Spaces:
Running
Running
Update app.py
Browse files
app.py
CHANGED
@@ -7,27 +7,19 @@ ASR_API_URL = os.getenv("ASR_API_URL")
|
|
7 |
AUTH_TOKEN = os.getenv("AUTH_TOKEN")
|
8 |
|
9 |
def transcribe_audio(file_path: str):
|
10 |
-
"""
|
11 |
-
دریافت مسیر فایل صوتی (بهدلیل type='filepath') و برگرداندن
|
12 |
-
رونوشت و مدت زمان پردازش.
|
13 |
-
"""
|
14 |
if not ASR_API_URL or not AUTH_TOKEN:
|
15 |
return "❌ Error: ASR_API_URL or AUTH_TOKEN is not set.", ""
|
16 |
-
|
17 |
headers = {
|
18 |
"accept": "application/json",
|
19 |
"Authorization": f"Bearer {AUTH_TOKEN}",
|
20 |
}
|
21 |
files = {"file": (file_path, open(file_path, "rb"), "audio/mpeg")}
|
22 |
-
|
23 |
start = time.time()
|
24 |
try:
|
25 |
resp = requests.post(ASR_API_URL, headers=headers, files=files)
|
26 |
except Exception as e:
|
27 |
return f"❌ Error: {e}", ""
|
28 |
-
|
29 |
elapsed = time.time() - start
|
30 |
-
|
31 |
if resp.status_code == 200:
|
32 |
data = resp.json()
|
33 |
transcript = data.get("transcription", "No transcription returned.")
|
@@ -36,79 +28,21 @@ def transcribe_audio(file_path: str):
|
|
36 |
else:
|
37 |
return f"❌ Error: {resp.status_code}, {resp.text}", ""
|
38 |
|
39 |
-
|
40 |
-
|
41 |
-
|
42 |
-
|
43 |
-
|
44 |
-
|
45 |
-
|
46 |
-
|
47 |
-
""
|
48 |
-
|
49 |
-
|
50 |
-
|
51 |
-
""
|
52 |
-
<div id="gooya-title">
|
53 |
-
<h1 style='margin-bottom:10px;font-weight:800;font-size:2rem;'>
|
54 |
-
Gooya ASR <span style="font-size:1.1rem;font-weight:400;opacity:.8;">v1.4</span>
|
55 |
-
</h1>
|
56 |
-
<p style='font-size:1.12rem;margin-bottom:2px;'>
|
57 |
-
High-performance Persian Speech-to-Text
|
58 |
-
</p>
|
59 |
-
<p style='font-size:.98rem;color:#c6e8fa'>
|
60 |
-
Upload or record a Persian audio file (max 30 s) and instantly receive the transcription.
|
61 |
-
</p>
|
62 |
-
</div>
|
63 |
-
"""
|
64 |
-
)
|
65 |
-
|
66 |
-
with gr.Row():
|
67 |
-
with gr.Column():
|
68 |
-
audio = gr.Audio(
|
69 |
-
label="Audio Input (upload or record, up to 30 s)",
|
70 |
-
type="filepath",
|
71 |
-
sources=["upload", "microphone"],
|
72 |
-
)
|
73 |
-
with gr.Column():
|
74 |
-
inference_time = gr.Label(
|
75 |
-
label="⏱️ Processing Time", elem_classes="gooya-badge"
|
76 |
-
)
|
77 |
-
transcription = gr.Textbox(
|
78 |
-
label="📝 Transcription",
|
79 |
-
lines=5,
|
80 |
-
show_copy_button=True,
|
81 |
-
placeholder="The transcription will appear here...",
|
82 |
-
elem_id="gooya-textbox",
|
83 |
-
)
|
84 |
-
|
85 |
-
with gr.Row():
|
86 |
-
btn_transcribe = gr.Button("Transcribe", variant="primary")
|
87 |
-
btn_clear = gr.Button("Clear", variant="secondary")
|
88 |
-
|
89 |
-
gr.Markdown(
|
90 |
-
"""
|
91 |
-
**دستورالعملها**
|
92 |
-
- حداکثر طول صدا: **۳۰ ثانیه**
|
93 |
-
- صدا باید فارسی باشد.
|
94 |
-
- نتیجهی رونویسی و زمان پردازش بلافاصله نمایش داده میشود.
|
95 |
-
|
96 |
-
برای مشاهده بنچمارکها به [Persian ASR Leaderboard](https://huggingface.co/spaces/navidved/open_persian_asr_leaderboard) مراجعه کنید.
|
97 |
-
"""
|
98 |
-
)
|
99 |
-
|
100 |
-
btn_transcribe.click(
|
101 |
-
transcribe_audio,
|
102 |
-
inputs=audio,
|
103 |
-
outputs=[transcription, inference_time],
|
104 |
-
)
|
105 |
-
|
106 |
-
btn_clear.click(
|
107 |
-
lambda: ("", "", None),
|
108 |
-
inputs=None,
|
109 |
-
outputs=[transcription, inference_time, audio],
|
110 |
)
|
|
|
111 |
|
112 |
if __name__ == "__main__":
|
113 |
-
|
114 |
-
demo.launch(share=True)
|
|
|
7 |
AUTH_TOKEN = os.getenv("AUTH_TOKEN")
|
8 |
|
9 |
def transcribe_audio(file_path: str):
|
|
|
|
|
|
|
|
|
10 |
if not ASR_API_URL or not AUTH_TOKEN:
|
11 |
return "❌ Error: ASR_API_URL or AUTH_TOKEN is not set.", ""
|
|
|
12 |
headers = {
|
13 |
"accept": "application/json",
|
14 |
"Authorization": f"Bearer {AUTH_TOKEN}",
|
15 |
}
|
16 |
files = {"file": (file_path, open(file_path, "rb"), "audio/mpeg")}
|
|
|
17 |
start = time.time()
|
18 |
try:
|
19 |
resp = requests.post(ASR_API_URL, headers=headers, files=files)
|
20 |
except Exception as e:
|
21 |
return f"❌ Error: {e}", ""
|
|
|
22 |
elapsed = time.time() - start
|
|
|
23 |
if resp.status_code == 200:
|
24 |
data = resp.json()
|
25 |
transcript = data.get("transcription", "No transcription returned.")
|
|
|
28 |
else:
|
29 |
return f"❌ Error: {resp.status_code}, {resp.text}", ""
|
30 |
|
31 |
+
def main():
|
32 |
+
iface = gr.Interface(
|
33 |
+
fn=transcribe_audio,
|
34 |
+
inputs=gr.Audio(label="Audio Input (upload or record, up to 30 s)", type="filepath", sources=["upload", "microphone"]),
|
35 |
+
outputs=[
|
36 |
+
gr.Textbox(label="📝 Transcription", lines=5, show_copy_button=True, placeholder="The transcription will appear here..."),
|
37 |
+
gr.Label(label="⏱️ Processing Time")
|
38 |
+
],
|
39 |
+
title="Gooya ASR v1.4",
|
40 |
+
description="High-performance Persian Speech-to-Text. Upload or record a Persian audio file (max 30 s) and instantly receive the transcription.",
|
41 |
+
examples=[],
|
42 |
+
cache_examples=False,
|
43 |
+
allow_flagging="never"
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
44 |
)
|
45 |
+
iface.launch()
|
46 |
|
47 |
if __name__ == "__main__":
|
48 |
+
main()
|
|