Update app.py

app.py

@@ -1,130 +1,87 @@
 import gradio as gr
-import requests, os, time, pathlib
+import requests
+import os
+import time
 
-ASR_API_URL = os.getenv("ASR_API_URL")
-AUTH_TOKEN  = os.getenv("AUTH_TOKEN")
-
-# ---------- کمکى: استخراج مسیر واقعی فایل ----------
-def extract_path(audio_value):
-    """
-    audio_value می‌تواند None، رشته یا دیکشنری باشد.
-    خروجی: مسیر فایل یا None
-    """
-    if not audio_value:
-        return None
-    if isinstance(audio_value, dict):
-        return audio_value.get("path")            # گرادیو ≥ 4
-    # برای سازگاری با نسخه‌های قدیم‌تر که str برمی‌گردانند
-    return audio_value if isinstance(audio_value, (str, pathlib.Path)) else None
-
-# ---------- پردازش ----------
-def transcribe_audio(audio_value):
-    file_path = extract_path(audio_value)
-    if not file_path:
-        return "❌ فایل صوتی هنوز آماده نیست.", ""
+ASR_API_URL = os.getenv('ASR_API_URL')
+AUTH_TOKEN = os.getenv('AUTH_TOKEN')
 
+def transcribe_audio(file_path):
     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}",
+        'accept': 'application/json',
+        'Authorization': f'Bearer {AUTH_TOKEN}',
     }
-
-    start = time.time()
+    files = {
+        'file': (file_path, open(file_path, 'rb'), 'audio/mpeg'),
+    }
+    start_time = time.time()
     try:
-        with open(file_path, "rb") as f:
-            files = {"file": (os.path.basename(file_path), f, "audio/mpeg")}
-            response = requests.post(ASR_API_URL, headers=headers, files=files)
+        response = requests.post(ASR_API_URL, headers=headers, files=files)
     except Exception as e:
-        return f"❌ Error: {e}", ""
+        return f"❌ Error: {str(e)}", ""
+    inference_time = time.time() - start_time
 
-    elapsed = time.time() - start
     if response.status_code == 200:
-        data = response.json()
-        txt = data.get("transcription", "No transcription returned.")
-        elapsed = data.get("time", elapsed)
-        return txt, f"{elapsed:.2f} seconds"
-
-    return f"❌ Error {response.status_code}: {response.text}", ""
-
-
-# ---------- رابط کاربری ----------
-CUSTOM_CSS = """
-#gooya-title {color:white; background:linear-gradient(90deg,#224CA5 0%,#2CD8D5 100%);
-              border-radius:12px;padding:20px 10px;margin-bottom:12px;}
-.gooya-badge {display:inline-block;background:#224CA5;color:#fff;border-radius:16px;
-              padding:6px 16px;font-size:0.97rem;margin-top:4px;}
-"""
-
-with gr.Blocks(css=CUSTOM_CSS) as demo:
-
-    gr.HTML("""
-    <div id="gooya-title">
-        <h1 style='margin-bottom:10px;font-weight:800;font-size:2rem;'>Gooya ASR
-            <span style="font-size:1.1rem;font-weight:400;opacity:0.8;'>v1.4</span></h1>
-        <p style='font-size:1.12rem;margin-bottom:2px;'>High-performance Persian Speech-to-Text</p>
-        <p style='font-size:0.98rem;color:#c6e8fa'>Upload or record a Persian audio file (max 30s) and instantly receive the transcription.</p>
-    </div>
-    """)
+        res = response.json()
+        transcription = res.get("transcription", "No transcription returned.")
+        inference_time_str = f"{res.get('time', inference_time):.2f} seconds"
+        return transcription, inference_time_str
+    else:
+        return f"❌ Error: {response.status_code}, {response.text}", ""
+
+with gr.Blocks(css="""
+#gooya-title {color:white; background: linear-gradient(90deg, #224CA5 0%, #2CD8D5 100%); border-radius: 12px; padding:20px 10px;margin-bottom:12px;}
+.gooya-badge {display:inline-block; background:#224CA5; color:#fff; border-radius:16px; padding:6px 16px; font-size:0.97rem; margin-top:4px;}
+#gooya-box {background:#F7FAFF; border:1px solid #e7e9ef; border-radius:14px; padding:22px 18px; margin-top:12px;}
+""") as demo:
+    gr.HTML("""<div id="gooya-title">
+    <h1 style='margin-bottom:10px;font-weight:800;font-size:2rem;'>Gooya ASR <span style="font-size:1.1rem; font-weight:400; opacity:0.8;">v1.4</span></h1>
+    <p style='font-size:1.12rem; margin-bottom:2px;'>High-performance Persian Speech-to-Text</p>
+    <p style='font-size:0.98rem; color:#c6e8fa'>Upload or record a Persian audio file (max 30s) and instantly receive the transcription.</p>
+    </div>""")
 
     with gr.Row():
         with gr.Column():
             audio = gr.Audio(
                 label="Audio Input (Upload or record, up to 30s)",
                 type="filepath",
-                sources=["upload", "microphone"],
                 show_label=True,
+                sources=["upload", "microphone"]
             )
         with gr.Column():
-            inf_time_lbl  = gr.Label(label="⏱️ Processing Time", elem_classes="gooya-badge")
+            inference_time = gr.Label(label="⏱️ Processing Time", elem_classes="gooya-badge")
             transcription = gr.Textbox(
                 label="📝 Transcription",
                 lines=5,
                 show_copy_button=True,
                 placeholder="The transcription will appear here...",
-                elem_id="gooya-textbox",
+                elem_id="gooya-textbox"
             )
 
     with gr.Row():
-        submit_btn = gr.Button("Transcribe", variant="primary", interactive=False)
-        clear_btn  = gr.Button("Clear",      variant="secondary")
+        submit_btn = gr.Button("Transcribe", variant="primary")
+        clear_btn = gr.Button("Clear", variant="secondary")
 
-    # ---------- فعال / غیرفعال کردن دکمه ----------
-    def toggle_btn(audio_value):
-        return gr.Button.update(interactive=bool(extract_path(audio_value)))
+    gr.Markdown("""
+**Instructions:**  
+- Maximum audio length: **30 seconds**
+- Input audio should be in Persian.
+- The transcription and processing time will be displayed instantly.
 
-    # بعد از آپلود فایل
-    audio.change(toggle_btn,      inputs=audio, outputs=submit_btn, queue=False)
-    # بعد از پایان ضبط میکروفون
-    audio.stop_recording(toggle_btn, inputs=audio, outputs=submit_btn, queue=False)
+For performance benchmarks, visit: [Persian ASR Leaderboard](https://huggingface.co/spaces/navidved/open_persian_asr_leaderboard)
+""")
 
-    # ---------- پردازش ----------
     submit_btn.click(
         transcribe_audio,
         inputs=audio,
-        outputs=[transcription, inf_time_lbl],
-    ).then(
-        lambda: gr.Button.update(interactive=False),
-        None,
-        submit_btn,
-        queue=False,
+        outputs=[transcription, inference_time]
     )
-
-    # ---------- پاک‌کردن ----------
-    def clear_all():
-        return (
-            "",                                        # متن
-            "",                                        # زمان
-            gr.Audio.update(value=None),               # پاک کردن کامپوننت Audio
-            gr.Button.update(interactive=False),       # غیر فعال کردن دکمه
-        )
-
     clear_btn.click(
-        clear_all,
+        lambda: ("", ""),
         None,
-        [transcription, inf_time_lbl, audio, submit_btn],
-        queue=False,
+        [transcription, inference_time, audio]
     )
 
-demo.launch()
+demo.launch(share=True)