Create ADI voice persona lab Space

.gitattributes

@@ -3,20 +3,16 @@
 *.bin filter=lfs diff=lfs merge=lfs -text
 *.bz2 filter=lfs diff=lfs merge=lfs -text
 *.ckpt filter=lfs diff=lfs merge=lfs -text
-*.ftz filter=lfs diff=lfs merge=lfs -text
 *.gz filter=lfs diff=lfs merge=lfs -text
 *.h5 filter=lfs diff=lfs merge=lfs -text
 *.joblib filter=lfs diff=lfs merge=lfs -text
 *.lfs.* filter=lfs diff=lfs merge=lfs -text
-*.mlmodel filter=lfs diff=lfs merge=lfs -text
 *.model filter=lfs diff=lfs merge=lfs -text
 *.msgpack filter=lfs diff=lfs merge=lfs -text
 *.npy filter=lfs diff=lfs merge=lfs -text
 *.npz filter=lfs diff=lfs merge=lfs -text
 *.onnx filter=lfs diff=lfs merge=lfs -text
-*.ot filter=lfs diff=lfs merge=lfs -text
 *.parquet filter=lfs diff=lfs merge=lfs -text
-*.pb filter=lfs diff=lfs merge=lfs -text
 *.pickle filter=lfs diff=lfs merge=lfs -text
 *.pkl filter=lfs diff=lfs merge=lfs -text
 *.pt filter=lfs diff=lfs merge=lfs -text
@@ -33,3 +29,5 @@ saved_model/**/* filter=lfs diff=lfs merge=lfs -text
 *.zip filter=lfs diff=lfs merge=lfs -text
 *.zst filter=lfs diff=lfs merge=lfs -text
 *tfevents* filter=lfs diff=lfs merge=lfs -text
+*.mp3 filter=lfs diff=lfs merge=lfs -text
+*.wav filter=lfs diff=lfs merge=lfs -text

README.md

@@ -1,13 +1,11 @@
 ---
-title: Adi Voice Persona Lab
-emoji: 🐢
-colorFrom: yellow
+title: ADI Voice Persona Lab
+emoji: 🎛️
+colorFrom: gray
 colorTo: green
 sdk: gradio
 sdk_version: 6.19.0
-python_version: '3.13'
+python_version: '3.12'
 app_file: app.py
 pinned: false
 ---
-
-Check out the configuration reference at https://huggingface.co/docs/hub/spaces-config-reference

