Spaces:
Runtime error
Runtime error
Update app.py
Browse filesTambah china tradisional dan vietnam
app.py
CHANGED
@@ -3,6 +3,7 @@ from transformers import pipeline
|
|
3 |
import langdetect
|
4 |
import logging
|
5 |
import os
|
|
|
6 |
|
7 |
# Atur direktori cache
|
8 |
os.environ["HF_HOME"] = "/app/cache"
|
@@ -20,7 +21,9 @@ MODEL_MAP = {
|
|
20 |
"fr": "Helsinki-NLP/opus-mt-fr-en", # Prancis ke Inggris
|
21 |
"es": "Helsinki-NLP/opus-mt-es-en", # Spanyol ke Inggris
|
22 |
"ja": "Helsinki-NLP/opus-mt-ja-en", # Jepang ke Inggris
|
23 |
-
"zh": "Helsinki-NLP/opus-mt-zh-en", # Mandarin ke Inggris
|
|
|
|
|
24 |
}
|
25 |
|
26 |
# Inisialisasi pipeline untuk setiap model
|
@@ -34,10 +37,21 @@ except Exception as e:
|
|
34 |
logger.error(f"Model initialization failed: {str(e)}")
|
35 |
raise Exception(f"Model initialization failed: {str(e)}")
|
36 |
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
37 |
def detect_language(text: str) -> str:
|
38 |
"""Deteksi bahasa dari teks menggunakan langdetect."""
|
39 |
try:
|
40 |
lang = langdetect.detect(text)
|
|
|
|
|
|
|
41 |
return lang if lang in MODEL_MAP else "en" # Default ke Inggris jika tidak didukung
|
42 |
except Exception as e:
|
43 |
logger.warning(f"Language detection failed: {str(e)}, defaulting to English")
|
@@ -54,11 +68,16 @@ async def translate(text: str):
|
|
54 |
# Jika sudah Inggris atau bahasa tidak didukung, kembalikan teks asli
|
55 |
if source_lang == "en":
|
56 |
return {"translated_text": text}
|
|
|
|
|
|
|
|
|
|
|
57 |
# Terjemahkan ke Inggris
|
58 |
translator = translators.get(source_lang)
|
59 |
if not translator:
|
60 |
raise HTTPException(status_code=400, detail=f"Translation not supported for language: {source_lang}")
|
61 |
-
result = translator(
|
62 |
return {"translated_text": result[0]["translation_text"]}
|
63 |
except Exception as e:
|
64 |
logger.error(f"Processing failed: {str(e)}")
|
|
|
3 |
import langdetect
|
4 |
import logging
|
5 |
import os
|
6 |
+
import opencc # Untuk konversi Mandarin Sederhana/Tradisional
|
7 |
|
8 |
# Atur direktori cache
|
9 |
os.environ["HF_HOME"] = "/app/cache"
|
|
|
21 |
"fr": "Helsinki-NLP/opus-mt-fr-en", # Prancis ke Inggris
|
22 |
"es": "Helsinki-NLP/opus-mt-es-en", # Spanyol ke Inggris
|
23 |
"ja": "Helsinki-NLP/opus-mt-ja-en", # Jepang ke Inggris
|
24 |
+
"zh-CN": "Helsinki-NLP/opus-mt-zh-en", # Mandarin Sederhana ke Inggris
|
25 |
+
"zh-TW": "Helsinki-NLP/opus-mt-zh-en", # Mandarin Tradisional ke Inggris
|
26 |
+
"vi": "Helsinki-NLP/opus-mt-vi-en", # Vietnam ke Inggris
|
27 |
}
|
28 |
|
29 |
# Inisialisasi pipeline untuk setiap model
|
|
|
37 |
logger.error(f"Model initialization failed: {str(e)}")
|
38 |
raise Exception(f"Model initialization failed: {str(e)}")
|
39 |
|
40 |
+
# Inisialisasi konverter OpenCC untuk Mandarin Tradisional ke Sederhana
|
41 |
+
converter = opencc.OpenCC('t2s') # Tradisional ke Sederhana
|
42 |
+
|
43 |
+
def is_traditional_chinese(text: str) -> bool:
|
44 |
+
"""Cek apakah teks menggunakan karakter Mandarin Tradisional."""
|
45 |
+
traditional_chars = set('繁體字') # Contoh karakter Tradisional
|
46 |
+
return any(char in traditional_chars for char in text)
|
47 |
+
|
48 |
def detect_language(text: str) -> str:
|
49 |
"""Deteksi bahasa dari teks menggunakan langdetect."""
|
50 |
try:
|
51 |
lang = langdetect.detect(text)
|
52 |
+
# Jika terdeteksi sebagai Mandarin (zh), cek apakah Tradisional atau Sederhana
|
53 |
+
if lang == "zh":
|
54 |
+
return "zh-TW" if is_traditional_chinese(text) else "zh-CN"
|
55 |
return lang if lang in MODEL_MAP else "en" # Default ke Inggris jika tidak didukung
|
56 |
except Exception as e:
|
57 |
logger.warning(f"Language detection failed: {str(e)}, defaulting to English")
|
|
|
68 |
# Jika sudah Inggris atau bahasa tidak didukung, kembalikan teks asli
|
69 |
if source_lang == "en":
|
70 |
return {"translated_text": text}
|
71 |
+
# Jika Mandarin Tradisional, konversi ke Sederhana
|
72 |
+
input_text = text
|
73 |
+
if source_lang == "zh-TW":
|
74 |
+
input_text = converter.convert(text)
|
75 |
+
logger.info("Converted Traditional Chinese to Simplified Chinese")
|
76 |
# Terjemahkan ke Inggris
|
77 |
translator = translators.get(source_lang)
|
78 |
if not translator:
|
79 |
raise HTTPException(status_code=400, detail=f"Translation not supported for language: {source_lang}")
|
80 |
+
result = translator(input_text)
|
81 |
return {"translated_text": result[0]["translation_text"]}
|
82 |
except Exception as e:
|
83 |
logger.error(f"Processing failed: {str(e)}")
|