Spaces:
Sleeping
Sleeping
commit gradio new
Browse files- Berhenti.mp3 +0 -0
- Dimulai.mp3 +0 -0
- app.py +732 -76
- wordlist.lst +0 -0
Berhenti.mp3
ADDED
Binary file (27.2 kB). View file
|
|
Dimulai.mp3
ADDED
Binary file (26.8 kB). View file
|
|
app.py
CHANGED
@@ -1,128 +1,784 @@
|
|
1 |
import gradio as gr
|
|
|
2 |
import os
|
3 |
import requests
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
4 |
from dotenv import load_dotenv
|
5 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
6 |
load_dotenv()
|
7 |
API_TRANSCRIBE = os.getenv("API_TRANSCRIBE")
|
8 |
API_TEXT = os.getenv("API_TEXT")
|
9 |
|
10 |
-
#
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
11 |
def handle_audio(audio_file):
|
|
|
12 |
if audio_file is None:
|
13 |
-
return "", "", ""
|
14 |
-
|
15 |
-
|
16 |
-
|
17 |
-
|
18 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
19 |
|
20 |
def handle_text(dialogue):
|
|
|
21 |
if not dialogue.strip():
|
22 |
-
return "", "", ""
|
23 |
-
|
24 |
-
|
25 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
26 |
|
27 |
-
# ==== Function to toggle inputs with refresh ====
|
28 |
def toggle_inputs_with_refresh(choice):
|
29 |
-
#
|
30 |
return (
|
31 |
-
gr.update(visible=(choice == "Upload Audio"), value=None),
|
32 |
-
gr.update(visible=(choice == "Realtime Recording"), value=None), #
|
33 |
-
gr.update(visible=(choice == "Input Teks"), value=""), #
|
34 |
-
gr.update(
|
35 |
-
gr.update(
|
36 |
-
gr.update(
|
|
|
|
|
|
|
|
|
|
|
37 |
)
|
38 |
|
39 |
-
# ==== Function to clear all data ====
|
40 |
def clear_all_data():
|
41 |
return (
|
42 |
-
gr.update(value=None), #
|
43 |
-
gr.update(value=None), #
|
44 |
-
gr.update(value=""), #
|
45 |
-
gr.update(value=""), #
|
46 |
-
gr.update(value=""), #
|
47 |
-
gr.update(value="")
|
|
|
|
|
|
|
|
|
48 |
)
|
49 |
|
50 |
-
# ==== Combined processing function ====
|
51 |
def process_data(choice, audio_upload, audio_record, text_input):
|
|
|
|
|
|
|
|
|
|
|
52 |
if choice == "Upload Audio":
|
53 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
54 |
elif choice == "Realtime Recording":
|
55 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
56 |
elif choice == "Input Teks":
|
57 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
58 |
else:
|
59 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
60 |
|
61 |
-
|
62 |
-
|
63 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
64 |
|
65 |
with gr.Row():
|
66 |
-
|
67 |
-
|
68 |
-
|
69 |
-
|
70 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
71 |
)
|
72 |
-
|
73 |
-
|
74 |
-
|
75 |
-
|
76 |
-
|
77 |
-
|
78 |
-
|
79 |
-
|
80 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
81 |
)
|
82 |
-
|
83 |
-
|
84 |
-
|
85 |
-
|
86 |
-
|
87 |
-
|
88 |
-
visible=False
|
89 |
)
|
90 |
-
|
91 |
-
|
92 |
-
|
93 |
-
|
94 |
-
|
95 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
96 |
)
|
97 |
|
98 |
-
#
|
99 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
100 |
|
101 |
-
#
|
102 |
-
|
103 |
-
|
104 |
-
|
|
|
|
|
105 |
|
106 |
-
#
|
107 |
-
# Event untuk dropdown change - akan clear semua data
|
108 |
input_choice.change(
|
109 |
-
fn=
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
110 |
inputs=input_choice,
|
111 |
-
outputs=[
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
112 |
)
|
113 |
|
114 |
-
# Event untuk tombol clear - akan clear semua data
|
115 |
clear_button.click(
|
116 |
fn=clear_all_data,
|
117 |
-
outputs=[
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
118 |
)
|
119 |
|
120 |
-
# Single event handler untuk semua jenis input
|
121 |
process_button.click(
|
122 |
fn=process_data,
|
123 |
inputs=[input_choice, audio_upload, audio_record, text_input],
|
124 |
-
outputs=[
|
125 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
126 |
)
|
127 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
128 |
app.launch()
|
|
|
1 |
import gradio as gr
|
2 |
+
import threading
|
3 |
import os
|
4 |
import requests
|
5 |
+
import string
|
6 |
+
import pygame
|
7 |
+
import time
|
8 |
+
from pydub import AudioSegment
|
9 |
+
from nltk.tokenize import word_tokenize