app.py

@@ -0,0 +1,162 @@
+import asyncio
+import re
+import tempfile
+import zipfile
+from pathlib import Path
+
+import edge_tts
+import gradio as gr
+
+
+PERSONAS = {
+    "ADI calm": {
+        "prefix": "Here is the calm version.",
+        "style": "steady, concise, and reassuring",
+    },
+    "Technical mentor": {
+        "prefix": "Here is the technical mentor version.",
+        "style": "precise, capable, and lightly explanatory",
+    },
+    "Operator mode": {
+        "prefix": "Here is the operator mode version.",
+        "style": "brief, action-oriented, and status-aware",
+    },
+    "Warm companion": {
+        "prefix": "Here is the warm companion version.",
+        "style": "friendly, grounded, and conversational",
+    },
+    "Launch demo": {
+        "prefix": "Here is the launch demo version.",
+        "style": "polished, confident, and presentation-ready",
+    },
+}
+
+VOICE_CHOICES = [
+    "en-US-AriaNeural",
+    "en-US-JennyNeural",
+    "en-US-GuyNeural",
+    "en-US-DavisNeural",
+    "en-US-AvaNeural",
+    "en-GB-SoniaNeural",
+    "en-GB-RyanNeural",
+]
+
+
+def clean(text):
+    return re.sub(r"\s+", " ", (text or "").strip())
+
+
+def rewrite_for_persona(text, persona, brevity):
+    base = clean(text)
+    if not base:
+        raise gr.Error("Add text for ADI to speak first.")
+
+    style = PERSONAS[persona]["style"]
+    prefix = PERSONAS[persona]["prefix"]
+    sentences = re.split(r"(?<=[.!?])\s+", base)
+    sentences = [sentence for sentence in sentences if sentence]
+
+    if brevity == "short":
+        body = " ".join(sentences[:2])
+    elif brevity == "expanded":
+        body = base
+        body += f" The intended delivery is {style}, with enough detail to feel useful."
+    else:
+        body = base
+
+    if persona == "Operator mode":
+        body = f"Status: ready. {body} Next step: tell me what you want to run."
+    elif persona == "Technical mentor":
+        body = f"{body} The key idea is to make the next action obvious."
+    elif persona == "Warm companion":
+        body = f"{body} I can stay with you while we tune it."
+    elif persona == "Launch demo":
+        body = f"{body} This is ready to show as a live ADI capability."
+
+    return f"{prefix} {body}"
+
+
+async def synthesize_one(text, voice, rate, pitch):
+    output = Path(tempfile.NamedTemporaryFile(suffix=".mp3", delete=False).name)
+    communicate = edge_tts.Communicate(text, voice, rate=rate, pitch=pitch)
+    await communicate.save(str(output))
+    return str(output)
+
+
+def synthesize(text, persona, voice, rate, pitch, brevity):
+    script = rewrite_for_persona(text, persona, brevity)
+    audio_path = asyncio.run(synthesize_one(script, voice, rate, pitch))
+    return script, audio_path
+
+
+def compare(text, voice_a, persona_a, voice_b, persona_b, voice_c, persona_c, rate, pitch, brevity):
+    script_a, audio_a = synthesize(text, persona_a, voice_a, rate, pitch, brevity)
+    script_b, audio_b = synthesize(text, persona_b, voice_b, rate, pitch, brevity)
+    script_c, audio_c = synthesize(text, persona_c, voice_c, rate, pitch, brevity)
+
+    zip_path = Path(tempfile.NamedTemporaryFile(suffix=".zip", delete=False).name)
+    with zipfile.ZipFile(zip_path, "w", zipfile.ZIP_DEFLATED) as bundle:
+        for idx, (script, audio) in enumerate(((script_a, audio_a), (script_b, audio_b), (script_c, audio_c)), 1):
+            bundle.write(audio, f"adi_voice_take_{idx}.mp3")
+            bundle.writestr(f"adi_voice_take_{idx}.txt", script)
+
+    return script_a, audio_a, script_b, audio_b, script_c, audio_c, str(zip_path)
+
+
+with gr.Blocks(title="ADI Voice Persona Lab", fill_width=True) as demo:
+    gr.Markdown("# ADI Voice Persona Lab")
+    source = gr.Textbox(
+        label="Base ADI response",
+        lines=6,
+        value=(
+            "Hello, I am ADI. I can help inspect models, compare prompts, generate training data, "
+            "and turn your ideas into working demos on Hugging Face Spaces."
+        ),
+    )
+
+    with gr.Row():
+        rate = gr.Dropdown(
+            choices=["-20%", "-10%", "+0%", "+10%", "+20%"],
+            value="+0%",
+            label="Rate",
+        )
+        pitch = gr.Dropdown(
+            choices=["-10Hz", "+0Hz", "+10Hz", "+20Hz"],
+            value="+0Hz",
+            label="Pitch",
+        )
+        brevity = gr.Radio(
+            choices=["short", "normal", "expanded"],
+            value="normal",
+            label="Script length",
+        )
+
+    with gr.Row():
+        with gr.Column():
+            persona_a = gr.Dropdown(list(PERSONAS), value="ADI calm", label="Persona A")
+            voice_a = gr.Dropdown(VOICE_CHOICES, value="en-US-AriaNeural", label="Voice A")
+            script_a = gr.Textbox(label="Script A", lines=7)
+            audio_a = gr.Audio(label="Audio A", type="filepath")
+        with gr.Column():
+            persona_b = gr.Dropdown(list(PERSONAS), value="Technical mentor", label="Persona B")
+            voice_b = gr.Dropdown(VOICE_CHOICES, value="en-US-GuyNeural", label="Voice B")
+            script_b = gr.Textbox(label="Script B", lines=7)
+            audio_b = gr.Audio(label="Audio B", type="filepath")
+        with gr.Column():
+            persona_c = gr.Dropdown(list(PERSONAS), value="Launch demo", label="Persona C")
+            voice_c = gr.Dropdown(VOICE_CHOICES, value="en-US-AvaNeural", label="Voice C")
+            script_c = gr.Textbox(label="Script C", lines=7)
+            audio_c = gr.Audio(label="Audio C", type="filepath")
+
+    run = gr.Button("Generate Takes", variant="primary")
+    bundle = gr.File(label="Download all takes")
+
+    run.click(
+        compare,
+        inputs=[source, voice_a, persona_a, voice_b, persona_b, voice_c, persona_c, rate, pitch, brevity],
+        outputs=[script_a, audio_a, script_b, audio_b, script_c, audio_c, bundle],
+    )
+
+
+if __name__ == "__main__":
+    demo.launch()

requirements.txt

@@ -0,0 +1,2 @@
+gradio==6.19.0
+edge-tts==7.2.8