Google Speech Recognition service: {e}"
|
| 76 |
|
| 77 |
+
# Function to get IPA transcription
|
| 78 |
+
def ipa_transcription(sentence):
|
| 79 |
+
try:
|
| 80 |
+
return epi.transliterate(sentence)
|
| 81 |
+
except Exception as e:
|
| 82 |
+
return f"Error during IPA transcription: {e}"
|
| 83 |
+
|
| 84 |
# Step 2: Create pronunciation audio for incorrect words (locally)
|
| 85 |
def create_pronunciation_audio(word):
|
| 86 |
try:
|
|
|
|
| 91 |
except Exception as e:
|
| 92 |
return f"Failed to create pronunciation audio: {e}"
|
| 93 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 94 |
# Step 3: Compare the transcribed text with the input paragraph
|
| 95 |
def compare_texts(reference_text, transcribed_text):
|
| 96 |
reference_words = reference_text.split()
|
|
|
|
| 113 |
|
| 114 |
html_output += f"<strong>Quality Score:</strong> {similarity_score}%<br>"
|
| 115 |
html_output += f"<strong>Transcribed Text:</strong> {transcribed_text}<br>"
|
| 116 |
+
html_output += f"<strong>IPA Transcription:</strong> {ipa_transcription(reference_text)}<br>" # Display IPA transcription
|
|
|
|
|
|
|
| 117 |
html_output += "<strong>Word Score List:</strong><br>"
|
| 118 |
|
| 119 |
# Generate colored word score list
|