Create app.py

app.py

@@ -0,0 +1,75 @@
+#@markdown Output: Accuracy Score
+
+import gradio as gr
+import speech_recognition as sr
+from Levenshtein import ratio
+import tempfile
+import numpy as np
+import soundfile as sf
+import pandas as pd
+
+# Sample dataframe with sentences ordered from easy to hard
+data = {
+    "Sentences": [
+        "A stitch in time saves nine.",
+        "To be or not to be, that is the question.",
+        "Five cats were living in safe caves.",
+        "Hives give shelter to bees in large caves.",
+        "His decision to plant a rose was amazing.",
+        "She sells sea shells by the sea shore.",
+        "The colorful parrot likes rolling berries.",
+        "Time flies like an arrow; fruit flies like a banana.",
+        "Good things come to those who wait.",
+        "All human beings are born free and equal in dignity and rights."
+    ]
+}
+df = pd.DataFrame(data)
+
+def transcribe_audio(file_info):
+    r = sr.Recognizer()
+    with tempfile.NamedTemporaryFile(delete=True, suffix=".wav") as tmpfile:
+        sf.write(tmpfile.name, data=file_info[1], samplerate=44100, format='WAV')
+        tmpfile.seek(0)
+        with sr.AudioFile(tmpfile.name) as source:
+            audio_data = r.record(source)
+    try:
+        text = r.recognize_google(audio_data)
+        return text
+    except sr.UnknownValueError:
+        return "Could not understand audio"
+    except sr.RequestError as e:
+        return f"Could not request results; {e}"
+
+def pronunciation_correction(expected_text, file_info):
+    user_spoken_text = transcribe_audio(file_info)
+    similarity = ratio(expected_text.lower(), user_spoken_text.lower())
+    description = f"{similarity:.2f}"
+
+    if similarity >= 0.9:
+        feedback = "Excellent pronunciation!"
+    elif similarity >= 0.7:
+        feedback = "Good pronunciation!"
+    elif similarity >= 0.5:
+        feedback = "Needs improvement."
+    else:
+        feedback = "Poor pronunciation, try to focus more on clarity."
+
+    return feedback, description
+
+with gr.Blocks() as app:
+    with gr.Row():
+        sentence_dropdown = gr.Dropdown(choices=df['Sentences'].tolist(), label="Select a Sentence")
+        selected_sentence_output = gr.Textbox(label="Selected Text", interactive=False)
+    audio_input = gr.Audio(label="Upload Audio File", type="numpy")
+    check_pronunciation_button = gr.Button("Check Pronunciation")
+    pronunciation_feedback = gr.Textbox(label="Pronunciation Feedback")
+    pronunciation_score = gr.Number(label="Pronunciation Accuracy Score: 0 (No Match) ~ 1 (Perfect)")
+
+    sentence_dropdown.change(lambda x: x, inputs=sentence_dropdown, outputs=selected_sentence_output)
+    check_pronunciation_button.click(
+        pronunciation_correction,
+        inputs=[sentence_dropdown, audio_input],
+        outputs=[pronunciation_feedback, pronunciation_score]
+    )
+
+app.launch(debug=True)