Spaces:
Running
Running
| import os, time, requests, gradio as gr | |
| print("Gradio version:", gr.version) | |
| # ---------- Environment Variables ---------- | |
| ASR_API_URL = os.getenv("ASR_API_URL") | |
| AUTH_TOKEN = os.getenv("AUTH_TOKEN") | |
| if not ASR_API_URL or not AUTH_TOKEN: | |
| print("⚠️ ASR_API_URL or AUTH_TOKEN is not set; API calls will fail.") | |
| # ---------- Core Transcription Function ---------- | |
| def transcribe_audio(file_path: str): | |
| if not ASR_API_URL or not AUTH_TOKEN: | |
| return "❌ Error: ASR_API_URL or AUTH_TOKEN is not set.", "" | |
| headers = { | |
| "accept": "application/json", | |
| "Authorization": f"Bearer {AUTH_TOKEN}", | |
| } | |
| start = time.time() | |
| try: | |
| with open(file_path, "rb") as f: | |
| files = {"file": (os.path.basename(file_path), f, "audio/mpeg")} | |
| resp = requests.post(ASR_API_URL, headers=headers, files=files, timeout=120) | |
| except Exception as e: | |
| return f"❌ Error while calling ASR API: {e}", "" | |
| elapsed = time.time() - start | |
| if resp.status_code == 200: | |
| data = resp.json() | |
| text = data.get("transcription", "No transcription returned.") | |
| return text, f"{data.get('time', elapsed):.2f} s" | |
| return f"❌ Error: {resp.status_code}, {resp.text}", "" | |
| # ---------- Styling ---------- | |
| VIOLET_MAIN = "#7F3FBF" | |
| VIOLET_LIGHT = "#C3A6FF" | |
| custom_css = f""" | |
| #gooya-title {{ | |
| color:#fff; | |
| background:linear-gradient(90deg,{VIOLET_MAIN} 0%,{VIOLET_LIGHT} 100%); | |
| border-radius:12px;padding:20px 10px;margin-bottom:12px; | |
| }} | |
| .gooya-badge {{ | |
| display:inline-block;background:{VIOLET_MAIN};color:#fff; | |
| border-radius:16px;padding:6px 16px;font-size:.97rem;margin-top:4px; | |
| }} | |
| """ | |
| # ---------- UI ---------- | |
| with gr.Blocks(css=custom_css, title="Gooya ASR v1.4") as demo: | |
| with gr.Row(): | |
| with gr.Column(): | |
| audio_input = gr.Audio( | |
| label="Audio Input (upload or record, up to 30 s)", | |
| type="filepath", | |
| sources=["upload", "microphone"], | |
| ) | |
| with gr.Column(): | |
| processing_time_tb = gr.Textbox( | |
| label="⏱️ Processing Time", | |
| interactive=False, | |
| elem_classes="gooya-badge", | |
| ) | |
| transcription_tb = gr.Textbox( | |
| label="📝 Transcription", | |
| lines=5, | |
| show_copy_button=True, | |
| placeholder="The transcription will appear here...", | |
| elem_id="gooya-textbox", | |
| ) | |
| with gr.Row(): | |
| btn_transcribe = gr.Button("Transcribe", variant="primary") | |
| btn_clear = gr.Button("Clear", variant="secondary") | |
| gr.Markdown( | |
| """ | |
| Guidelines | |
| - Maximum audio length: 30 seconds | |
| - Audio content should be in Persian. | |
| - Both transcription and processing time are displayed immediately. | |
| See the [Persian ASR Leaderboard](https://huggingface.co/spaces/navidved/open_persian_asr_leaderboard) for benchmarks. | |
| """ | |
| ) | |
| # ---------- Callbacks ---------- | |
| btn_transcribe.click( | |
| fn=transcribe_audio, | |
| inputs=[audio_input], | |
| outputs=[transcription_tb, processing_time_tb], | |
| ) | |
| btn_clear.click( | |
| lambda: ("", "", None), | |
| inputs=None, | |
| outputs=[transcription_tb, processing_time_tb, audio_input], | |
| ) | |
| # ---------- Launch ---------- | |
| if name == "main": | |
| demo.queue().launch(debug=True, share=False) |