Update app.py

app.py

@@ -1,94 +1,18 @@
-# app.py - Speech to ASL Avatar on Hugging Face Spaces
-
 import gradio as gr
-import whisper
-import requests
-import tempfile
-import os
-
-# Load API key from HF Space secrets (set in Settings → Secrets)
-API_KEY = os.environ.get("SIGN_SPEAK_API_KEY")
-if not API_KEY:
-    raise ValueError("SIGN_SPEAK_API_KEY not set in Space secrets!")
-
-BASE_URL = "https://api.sign-speak.com"
-PRODUCE_SIGN_URL = f"{BASE_URL}/produce-sign"
-
-def get_sign_language(text: str, request_class="BLOCKING", identity="MALE"):
-    headers = {
-        "X-api-key": API_KEY,
-        "Content-Type": "application/json"
-    }
-    payload = {
-        "english": text.strip(),
-        "request_class": request_class.upper(),
-        "identity": identity.upper(),
-        # Optional: add "model_version": "SLP.2.xs" for smaller/faster if needed
-    }
-    response = requests.post(PRODUCE_SIGN_URL, json=payload, headers=headers)
-    
-    if response.status_code == 200:
-        # Save MP4 bytes to temporary file (Gradio Video needs filepath)
-        with tempfile.NamedTemporaryFile(delete=False, suffix=".mp4") as tmp:
-            tmp.write(response.content)
-            return tmp.name
-    elif response.status_code == 202:
-        data = response.json()
-        batch_id = data.get("batch_id")
-        raise ValueError(f"Batch processing started (ID: {batch_id}). Video will be ready later – check logs or add polling.")
-    else:
-        raise ValueError(f"Sign-Speak API error {response.status_code}: {response.text}")
 
-def transcribe_and_translate(audio_filepath):
-    if audio_filepath is None:
-        return "No audio recorded.", None
-    
-    try:
-        # Load Whisper model (use "base" or "small" if "medium" is too slow on CPU)
-        model = whisper.load_model("small")
-        
-        # Transcribe
-        result = model.transcribe(audio_filepath, language="en")
-        text = result["text"].strip()
-        
-        if not text:
-            return "No speech detected in the recording.", None
-        
-        # Get ASL avatar video
-        video_path = get_sign_language(text)
-        
-        return f"Transcribed: \"{text}\"", video_path
-    
-    except Exception as e:
-        return f"Error: {str(e)}", None
+def greet(audio):
+    return "Audio received! (length: {} seconds)".format(len(audio) if audio else 0)
 
-# Gradio UI
-with gr.Blocks(title="Speech → ASL Avatar Translator") as demo:
-    gr.Markdown("""
-    # Speech to ASL Avatar
-    1. Record your voice using the microphone below  
-    2. Click **Translate**  
-    3. Whisper transcribes → Sign-Speak generates ASL signing video  
-    """)
-    
-    with gr.Row():
-        audio_input = gr.Audio(
-            sources=["microphone"],          # ← Fixed: "sources" (list), not "source"
-            type="filepath",
-            label="Speak here (click record)",
-            format="wav"                     # Helps Whisper compatibility
-        )
-        submit_btn = gr.Button("Translate", variant="primary")
-    
-    transcript_output = gr.Textbox(label="Transcribed Text / Status", lines=3)
-    video_output = gr.Video(label="ASL Avatar Signing Video", autoplay=True)
-    
-    # Wire up the button
-    submit_btn.click(
-        fn=transcribe_and_translate,
-        inputs=audio_input,
-        outputs=[transcript_output, video_output]
-    )
+demo = gr.Interface(
+    fn=greet,
+    inputs=gr.Audio(
+        sources=["microphone"],
+        type="numpy",           # or "filepath"
+        label="Record something",
+        format="wav"
+    ),
+    outputs="text",
+    title="Mic Test"
+)
 
-# Launch (HF Spaces ignores server_name/port)
 demo.launch()