|
10 |
+
import nltk
|
11 |
+
from nltk.corpus import words, stopwords
|
12 |
from dotenv import load_dotenv
|
13 |
|
14 |
+
# Initialize pygame mixer for sound playback
|
15 |
+
pygame.mixer.init()
|
16 |
+
|
17 |
+
# Download resource NLTK (hanya sekali)
|
18 |
+
nltk.download('punkt')
|
19 |
+
nltk.download('words')
|
20 |
+
nltk.download('stopwords')
|
21 |
+
|
22 |
load_dotenv()
|
23 |
API_TRANSCRIBE = os.getenv("API_TRANSCRIBE")
|
24 |
API_TEXT = os.getenv("API_TEXT")
|
25 |
|
26 |
+
# Path ke file sound yang akan diputar (sesuaikan dengan file Anda)
|
27 |
+
NOTIFICATION_SOUND_PATH = "Berhenti.mp3" # Sound ketika recording selesai
|
28 |
+
START_RECORDING_SOUND_PATH = "Dimulai.mp3" # Sound ketika mulai recording
|
29 |
+
|
30 |
+
english_words = set(words.words())
|
31 |
+
indonesian_stopwords = set(stopwords.words('indonesian'))
|
32 |
+
|
33 |
+
def load_indonesian_wordlist(filepath='wordlist.lst'):
|
34 |
+
try:
|
35 |
+
with open(filepath, encoding='utf-8') as f:
|
36 |
+
return set(line.strip().lower() for line in f if line.strip())
|
37 |
+
except UnicodeDecodeError:
|
38 |
+
try:
|
39 |
+
with open(filepath, encoding='latin-1') as f:
|
40 |
+
return set(line.strip().lower() for line in f if line.strip())
|
41 |
+
except Exception:
|
42 |
+
return set()
|
43 |
+
except Exception:
|
44 |
+
return set()
|
45 |
+
|
46 |
+
indonesian_words = load_indonesian_wordlist()
|
47 |
+
valid_words = english_words.union(indonesian_words)
|
48 |
+
|
49 |
+
def contains_medical_terms_auto_threshold(text, medical_words):
|
50 |
+
tokens = word_tokenize(text.lower())
|
51 |
+
tokens = [w.strip(string.punctuation) for w in tokens if w.isalpha()]
|
52 |
+
if not tokens:
|
53 |
+
return False
|
54 |
+
medical_count = sum(1 for w in tokens if w in medical_words)
|
55 |
+
ratio = medical_count / len(tokens)
|
56 |
+
threshold = 0.4 if len(tokens) <= 5 else 0.1
|
57 |
+
return ratio >= threshold
|
58 |
+
|
59 |
+
medical_words = load_indonesian_wordlist('wordlist.lst')
|
60 |
+
|
61 |
+
MAX_DURATION_SECONDS = 600
|
62 |
+
|
63 |
+
def validate_audio_duration(audio_file):
|
64 |
+
try:
|
65 |
+
audio = AudioSegment.from_file(audio_file)
|
66 |
+
duration_sec = len(audio) / 1000.0
|
67 |
+
if duration_sec > MAX_DURATION_SECONDS:
|
68 |
+
return False, duration_sec
|
69 |
+
return True, duration_sec
|
70 |
+
except Exception as e:
|
71 |
+
return False, -1
|
72 |
+
|
73 |
+
def play_notification_sound():
|
74 |
+
"""Function untuk memainkan sound notification ketika recording selesai"""
|
75 |
+
try:
|
76 |
+
if os.path.exists(NOTIFICATION_SOUND_PATH):
|
77 |
+
pygame.mixer.music.load(NOTIFICATION_SOUND_PATH)
|
78 |
+
pygame.mixer.music.play()
|
79 |
+
print("π Playing completion notification sound...")
|
80 |
+
else:
|
81 |
+
print(f"β οΈ File completion sound tidak ditemukan: {NOTIFICATION_SOUND_PATH}")
|
82 |
+
except Exception as e:
|
83 |
+
print(f"β Error playing completion sound: {e}")
|
84 |
+
|
85 |
+
def play_start_recording_sound():
|
86 |
+
"""Function untuk memainkan sound ketika mulai recording"""
|
87 |
+
try:
|
88 |
+
if os.path.exists(START_RECORDING_SOUND_PATH):
|
89 |
+
# Menggunakan Sound effect untuk play bersamaan tanpa interrupt music
|
90 |
+
start_sound = pygame.mixer.Sound(START_RECORDING_SOUND_PATH)
|
91 |
+
start_sound.play()
|
92 |
+
print("π΅ Playing start recording sound...")
|
93 |
+
else:
|
94 |
+
print(f"β οΈ File start recording sound tidak ditemukan: {START_RECORDING_SOUND_PATH}")
|
95 |
+
except Exception as e:
|
96 |
+
print(f"β Error playing start recording sound: {e}")
|
97 |
+
|
98 |
+
def start_recording():
|
99 |
+
"""Function yang dipanggil ketika tombol record ditekan"""
|
100 |
+
print("ποΈ Recording started...")
|
101 |
+
# Play start recording sound
|
102 |
+
threading.Thread(target=play_start_recording_sound, daemon=True).start()
|
103 |
+
return "ποΈ Sedang merekam... Klik stop untuk menyelesaikan"
|
104 |
+
|
105 |
+
def stop_recording(audio):
|
106 |
+
"""Function yang dipanggil ketika recording selesai"""
|
107 |
+
if audio is not None:
|
108 |
+
print("β
Recording completed!")
