Update translation.py

translation.py

@@ -2,9 +2,15 @@ import streamlit as st
 from transformers import MarianTokenizer, MarianMTModel
 import torch
 
+LANGUAGES = {
+    "en": ("English", "English"), "fr": ("Français", "French"), "es": ("Español", "Spanish"),
+    "de": ("Deutsch", "German"), "hi": ("हिन्दी", "Hindi"), "zh": ("中文", "Chinese"),
+    "ar": ("العربية", "Arabic"), "ru": ("Русский", "Russian"), "ja": ("日本語", "Japanese")
+}
+
 @st.cache_resource
 def _load_default_model():
-    model_name = "Helsinki-NLP/opus-mt-en-fr"
+    model_name = "Helsinki-NLP/opus-mt-en-hi"  
     tokenizer = MarianTokenizer.from_pretrained(model_name)
     model = MarianMTModel.from_pretrained(model_name)
     return tokenizer, model
@@ -19,21 +25,17 @@ def load_model(source_lang, target_lang):
         model = MarianMTModel.from_pretrained(model_name)
         return tokenizer, model
     except Exception as e:
-        st.warning(f"No direct model for {source_lang} to {target_lang}. Falling back to English buffer.")
+        st.warning(f"No direct model for {source_lang} to {target_lang}. Using en-hi fallback.")
         return _load_default_model()
 
 @st.cache_data(ttl=3600)
 def translate_cached(text, source_lang, target_lang):
-    src_code = {"English": "en", "French": "fr", "Spanish": "es", "German": "de",
-                "Hindi": "hi", "Chinese": "zh", "Arabic": "ar", "Russian": "ru", "Japanese": "ja"}.get(source_lang, "en")
-    tgt_code = {"English": "en", "French": "fr", "Spanish": "es", "German": "de",
-                "Hindi": "hi", "Chinese": "zh", "Arabic": "ar", "Russian": "ru", "Japanese": "ja"}.get(target_lang, "fr")
-    tokenizer, model = load_model(src_code, tgt_code)
+    tokenizer, model = load_model(source_lang, target_lang)
     inputs = tokenizer(text, return_tensors="pt", padding=True, truncation=True, max_length=500)
     with torch.no_grad():
         translated = model.generate(**inputs, max_length=500, num_beams=2, early_stopping=True)
     translated_text = tokenizer.decode(translated[0], skip_special_tokens=True)
-    return translated_text if translated_text.strip() else text  # Fallback to input if empty
+    return translated_text if translated_text.strip() and len(translated_text.split()) >= 2 else text
 
 def translate(text, source_lang, target_lang):
     if not text:
@@ -43,6 +45,3 @@ def translate(text, source_lang, target_lang):
     except Exception as e:
         st.error(f"Translation error: {str(e)}. Using input as fallback.")
         return text
-
-LANGUAGES = {"English": "en", "French": "fr", "Spanish": "es", "German": "de",
-             "Hindi": "hi", "Chinese": "zh", "Arabic": "ar", "Russian": "ru", "Japanese": "ja"}