Update app.py

app.py

@@ -1,105 +1,87 @@
-import os, time, requests, gradio as gr
+import gradio as gr
+import requests
+import os
+import time
 
-print("Gradio version:", gr.version)
+ASR_API_URL = os.getenv('ASR_API_URL')
+AUTH_TOKEN = os.getenv('AUTH_TOKEN')
 
-# ---------- 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):
+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")}
-            resp = requests.post(ASR_API_URL, headers=headers, files=files, timeout=120)
+        response = requests.post(ASR_API_URL, headers=headers, files=files)
     except Exception as e:
-        return f"❌ Error while calling ASR API: {e}", ""
+        return f"❌ Error: {str(e)}", ""
+    inference_time = time.time() - start_time
 
-    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}", ""
+    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
+    else:
+        return f"❌ Error: {response.status_code}, {response.text}", ""
 
-# ---------- Styling ----------
-VIOLET_MAIN = "#7F3FBF"  
-VIOLET_LIGHT = "#C3A6FF"
+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>""")
 
-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)",
+            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():
-            processing_time_tb = gr.Textbox(
-                label="⏱️ Processing Time",
-                interactive=False,
-                elem_classes="gooya-badge",
-            )
-            transcription_tb = gr.Textbox(
+            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():
-        btn_transcribe = gr.Button("Transcribe", variant="primary")
-        btn_clear      = gr.Button("Clear",      variant="secondary")
+        submit_btn = gr.Button("Transcribe", variant="primary")
+        clear_btn = 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.  
+    gr.Markdown("""
+**Instructions:**  
+- Maximum audio length: **30 seconds**
+- Input audio should be in Persian.
+- The transcription and processing time will be displayed instantly.
 
-See the [Persian ASR Leaderboard](https://huggingface.co/spaces/navidved/open_persian_asr_leaderboard) for benchmarks.
-"""
-    )
+For performance benchmarks, visit: [Persian ASR Leaderboard](https://huggingface.co/spaces/navidved/open_persian_asr_leaderboard)
+""")
 
-    # ---------- Callbacks ----------
-    btn_transcribe.click(
-        fn=transcribe_audio,
-        inputs=[audio_input],
-        outputs=[transcription_tb, processing_time_tb],
+    submit_btn.click(
+        transcribe_audio,
+        inputs=audio,
+        outputs=[transcription, inference_time]
     )
-
-    btn_clear.click(
-        lambda: ("", "", None),
-        inputs=None,
-        outputs=[transcription_tb, processing_time_tb, audio_input],
+    clear_btn.click(
+        lambda: ("", ""),
+        None,
+        [transcription, inference_time, audio]
     )
 
-# ---------- Launch ----------
-if name == "main":
-    demo.queue().launch(debug=True, share=False)
+demo.launch(share=True)