|
109 |
+
# Play notification sound when recording is completed
|
110 |
+
threading.Thread(target=play_notification_sound, daemon=True).start()
|
111 |
+
return "β
Recording selesai! Audio siap diproses"
|
112 |
+
else:
|
113 |
+
print("β No audio recorded")
|
114 |
+
return "β Tidak ada audio yang direkam"
|
115 |
+
|
116 |
+
def test_microphone():
|
117 |
+
"""Function untuk test microphone"""
|
118 |
+
print("π§ Testing microphone...")
|
119 |
+
return "π§ Testing microphone... Silakan coba record lagi"
|
120 |
+
|
121 |
+
def reset_recording_status():
|
122 |
+
"""Function untuk reset status recording"""
|
123 |
+
return "π± Siap untuk merekam - Klik tombol record"
|
124 |
+
|
125 |
def handle_audio(audio_file):
|
126 |
+
"""Handle audio processing - returns (validation_message, transcript, soap, tags)"""
|
127 |
if audio_file is None:
|
128 |
+
return "β Tidak ada file audio", "", "", ""
|
129 |
+
|
130 |
+
valid, duration = validate_audio_duration(audio_file)
|
131 |
+
if not valid:
|
132 |
+
if duration == -1:
|
133 |
+
msg = "β οΈ Gagal memproses file audio."
|
134 |
+
else:
|
135 |
+
msg = f"β οΈ Durasi rekaman terlalu panjang ({duration:.1f}s). Maksimal {MAX_DURATION_SECONDS}s."
|
136 |
+
return msg, "", "", ""
|
137 |
+
|
138 |
+
try:
|
139 |
+
with open(audio_file, "rb") as f:
|
140 |
+
files = {"audio": f}
|
141 |
+
response = requests.post(API_TRANSCRIBE, files=files)
|
142 |
+
result = response.json()
|
143 |
+
|
144 |
+
transcription = result.get("transcription", "")
|
145 |
+
soap_content = result.get("soap_content", "")
|
146 |
+
tags_content = result.get("tags_content", "")
|
147 |
+
|
148 |
+
if not transcription and not soap_content and not tags_content:
|
149 |
+
return "β οΈ Tidak ada hasil dari proses audio", "", "", ""
|
150 |
+
|
151 |
+
return "", transcription, soap_content, tags_content
|
152 |
+
|
153 |
+
except Exception as e:
|
154 |
+
return f"β Error processing audio: {str(e)}", "", "", ""
|
155 |
|
156 |
def handle_text(dialogue):
|
157 |
+
"""Handle text processing - returns (validation_message, transcript, soap, tags)"""
|
158 |
if not dialogue.strip():
|
159 |
+
return "β οΈ Teks tidak boleh kosong", "", "", ""
|
160 |
+
|
161 |
+
if not contains_medical_terms_auto_threshold(dialogue, medical_words):
|
162 |
+
return "β οΈ Teks tidak mengandung istilah medis yang cukup untuk diproses.", "", "", ""
|
163 |
+
|
164 |
+
try:
|
165 |
+
response = requests.post(API_TEXT, json={"dialogue": dialogue})
|
166 |
+
result = response.json()
|
167 |
+
|
168 |
+
soap_content = result.get("soap_content", "")
|
169 |
+
tags_content = result.get("tags_content", "")
|
170 |
+
|
171 |
+
if not soap_content and not tags_content:
|
172 |
+
return "β οΈ Tidak ada hasil dari proses teks", "", "", ""
|
173 |
+
|
174 |
+
return "", dialogue, soap_content, tags_content
|
175 |
+
|
176 |
+
except Exception as e:
|
177 |
+
return f"β Error processing text: {str(e)}", "", "", ""
|
178 |
|
|
|
179 |
def toggle_inputs_with_refresh(choice):
|
180 |
+
# Tampilkan input dan validasi yang sesuai, sembunyikan lainnya
|
181 |
return (
|
182 |
+
gr.update(visible=(choice == "Upload Audio"), value=None), # audio upload
|
183 |
+
gr.update(visible=(choice == "Realtime Recording"), value=None), # audio record
|
184 |
+
gr.update(visible=(choice == "Input Teks"), value=""), # text input
|
185 |
+
gr.update(visible=(choice == "Upload Audio")), # validasi upload
|
186 |
+
gr.update(visible=(choice == "Realtime Recording")), # validasi realtime
|
187 |
+
gr.update(visible=(choice == "Input Teks")), # validasi teks
|
188 |
+
gr.update(visible=(choice == "Realtime Recording")), # recording status group
|
189 |
+
gr.update(visible=(choice == "Realtime Recording")), # record audio group
|
190 |
+
gr.update(value=""), # transcript
|
191 |
+
gr.update(value=""), # soap
|
192 |
+
gr.update(value=""), # tags
|
193 |
)
|
194 |
|
|
|
195 |
def clear_all_data():
|
196 |
return (
|
197 |
+
gr.update(value=None), # audio_upload
|
198 |
+
gr.update(value=None), # audio_record
|
199 |
+
gr.update(value=""), # text_input
|
200 |
+
gr.update(value=""), # validation_upload
|
201 |
+
gr.update(value=""), # validation_realtime
|
202 |
+
gr.update(value=""), # validation_text
|
203 |
+
gr.update(value="π± Siap untuk merekam"), # recording_status
|
204 |
+
gr.update(value=""), # transcript_output
|
205 |
+
gr.update(value=""), # soap_output
|
206 |
+
gr.update(value=""), # tags_output
|
207 |
)
|
208 |
|
|
|
209 |
def process_data(choice, audio_upload, audio_record, text_input):
|
210 |
+
"""
|
211 |
+
Process data based on choice and return results in correct order:
|
212 |
+
Returns: (validation_upload, validation_realtime, validation_text, transcript, soap, tags)
|
213 |
+
"""
|
214 |
+
|
215 |
if choice == "Upload Audio":
|
216 |
+
# Process upload audio
|
217 |
+
validation_msg, transcript, soap, tags = handle_audio(audio_upload)
|
218 |
+
return (
|
219 |
+
validation_msg, # validation_upload
|
220 |
+
"", # validation_realtime (empty)
|
221 |
+
"", # validation_text (empty)
|
222 |
+
transcript, # transcript_output
|
223 |
+
soap, # soap_output
|
224 |
+
tags # tags_output
|
225 |
+
)
|
226 |
+
|
227 |
elif choice == "Realtime Recording":
|
228 |
+
# Process realtime recording
|
229 |
+
validation_msg, transcript, soap, tags = handle_audio(audio_record)
|
230 |
+
return (
|
231 |
+
"", # validation_upload (empty)
|
232 |
+
validation_msg, # validation_realtime
|
233 |
+
"", # validation_text (empty)
|
234 |
+
transcript, # transcript_output
|
235 |
+
soap, # soap_output
|
236 |
+
tags # tags_output
|
237 |
+
)
|
238 |
+
|
239 |
elif choice == "Input Teks":
|
240 |
+
# Process text input
|
241 |
+
validation_msg, transcript, soap, tags = handle_text(text_input)
|
242 |
+
return (
|
243 |
+
"", # validation_upload (empty)
|
244 |
+
"", # validation_realtime (empty)
|
245 |
+
validation_msg, # validation_text
|
246 |
+
transcript, # transcript_output (will be same as input for text)
|
247 |
+
soap, # soap_output
|
248 |
+
tags # tags_output
|
249 |
+
)
|
250 |
+
|
251 |
else:
|
252 |
+
# Default case - clear all
|
253 |
+
return ("", "", "", "", "", "")
|
254 |
+
|
255 |
+
# Custom CSS untuk tampilan modern dengan alignment yang diperbaiki
|
256 |
+
modern_css = """
|
257 |
+
<style>
|
258 |
+
/* Background gradient yang modern */
|
259 |
+
.gradio-container {
|
260 |
+
background: linear-gradient(135deg, #667eea 0%, #764ba2 100%);
|
261 |
+
min-height: 100vh;
|
262 |
+
font-family: 'Segoe UI', Tahoma, Geneva, Verdana, sans-serif;
|
263 |
+
}
|
264 |
+
|
265 |
+
/* Header styling */
|
266 |
+
.main-header {
|
267 |
+
text-align: center;
|
268 |
+
padding: 2rem 0;
|
269 |
+
color: white;
|
270 |
+
margin-bottom: 2rem;
|
271 |
+
}
|
272 |
+
|
273 |
+
.main-header h1 {
|
274 |
+
font-size: 3rem;
|
275 |
+
font-weight: 700;
|
276 |
+
margin-bottom: 0.5rem;
|
277 |
+
text-shadow: 0 2px 4px rgba(0,0,0,0.3);
|
278 |
+
}
|
279 |
+
|
280 |
+
.main-header p {
|
281 |
+
font-size: 1.2rem;
|
282 |
+
opacity: 0.9;
|
283 |
+
font-weight: 300;
|
284 |
+
}
|
285 |
+
|
286 |
+
/* Card styling */
|
287 |
+
.input-card {
|
288 |
+
background: rgba(255, 255, 255, 0.95);
|
289 |
+
border-radius: 20px;
|
290 |
+
padding: 2rem;
|
291 |
+
margin: 0.8rem 0;
|
292 |
+
box-shadow: 0 8px 32px rgba(0,0,0,0.1);
|
293 |
+
backdrop-filter: blur(10px);
|
294 |
+
border: 1px solid rgba(255,255,255,0.2);
|
295 |
+
}
|
296 |
+
|
297 |
+
.output-card {
|
298 |
+
background: rgba(255, 255, 255, 0.95);
|
299 |
+
border-radius: 20px;
|
300 |
+
padding: 2rem;
|
301 |
+
box-shadow: 0 8px 32px rgba(0,0,0,0.1);
|
302 |
+
backdrop-filter: blur(10px);
|
303 |
+
border: 1px solid rgba(255,255,255,0.2);
|
304 |
+
}
|
305 |
+
|
306 |
+
/* Record audio section with padding */
|
307 |
+
.record-audio-section {
|
308 |
+
background: rgba(255, 255, 255, 0.95);
|
309 |
+
border-radius: 20px;
|
310 |
+
padding: 2rem;
|
311 |
+
margin: 0.8rem 0;
|
312 |
+
box-shadow: 0 8px 32px rgba(0,0,0,0.1);
|
313 |
+
backdrop-filter: blur(10px);
|
314 |
+
border: 1px solid rgba(255,255,255,0.2);
|
315 |
+
}
|
316 |
+
|
317 |
+
/* Output header styling - aligned properly */
|
318 |
+
.output-header {
|
319 |
+
text-align: left;
|
320 |
+
margin-bottom: 1.5rem;
|
321 |
+
padding-left: 0;
|
322 |
+
}
|
323 |
+
|
324 |
+
.output-header h3 {
|
325 |
+
font-size: 1.5rem;
|
326 |
+
font-weight: 600;
|
327 |
+
color: #2d3436;
|
328 |
+
margin: 0;
|
329 |
+
display: flex;
|
330 |
+
align-items: center;
|
331 |
+
}
|
332 |
+
|
333 |
+
/* Spacing adjustments */
|
334 |
+
.input-section {
|
335 |
+
margin-bottom: 1rem;
|
336 |
+
}
|
337 |
+
|
338 |
+
.button-section {
|
339 |
+
margin-top: 0.5rem;
|
340 |
+
margin-bottom: 1rem;
|
341 |
+
}
|
342 |
+
|
343 |
+
/* Updated output section styling with proper alignment */
|
344 |
+
.output-section {
|
345 |
+
background: rgba(255, 255, 255, 0.98);
|
346 |
+
border-radius: 15px;
|
347 |
+
padding: 1.5rem;
|
348 |
+
margin: 1rem 0 0 0;
|
349 |
+
box-shadow: 0 4px 20px rgba(0,0,0,0.08);
|
350 |
+
}
|
351 |
+
|
352 |
+
/* Ensure all output components are aligned consistently */
|
353 |
+
.output-container {
|
354 |
+
display: flex;
|
355 |
+
flex-direction: column;
|
356 |
+
gap: 1rem;
|
357 |
+
}
|
358 |
|
359 |
+
/* Button styling */
|
360 |
+
.record-btn {
|
361 |
+
background: linear-gradient(45deg, #ff6b6b, #ee5a24);
|
362 |
+
border: none;
|
363 |
+
border-radius: 50px;
|
364 |
+
padding: 1rem 2rem;
|
365 |
+
color: white;
|
366 |
+
font-weight: 600;
|
367 |
+
font-size: 1.1rem;
|
368 |
+
cursor: pointer;
|
369 |
+
transition: all 0.3s ease;
|
370 |
+
box-shadow: 0 4px 15px rgba(238, 90, 36, 0.4);
|
371 |
+
}
|
372 |
+
|
373 |
+
.record-btn:hover {
|
374 |
+
transform: translateY(-2px);
|
375 |
+
box-shadow: 0 6px 20px rgba(238, 90, 36, 0.6);
|
376 |
+
}
|
377 |
+
|
378 |
+
.test-btn {
|
379 |
+
background: linear-gradient(45deg, #74b9ff, #0984e3);
|
380 |
+
border: none;
|
381 |
+
border-radius: 50px;
|
382 |
+
padding: 0.8rem 1.5rem;
|
383 |
+
color: white;
|
384 |
+
font-weight: 500;
|
385 |
+
cursor: pointer;
|
386 |
+
transition: all 0.3s ease;
|
387 |
+
}
|
388 |
+
|
389 |
+
.test-btn:hover {
|
390 |
+
transform: translateY(-1px);
|
391 |
+
box-shadow: 0 4px 15px rgba(116, 185, 255, 0.4);
|
392 |
+
}
|
393 |
+
|
394 |
+
.reset-btn {
|
395 |
+
background: linear-gradient(45deg, #a29bfe, #6c5ce7);
|
396 |
+
border: none;
|
397 |
+
border-radius: 15px;
|
398 |
+
padding: 0.4rem 1rem;
|
399 |
+
font-size: 0.8rem;
|
400 |
+
color: white;
|
401 |
+
font-weight: 500;
|
402 |
+
cursor: pointer;
|
403 |
+
transition: all 0.3s ease;
|
404 |
+
height: 2rem;
|
405 |
+
min-width: 80px;
|
406 |
+
}
|
407 |
+
|
408 |
+
.reset-btn:hover {
|
409 |
+
transform: translateY(-1px);
|
410 |
+
box-shadow: 0 2px 10px rgba(162, 155, 254, 0.4);
|
411 |
+
}
|
412 |
+
|
413 |
+
.process-btn {
|
414 |
+
background: linear-gradient(45deg, #00b894, #00cec9);
|
415 |
+
border: none;
|
416 |
+
border-radius: 50px;
|
417 |
+
padding: 1rem 2.5rem;
|
418 |
+
color: white;
|
419 |
+
font-weight: 600;
|
420 |
+
font-size: 1.2rem;
|
421 |
+
cursor: pointer;
|
422 |
+
transition: all 0.3s ease;
|
423 |
+
box-shadow: 0 4px 15px rgba(0, 184, 148, 0.4);
|
424 |
+
margin: 1rem 0;
|
425 |
+
}
|
426 |
+
|
427 |
+
.process-btn:hover {
|
428 |
+
transform: translateY(-2px);
|
429 |
+
box-shadow: 0 6px 20px rgba(0, 184, 148, 0.6);
|
430 |
+
}
|
431 |
+
|
432 |
+
/* Status indicator */
|
433 |
+
.status-indicator {
|
434 |
+
background: rgba(255, 255, 255, 0.95);
|
435 |
+
border: 2px solid #74b9ff;
|
436 |
+
color: #2d3436;
|
437 |
+
padding: 2rem;
|
438 |
+
border-radius: 15px;
|
439 |
+
font-weight: 500;
|
440 |
+
margin: 1rem 0;
|
441 |
+
box-shadow: 0 4px 15px rgba(116, 185, 255, 0.2);
|
442 |
+
}
|
443 |
+
|
444 |
+
/* Input styling */
|
445 |
+
.audio-input {
|
446 |
+
border: 2px dashed #ddd;
|
447 |
+
border-radius: 15px;
|
448 |
+
padding: 2rem;
|
449 |
+
text-align: center;
|
450 |
+
background: rgba(255,255,255,0.5);
|
451 |
+
transition: all 0.3s ease;
|
452 |
+
}
|
453 |
+
|
454 |
+
.audio-input:hover {
|
455 |
+
border-color: #74b9ff;
|
456 |
+
background: rgba(116, 185, 255, 0.1);
|
457 |
+
}
|
458 |
+
|
459 |
+
/* Ensure proper text alignment in outputs */
|
460 |
+
.output-section textarea {
|
461 |
+
text-align: left;
|
462 |
+
vertical-align: top;
|
463 |
+
resize: vertical;
|
464 |
+
}
|
465 |
+
|
466 |
+
/* Responsive design */
|
467 |
+
@media (max-width: 768px) {
|
468 |
+
.main-header h1 {
|
469 |
+
font-size: 2rem;
|
470 |
+
}
|
471 |
+
|
472 |
+
.input-card, .output-card, .record-audio-section {
|
473 |
+
padding: 2rem;
|
474 |
+
}
|
475 |
+
|
476 |
+
.record-btn, .process-btn {
|
477 |
+
width: 100%;
|
478 |
+
margin: 0.5rem 0;
|
479 |
+
}
|
480 |
+
|
481 |
+
.reset-btn {
|
482 |
+
min-width: 70px;
|
483 |
+
font-size: 0.75rem;
|
484 |
+
}
|
485 |
+
}
|
486 |
+
|
487 |
+
/* Animation untuk elemen yang muncul */
|
488 |
+
@keyframes fadeInUp {
|
489 |
+
from {
|
490 |
+
opacity: 0;
|
491 |
+
transform: translateY(30px);
|
492 |
+
}
|
493 |
+
to {
|
494 |
+
opacity: 1;
|
495 |
+
transform: translateY(0);
|
496 |
+
}
|
497 |
+
}
|
498 |
+
|
499 |
+
.input-card, .output-card, .record-audio-section {
|
500 |
+
animation: fadeInUp 0.6s ease-out;
|
501 |
+
}
|
502 |
+
|
503 |
+
/* Microphone icon animation */
|
504 |
+
@keyframes pulse {
|
505 |
+
0% {
|
506 |
+
transform: scale(1);
|
507 |
+
}
|
508 |
+
50% {
|
509 |
+
transform: scale(1.05);
|
510 |
+
}
|
511 |
+
100% {
|
512 |
+
transform: scale(1);
|
513 |
+
}
|
514 |
+
}
|
515 |
+
|
516 |
+
.recording-active {
|
517 |
+
animation: pulse 1s infinite;
|
518 |
+
}
|
519 |
+
</style>
|
520 |
+
"""
|
521 |
+
|
522 |
+
# Buat interface dengan theme modern
|
523 |
+
with gr.Blocks(
|
524 |
+
title="π©Ί SOAP AI - Modern Interface",
|
525 |
+
css=modern_css,
|
526 |
+
theme=gr.themes.Soft(
|
527 |
+
primary_hue="blue",
|
528 |
+
secondary_hue="cyan",
|
529 |
+
neutral_hue="slate",
|
530 |
+
font=gr.themes.GoogleFont("Inter")
|
531 |
+
)
|
532 |
+
) as app:
|
533 |
+
|
534 |
+
# Header
|
535 |
+
gr.HTML("""
|
536 |
+
<div class="main-header">
|
537 |
+
<h1>ποΈ Realtime Recording</h1>
|
538 |
+
<p>High Quality Audio Recording with Smart Notifications</p>
|
539 |
+
</div>
|
540 |
+
""")
|
541 |
|
542 |
with gr.Row():
|
543 |
+
with gr.Column(scale=8):
|
544 |
+
input_choice = gr.Dropdown(
|
545 |
+
choices=["Upload Audio", "Realtime Recording", "Input Teks"],
|
546 |
+
value="Realtime Recording",
|
547 |
+
label="π― Pilih Metode Input",
|
548 |
+
container=True,
|
549 |
+
elem_classes=["input-dropdown"]
|
550 |
+
)
|
551 |
+
with gr.Column(scale=2):
|
552 |
+
clear_button = gr.Button(
|
553 |
+
"ποΈ Clear",
|
554 |
+
variant="secondary",
|
555 |
+
size="sm",
|
556 |
+
elem_classes=["reset-btn"]
|
557 |
+
)
|
558 |
+
|
559 |
+
# Input Section - Upload Audio
|
560 |
+
with gr.Group(elem_classes=["input-card"], visible=False) as upload_audio_group:
|
561 |
+
gr.HTML("<h3>π Upload Audio File</h3>")
|
562 |
+
|
563 |
+
audio_upload = gr.Audio(
|
564 |
+
sources=["upload"],
|
565 |
+
label="π Upload File Audio",
|
566 |
+
type="filepath",
|
567 |
+
show_download_button=False,
|
568 |
+
show_share_button=False,
|
569 |
+
interactive=True,
|
570 |
+
elem_classes=["audio-input"]
|
571 |
)
|
572 |
+
|
573 |
+
