Spaces:
Running
Running
| from fastapi import FastAPI, HTTPException | |
| from transformers import M2M100ForConditionalGeneration, M2M100Tokenizer | |
| import langdetect | |
| import logging | |
| import os | |
| # Atur direktori cache untuk Hugging Face | |
| os.environ["HF_HOME"] = "/app/cache" | |
| os.environ["TRANSFORMERS_CACHE"] = "/app/cache" | |
| app = FastAPI() | |
| logging.basicConfig(level=logging.INFO) | |
| logger = logging.getLogger(__name__) | |
| try: | |
| logger.info("Loading translation model...") | |
| # Load model dan tokenizer M2M100 | |
| model_name = "facebook/m2m100_418M" | |
| model = M2M100ForConditionalGeneration.from_pretrained(model_name) | |
| tokenizer = M2M100Tokenizer.from_pretrained(model_name) | |
| logger.info("Model loaded successfully") | |
| except Exception as e: | |
| logger.error(f"Failed to load model: {str(e)}") | |
| raise Exception(f"Model initialization failed: {str(e)}") | |
| def detect_language(text: str) -> str: | |
| """Deteksi bahasa dari teks menggunakan langdetect.""" | |
| try: | |
| lang = langdetect.detect(text) | |
| # Konversi kode bahasa ke format M2M100 | |
| lang_map = { | |
| "id": "id", # Indonesia | |
| "fr": "fr", # Prancis | |
| "es": "es", # Spanyol | |
| "de": "de", # Jerman | |
| "ja": "ja", # Jepang | |
| "zh": "zh", # Mandarin | |
| "ru": "ru", # Rusia | |
| "th": "th", # Thailand | |
| } | |
| return lang_map.get(lang, "en") # Default ke Inggris jika tidak dikenali | |
| except Exception as e: | |
| logger.warning(f"Language detection failed: {str(e)}, defaulting to English") | |
| return "en" | |
| def translate_to_english(text: str, source_lang: str) -> str: | |
| """Terjemahkan teks ke Inggris menggunakan M2M100.""" | |
| try: | |
| # Set bahasa sumber | |
| tokenizer.src_lang = source_lang | |
| # Encode teks | |
| encoded = tokenizer(text, return_tensors="pt") | |
| # Generate terjemahan (ke Inggris) | |
| generated_tokens = model.generate(**encoded, forced_bos_token_id=tokenizer.get_lang_id("en")) | |
| # Decode hasil | |
| translated_text = tokenizer.batch_decode(generated_tokens, skip_special_tokens=True)[0] | |
| return translated_text | |
| except Exception as e: | |
| logger.error(f"Translation failed: {str(e)}") | |
| raise HTTPException(status_code=500, detail=f"Translation failed: {str(e)}") | |
| async def translate(text: str): | |
| if not text: | |
| raise HTTPException(status_code=400, detail="Text input is required") | |
| try: | |
| # Deteksi bahasa | |
| source_lang = detect_language(text) | |
| logger.info(f"Detected source language: {source_lang}") | |
| # Jika sudah Inggris, kembalikan teks asli | |
| if source_lang == "en": | |
| return {"translated_text": text} | |
| # Terjemahkan ke Inggris | |
| translated_text = translate_to_english(text, source_lang) | |
| return {"translated_text": translated_text} | |
| except Exception as e: | |
| logger.error(f"Processing failed: {str(e)}") | |
| raise HTTPException(status_code=500, detail=f"Processing failed: {str(e)}") |