Update app.py

app.py

@@ -42,25 +42,79 @@ Response Rules:
 app = Flask(__name__, template_folder="templates", static_folder="static")
 CORS(app)
 
-# Translation function with auto language detection
-def translate_text(text, target_lang, source_lang='auto'):
+# Language detection function
+def detect_language(text):
+    """
+    Detect language of text using multiple methods
+    """
+    try:
+        # Method 1: Try langdetect if available
+        try:
+            from langdetect import detect
+            detected = detect(text)
+            return detected
+        except ImportError:
+            pass
+        except:
+            pass
+        
+        # Method 2: Simple heuristic detection
+        common_words = {
+            'en': ['the', 'and', 'or', 'but', 'in', 'on', 'at', 'to', 'for', 'of', 'with', 'by', 'is', 'are', 'was', 'were'],
+            'es': ['el', 'la', 'de', 'que', 'y', 'en', 'un', 'es', 'se', 'no', 'te', 'lo', 'por', 'con', 'del', 'las'],
+            'fr': ['le', 'de', 'et', 'à', 'un', 'il', 'être', 'en', 'avoir', 'que', 'pour', 'du', 'ce', 'son', 'une'],
+            'de': ['der', 'die', 'und', 'in', 'den', 'von', 'zu', 'das', 'mit', 'sich', 'des', 'auf', 'für', 'ist'],
+            'it': ['il', 'di', 'che', 'e', 'la', 'per', 'in', 'un', 'è', 'con', 'non', 'da', 'del', 'sono'],
+            'pt': ['de', 'a', 'o', 'que', 'e', 'do', 'da', 'em', 'um', 'para', 'é', 'com', 'os', 'uma', 'ser'],
+            'ru': ['в', 'и', 'не', 'на', 'я', 'быть', 'с', 'а', 'как', 'его', 'к', 'он', 'что', 'то'],
+            'hi': ['है', 'का', 'के', 'में', 'और', 'को', 'से', 'पर', 'यह', 'वह', 'एक', 'जो'],
+            'ar': ['في', 'من', 'إلى', 'على', 'هذا', 'هذه', 'التي', 'الذي', 'أن', 'كان', 'مع'],
+            'zh': ['的', '是', '在', '了', '和', '有', '人', '我', '你', '他', '她', '它'],
+        }
+        
+        text_lower = text.lower()
+        scores = {}
+        
+        for lang, words in common_words.items():
+            score = sum(1 for word in words if word in text_lower)
+            if score > 0:
+                scores[lang] = score / len(words)  # Normalize by word count
+        
+        if scores:
+            return max(scores, key=scores.get)
+        
+        return 'en'  # Default to English if detection fails
+        
+    except Exception:
+        return 'en'  # Default to English
+
+# Translation function with proper language detection
+def translate_text(text, target_lang, source_lang=None):
     """
     Translate text to desired language using Microsoft Translator
     
     Args:
         text (str): Text to translate
         target_lang (str): Target language code (e.g., 'es', 'fr', 'de', 'hi', 'zh')
-        source_lang (str): Source language code (default: 'auto' for auto-detection)
+        source_lang (str): Source language code (if None, will auto-detect)
     
     Returns:
         str: Translated text or original text if translation fails
     """
     try:
-        # Skip translation if target language is English and text appears to be in English
+        # Skip translation if target language is English and source appears to be English
         if target_lang.lower() == 'en':
             return text
+        
+        # Detect source language if not provided
+        if source_lang is None or source_lang == 'auto':
+            source_lang = detect_language(text)
+        
+        # Skip translation if source and target are the same
+        if source_lang.lower() == target_lang.lower():
+            return text
             
-        # Create translator instance
+        # Create translator instance with detected/specified source language
         translator = Translator(to_lang=target_lang, from_lang=source_lang)
         translated = translator.translate(text)
         
@@ -72,7 +126,15 @@ def translate_text(text, target_lang, source_lang='auto'):
             
     except Exception as e:
         print(f"Translation error: {str(e)}")
-        return text  # