app

app.py

@@ -0,0 +1,86 @@
+import gradio as gr
+from rapidfuzz import fuzz, process
+
+
+def highlight_fuzzy_diff(user_text, original_text):
+    user_words = user_text.strip().split()
+    original_words = original_text.strip().split()
+    highlighted = []
+
+    used_indices = set()
+
+    for word in user_words:
+        match, score, idx = process.extractOne(word, original_words, scorer=fuzz.ratio)
+        if score > 85 and idx not in used_indices:
+            highlighted.append(
+                f"<span style='background-color:#f0f0f0; color:#333; padding:4px 10px; "
+                f"margin:4px; border-radius:999px; font-weight:500; display:inline-block;'>{word}</span>"
+            )
+            used_indices.add(idx)
+        else:
+            # added or incorrect word
+            highlighted.append(
+                f"<span style='background-color:#e6ffe6; color:#006600; padding:4px 10px; "
+                f"margin:4px; border-radius:999px; font-weight:500; display:inline-block;'>{word}</span>"
+            )
+
+    # find missing words
+    for i, word in enumerate(original_words):
+        if i not in used_indices:
+            highlighted.append(
+                f"<span style='background-color:#ffe6e6; color:#990000; padding:4px 10px; "
+                f"margin:4px; border-radius:999px; font-weight:500; display:inline-block;'>{word}</span>"
+            )
+
+    return " ".join(highlighted)
+
+
+def check_transcription(audio_file, user_text):
+    transcript_path = audio_file.replace(".mp3", ".txt").replace("audio", "transcripts")
+    with open(transcript_path, 'r') as f:
+        original = f.read().strip()
+
+    diff_html = highlight_fuzzy_diff(user_text, original)
+    score = round(fuzz.ratio(user_text, original))
+
+    # Emoji based on score
+    if score > 90:
+        emoji = "✅"
+        comment = "Great job! Keep it up."
+    elif score > 70:
+        emoji = "⚠️"
+        comment = "Not bad, but pay attention to some words."
+    else:
+        emoji = "❌"
+        comment = "Try again. There are too many mistakes."
+
+    feedback = f"""
+    <div style='font-size: 16px; margin-bottom: 6px;'>
+        Similarity: <b>{score}%</b> {emoji}
+    </div>
+    <div style='margin-top: 4px;'>{comment}</div>
+    """
+
+    return original, (
+            "<div style='margin-bottom: 10px;'><b>🟩 Extra / incorrect words</b>, "
+            "<b>🟥 Missing words</b>, <b>⬜ Correct words</b></div>" + diff_html
+    ), feedback
+
+
+audio_list = ["audio/lesson1.mp3", "audio/lesson2.mp3", "audio/lesson3.mp3"]
+
+gr.Interface(
+    fn=check_transcription,
+    inputs=[
+        gr.Dropdown(audio_list, label="Select an audio file"),
+        gr.Textbox(lines=6, label="Your transcription"),
+    ],
+    outputs=[
+        gr.Textbox(label="Original text"),
+        gr.HTML(label="Comparison (highlighted)"),
+        gr.HTML(label="Feedback and score"),
+    ],
+    title="🎧 English Transcription Checker",
+    description="Listen to the audio, type your transcription, and check how accurate it is.",
+    theme="soft"  # or "default" / "gradio/soft"
+).launch()

audio_generator.py

@@ -0,0 +1,21 @@
+import os
+
+from gtts import gTTS
+
+demo_texts = {
+    "lesson1": "Hello! My name is Anna. I live in London and I work as a teacher.",
+    "lesson2": "Every morning, I wake up at 7 o'clock and have a cup of coffee.",
+    "lesson3": "On weekends, I like to go to the park and read books in the sun."
+}
+
+os.makedirs("audio", exist_ok=True)
+os.makedirs("transcripts", exist_ok=True)
+
+for name, text in demo_texts.items():
+    with open(f"transcripts/{name}.txt", "w") as f:
+        f.write(text)
+
+    tts = gTTS(text)
+    tts.save(f"audio/{name}.mp3")
+
+print("Audio files and transcripts generated successfully.")

requirements.txt

@@ -0,0 +1 @@
+pip~=24.3.1

transcripts/lesson1.txt

@@ -0,0 +1 @@
+Hello! My name is Anna. I live in London and I work as a teacher.

transcripts/lesson2.txt

@@ -0,0 +1 @@
+Every morning, I wake up at 7 o'clock and have a cup of coffee.

transcripts/lesson3.txt

@@ -0,0 +1 @@
+On weekends, I like to go to the park and read books in the sun.