# Input Section - Record Audio with proper padding
|
574 |
+
with gr.Group(elem_classes=["record-audio-section"], visible=True) as record_audio_group:
|
575 |
+
gr.HTML("<h3>π΅ Record Your Audio</h3>")
|
576 |
+
|
577 |
+
audio_record = gr.Audio(
|
578 |
+
sources=["microphone"],
|
579 |
+
label="ποΈ Realtime Recording",
|
580 |
+
type="filepath",
|
581 |
+
show_download_button=True,
|
582 |
+
show_share_button=False,
|
583 |
+
interactive=True,
|
584 |
+
streaming=False,
|
585 |
+
elem_classes=["audio-input"]
|
586 |
+
)
|
587 |
+
|
588 |
+
# Input Section - Text Input (without record audio section)
|
589 |
+
with gr.Group(elem_classes=["input-card"], visible=False) as text_input_group:
|
590 |
+
gr.HTML("<h3>π Input Teks</h3>")
|
591 |
+
|
592 |
+
text_input = gr.Textbox(
|
593 |
+
label="π Masukkan Percakapan Dokter-Pasien",
|
594 |
+
lines=6,
|
595 |
+
placeholder="Ketik percakapan antara dokter dan pasien di sini...",
|
596 |
+
elem_classes=["text-input"]
|
597 |
+
)
|
598 |
+
|
599 |
+
# Status Section - hanya untuk Realtime Recording
|
600 |
+
recording_status_group = gr.Group(elem_classes=["status-indicator"], visible=True)
|
601 |
+
with recording_status_group:
|
602 |
+
gr.HTML("<h4>π Status Recording</h4>")
|
603 |
+
recording_status = gr.Textbox(
|
604 |
+
value="Siap untuk merekam",
|
605 |
+
interactive=False,
|
606 |
+
show_label=False,
|
607 |
+
lines=1,
|
608 |
+
elem_classes=["status-display"]
|
609 |
+
)
|
610 |
+
|
611 |
+
# Validation Section
|
612 |
+
validation_upload = gr.Textbox(
|
613 |
+
label="β οΈ Validasi Upload Audio",
|
614 |
+
lines=1,
|
615 |
+
interactive=False,
|
616 |
+
visible=False,
|
617 |
+
elem_classes=["validation-msg"]
|
618 |
)
|
619 |
+
validation_realtime = gr.Textbox(
|
620 |
+
label="β οΈ Validasi Realtime Recording",
|
621 |
+
lines=1,
|
622 |
+
interactive=False,
|
623 |
+
visible=True,
|
624 |
+
elem_classes=["validation-msg"]
|
|
|
625 |
)
|
626 |
+
validation_text = gr.Textbox(
|
627 |
+
label="β οΈ Validasi Input Teks",
|
628 |
+
lines=1,
|
629 |
+
interactive=False,
|
630 |
+
visible=False,
|
631 |
+
elem_classes=["validation-msg"]
|
632 |
+
)
|
633 |
+
|
634 |
+
# Process Button
|
635 |
+
process_button = gr.Button(
|
636 |
+
"π Proses ke SOAP",
|
637 |
+
variant="primary",
|
638 |
+
size="lg",
|
639 |
+
elem_classes=["process-btn"]
|
640 |
)
|
641 |
|
642 |
+
# Output Section with proper alignment - Fixed Layout
|
643 |
+
with gr.Group(elem_classes=["output-card"]):
|
644 |
+
# Header aligned properly
|
645 |
+
with gr.Row():
|
646 |
+
with gr.Column():
|
647 |
+
gr.HTML('<div class="output-header"><h3>π Hasil Analisis</h3></div>')
|
648 |
+
|
649 |
+
# All outputs in aligned container
|
650 |
+
with gr.Column(elem_classes=["output-container"]):
|
651 |
+
transcript_output = gr.Textbox(
|
652 |
+
label="π Hasil Transkripsi",
|
653 |
+
lines=4,
|
654 |
+
elem_classes=["output-section"]
|
655 |
+
)
|
656 |
+
|
657 |
+
soap_output = gr.Textbox(
|
658 |
+
label="π Ringkasan SOAP",
|
659 |
+
lines=8,
|
660 |
+
elem_classes=["output-section"]
|
661 |
+
)
|
662 |
+
|
663 |
+
tags_output = gr.Textbox(
|
664 |
+
label="π·οΈ Medical Tags",
|
665 |
+
lines=6,
|
666 |
+
elem_classes=["output-section"]
|
667 |
+
)
|
668 |
|
669 |
+
# Footer
|
670 |
+
gr.HTML("""
|
671 |
+
<div style="text-align: center; padding: 2rem; color: rgba(255,255,255,0.7);">
|
672 |
+
<p>Use via API π₯ β’ Built with Gradio π</p>
|
673 |
+
</div>
|
674 |
+
""")
|
675 |
|
676 |
+
# Event handlers untuk toggle inputs
|
|
|
677 |
input_choice.change(
|
678 |
+
fn=lambda choice: (
|
679 |
+
gr.update(visible=(choice == "Upload Audio")), # upload_audio_group
|
680 |
+
gr.update(visible=(choice == "Realtime Recording")), # record_audio_group
|
681 |
+
gr.update(visible=(choice == "Input Teks")), # text_input_group
|
682 |
+
