dora-vad 0.3.9__py3-none-any.whl → 0.3.10__py3-none-any.whl
This diff represents the content of publicly available package versions that have been released to one of the supported registries. The information contained in this diff is provided for informational purposes only and reflects changes between package versions as they appear in their respective public registries.
- dora_vad/main.py +12 -7
- {dora_vad-0.3.9.dist-info → dora_vad-0.3.10.dist-info}/METADATA +2 -2
- dora_vad-0.3.10.dist-info/RECORD +7 -0
- {dora_vad-0.3.9.dist-info → dora_vad-0.3.10.dist-info}/WHEEL +1 -1
- dora_vad-0.3.9.dist-info/RECORD +0 -7
- {dora_vad-0.3.9.dist-info → dora_vad-0.3.10.dist-info}/entry_points.txt +0 -0
- {dora_vad-0.3.9.dist-info → dora_vad-0.3.10.dist-info}/top_level.txt +0 -0
dora_vad/main.py
CHANGED
@@ -7,11 +7,11 @@ from dora import Node
|
|
7
7
|
from silero_vad import get_speech_timestamps, load_silero_vad
|
8
8
|
|
9
9
|
model = load_silero_vad()
|
10
|
-
MIN_SILENCE_DURATION_MS = int(os.getenv("MIN_SILENCE_DURATION_MS", "
|
10
|
+
MIN_SILENCE_DURATION_MS = int(os.getenv("MIN_SILENCE_DURATION_MS", "200"))
|
11
11
|
MIN_SPEECH_DURATION_MS = int(os.getenv("MIN_SPEECH_DURATION_MS", "300"))
|
12
|
-
|
13
|
-
|
14
|
-
|
12
|
+
THRESHOLD = float(os.getenv("THRESHOLD", "0.4"))
|
13
|
+
MAX_AUDIO_DURATION_S = float(os.getenv("MAX_AUDIO_DURATION_S", "75"))
|
14
|
+
MIN_AUDIO_SAMPLING_DURATION_MS = int(os.getenv("MIN_AUDIO_SAMPLING_DURATION_MS", "500"))
|
15
15
|
|
16
16
|
|
17
17
|
def main():
|
@@ -26,10 +26,11 @@ def main():
|
|
26
26
|
last_audios += [audio]
|
27
27
|
last_audios = last_audios[-100:]
|
28
28
|
audio = np.concatenate(last_audios)
|
29
|
+
sr = event["metadata"].get("sample_rate", 16000)
|
29
30
|
speech_timestamps = get_speech_timestamps(
|
30
31
|
torch.from_numpy(audio),
|
31
32
|
model,
|
32
|
-
threshold=
|
33
|
+
threshold=THRESHOLD,
|
33
34
|
min_speech_duration_ms=MIN_SPEECH_DURATION_MS,
|
34
35
|
min_silence_duration_ms=MIN_SILENCE_DURATION_MS,
|
35
36
|
)
|
@@ -37,16 +38,20 @@ def main():
|
|
37
38
|
# Check ig there is timestamp
|
38
39
|
if (
|
39
40
|
len(speech_timestamps) > 0
|
40
|
-
and len(
|
41
|
+
and len(audio) > MIN_AUDIO_SAMPLING_DURATION_MS * sr / 1000
|
41
42
|
):
|
42
43
|
# Check if the audio is not cut at the end. And only return if there is a long time spent
|
43
44
|
if speech_timestamps[-1]["end"] == len(audio):
|
45
|
+
node.send_output(
|
46
|
+
"timestamp_start",
|
47
|
+
pa.array([speech_timestamps[-1]["start"]]),
|
48
|
+
)
|
44
49
|
continue
|
45
50
|
audio = audio[0 : speech_timestamps[-1]["end"]]
|
46
51
|
node.send_output("audio", pa.array(audio))
|
47
52
|
last_audios = [audio[speech_timestamps[-1]["end"] :]]
|
48
53
|
|
49
54
|
# If there is no sound for too long return the audio
|
50
|
-
elif len(last_audios) >
|
55
|
+
elif len(last_audios) > MAX_AUDIO_DURATION_S:
|
51
56
|
node.send_output("audio", pa.array(audio))
|
52
57
|
last_audios = []
|
@@ -1,12 +1,12 @@
|
|
1
1
|
Metadata-Version: 2.2
|
2
2
|
Name: dora-vad
|
3
|
-
Version: 0.3.
