Update app.py

app.py

@@ -1,87 +1,95 @@
 import gradio as gr
-import requests
-import os
-import time
+import requests, os, time
 
-ASR_API_URL = os.getenv('ASR_API_URL')
-AUTH_TOKEN = os.getenv('AUTH_TOKEN')
+ASR_API_URL = os.getenv("ASR_API_URL")
+AUTH_TOKEN  = os.getenv("AUTH_TOKEN")
 
 def transcribe_audio(file_path):
+    if not file_path:
+        return "❌ Audio Loading ...", ""
     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}',
-    }
-    files = {
-        'file': (file_path, open(file_path, 'rb'), 'audio/mpeg'),
-    }
-    start_time = time.time()
+    
+    headers = {"accept": "application/json",
+               "Authorization": f"Bearer {AUTH_TOKEN}"}
+    files = {"file": (file_path, open(file_path, "rb"), "audio/mpeg")}
+    start = time.time()
     try:
         response = requests.post(ASR_API_URL, headers=headers, files=files)
     except Exception as e:
-        return f"❌ Error: {str(e)}", ""
-    inference_time = time.time() - start_time
-
+        return f"❌ Error: {e}", ""
+    
+    inf_time = time.time() - start
     if response.status_code == 200:
         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
+        text = res.get("transcription", "No transcription returned.")
+        inf_time = f"{res.get('time', inf_time):.2f} seconds"
+        return text, inf_time
     else:
-        return f"❌ Error: {response.status_code}, {response.text}", ""
+        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;}
+#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>
+        <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",
-                show_label=True,
-                sources=["upload", "microphone"]
-            )
+            audio = gr.Audio(label="Audio Input (Upload or record, up to 30s)",
+                             type="filepath",
+                             sources=["upload", "microphone"],
+                             show_label=True)
         with gr.Column():
-            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"
-            )
+            inf_time_lbl  = 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")
 
     with gr.Row():
-        submit_btn = gr.Button("Transcribe", variant="primary")
-        clear_btn = gr.Button("Clear", variant="secondary")
+        submit_btn = gr.Button("Transcribe", variant="primary", interactive=False)
+        clear_btn  = gr.Button("Clear",      variant="secondary")
 
     gr.Markdown("""
 **Instructions:**  
-- Maximum audio length: **30 seconds**
-- Input audio should be in Persian.
-- The transcription and processing time will be displayed instantly.
+- Maximum audio length: **30 seconds**  
+- Input audio should be in Persian.  
 
-For performance benchmarks, visit: [Persian ASR Leaderboard](https://huggingface.co/spaces/navidved/open_persian_asr_leaderboard)
+For benchmarks visit: [Persian ASR Leaderboard](https://huggingface.co/spaces/navidved/open_persian_asr_leaderboard)
 """)
 
+    def toggle_btn(path):
+        return gr.Button.update(interactive=bool(path))
+
+    audio.change(toggle_btn, inputs=audio, outputs=submit_btn)
+
     submit_btn.click(
         transcribe_audio,
         inputs=audio,
-        outputs=[transcription, inference_time]
+        outputs=[transcription, inf_time_lbl]
+    ).then(  
+        lambda: gr.Button.update(interactive=False),
+        None,
+        submit_btn
     )
+
     clear_btn.click(
-        lambda: ("", ""),
+        lambda: ("", "", None, False),
         None,
-        [transcription, inference_time, audio]
+        [transcription, inf_time_lbl, audio, submit_btn],
     )
 
-demo.launch(share=True)
+demo.launch()