gr.update(visible=(choice == "Upload Audio")), # validation_upload
|
683 |
+
gr.update(visible=(choice == "Realtime Recording")), # validation_realtime
|
684 |
+
gr.update(visible=(choice == "Input Teks")), # validation_text
|
685 |
+
gr.update(visible=(choice == "Realtime Recording")), # recording_status_group
|
686 |
+
gr.update(value=""), # transcript
|
687 |
+
gr.update(value=""), # soap
|
688 |
+
gr.update(value=""), # tags
|
689 |
+
),
|
690 |
inputs=input_choice,
|
691 |
+
outputs=[
|
692 |
+
upload_audio_group,
|
693 |
+
record_audio_group,
|
694 |
+
text_input_group,
|
695 |
+
validation_upload,
|
696 |
+
validation_realtime,
|
697 |
+
validation_text,
|
698 |
+
recording_status_group,
|
699 |
+
transcript_output,
|
700 |
+
soap_output,
|
701 |
+
tags_output,
|
702 |
+
],
|
703 |
+
)
|
704 |
+
|
705 |
+
# Event handlers untuk recording
|
706 |
+
audio_record.start_recording(
|
707 |
+
fn=start_recording,
|
708 |
+
outputs=recording_status
|
709 |
+
)
|
710 |
+
|
711 |
+
audio_record.stop_recording(
|
712 |
+
fn=stop_recording,
|
713 |
+
inputs=audio_record,
|
714 |
+
outputs=recording_status
|
715 |
)
|
716 |
|
|
|
717 |
clear_button.click(
|
718 |
fn=clear_all_data,
|
719 |
+
outputs=[
|
720 |
+
audio_upload,
|
721 |
+
audio_record,
|
722 |
+
text_input,
|
723 |
+
validation_upload,
|
724 |
+
validation_realtime,
|
725 |
+
validation_text,
|
726 |
+
recording_status,
|
727 |
+
transcript_output,
|
728 |
+
soap_output,
|
729 |
+
tags_output,
|
730 |
+
],
|
731 |
)
|
732 |
|
|
|
733 |
process_button.click(
|
734 |
fn=process_data,
|
735 |
inputs=[input_choice, audio_upload, audio_record, text_input],
|
736 |
+
outputs=[
|
737 |
+
validation_upload,
|
738 |
+
validation_realtime,
|
739 |
+
validation_text,
|
740 |
+
transcript_output,
|
741 |
+
soap_output,
|
742 |
+
tags_output,
|
743 |
+
],
|
744 |
+
show_progress="minimal",
|
745 |
)
|
746 |
|
747 |
+
# Startup information
|
748 |
+
if __name__ == "__main__":
|
749 |
+
print("π Starting Enhanced SOAP AI Application with Modern UI...")
|
750 |
+
print("π Setup Instructions:")
|
751 |
+
print("1. Install dependencies: pip install gradio pygame pydub nltk requests python-dotenv")
|
752 |
+
print(f"2. Place your sound files:")
|
753 |
+
print(f" - Start recording sound: {START_RECORDING_SOUND_PATH}")
|
754 |
+
print(f" - Completion sound: {NOTIFICATION_SOUND_PATH}")
|
755 |
+
print("3. Supported sound formats: WAV, MP3, OGG")
|
756 |
+
print("4. Make sure wordlist.lst file is available")
|
757 |
+
print("5. Set up your .env file with API_TRANSCRIBE and API_TEXT")
|
758 |
+
print()
|
759 |
+
|
760 |
+
# Cek apakah file sound ada
|
761 |
+
sounds_status = []
|
762 |
+
if os.path.exists(START_RECORDING_SOUND_PATH):
|
763 |
+
print(f"β
Start recording sound found: {START_RECORDING_SOUND_PATH}")
|
764 |
+
sounds_status.append("start")
|
765 |
+
else:
|
766 |
+
print(f"β οΈ Start recording sound not found: {START_RECORDING_SOUND_PATH}")
|
767 |
+
|
768 |
+
if os.path.exists(NOTIFICATION_SOUND_PATH):
|
769 |
+
print(f"β
Completion sound found: {NOTIFICATION_SOUND_PATH}")
|
770 |
+
sounds_status.append("completion")
|
771 |
+
else:
|
772 |
+
print(f"β οΈ Completion sound not found: {NOTIFICATION_SOUND_PATH}")
|
773 |
+
|
774 |
+
if not sounds_status:
|
775 |
+
print("π Note: Add sound files to enable audio notifications for realtime recording")
|
776 |
+
elif len(sounds_status) == 1:
|
777 |
+
print("π Note: Add the missing sound file for complete audio experience")
|
778 |
+
|
779 |
+
print("\nπ Application will start at: http://localhost:7860")
|
780 |
+
print("ποΈ Make sure to allow microphone access when using Realtime Recording!")
|
781 |
+
print("β¨ New Modern UI with enhanced visual experience!")
|
782 |
+
print()
|
783 |
+
|
784 |
app.launch()
|
wordlist.lst
ADDED
The diff for this file is too large to render.
See raw diff
|
|