|
3
|
+
Version: 0.3.10
|
4
4
|
Summary: Dora Node for Text translating using Argostranslate
|
5
5
|
Author-email: Haixuan Xavier Tao <tao.xavier@outlook.com>
|
6
6
|
License: MIT
|
7
7
|
Requires-Python: >=3.8
|
8
8
|
Description-Content-Type: text/markdown
|
9
|
-
Requires-Dist: dora-rs>=0.3.
|
9
|
+
Requires-Dist: dora-rs>=0.3.9
|
10
10
|
Requires-Dist: numpy<2.0.0
|
11
11
|
Requires-Dist: silero-vad>=5.1
|
12
12
|
|
@@ -0,0 +1,7 @@
|
|
1
|
+
dora_vad/__init__.py,sha256=HuSK3dnyI9Pb5QAuaKFwQQ3J5SIZnLcKHPJO0norGzc,353
|
2
|
+
dora_vad/main.py,sha256=z1-nS4_tnhTasV75Bsc6s2Der_ISbpjafmNJLYrsvws,2241
|
3
|
+
dora_vad-0.3.10.dist-info/METADATA,sha256=izl0Gjyg6Z1xQ2-N-b7jDn0LaZSR15aghlv-nwF-VE8,1461
|
4
|
+
dora_vad-0.3.10.dist-info/WHEEL,sha256=jB7zZ3N9hIM9adW7qlTAyycLYW9npaWKLRzaoVcLKcM,91
|
5
|
+
dora_vad-0.3.10.dist-info/entry_points.txt,sha256=X82-C0mpNH9bVjwn24mzfoWX5ssrK6W1Ylk1HPDR0ME,48
|
6
|
+
dora_vad-0.3.10.dist-info/top_level.txt,sha256=1EUgMajnck6tHR4mFagv_uv-t2A6Cabd2wYtEldKm2Q,9
|
7
|
+
dora_vad-0.3.10.dist-info/RECORD,,
|
dora_vad-0.3.9.dist-info/RECORD
DELETED
@@ -1,7 +0,0 @@
|
|
1
|
-
dora_vad/__init__.py,sha256=HuSK3dnyI9Pb5QAuaKFwQQ3J5SIZnLcKHPJO0norGzc,353
|
2
|
-
dora_vad/main.py,sha256=HfdT87XdtlZfoNRfKbWKl-ylVNFoq9t8acfrzB6n3wE,1913
|
3
|
-
dora_vad-0.3.9.dist-info/METADATA,sha256=B4iuvHVM6YfqEOqkyw-dZ6_kBJU1fjGU--2Ds6Bfq0g,1460
|
4
|
-
dora_vad-0.3.9.dist-info/WHEEL,sha256=In9FTNxeP60KnTkGw7wk6mJPYd_dQSjEZmXdBdMCI-8,91
|
5
|
-
dora_vad-0.3.9.dist-info/entry_points.txt,sha256=X82-C0mpNH9bVjwn24mzfoWX5ssrK6W1Ylk1HPDR0ME,48
|
6
|
-
dora_vad-0.3.9.dist-info/top_level.txt,sha256=1EUgMajnck6tHR4mFagv_uv-t2A6Cabd2wYtEldKm2Q,9
|
7
|
-
dora_vad-0.3.9.dist-info/RECORD,,
|
File without changes
|
File without changes
|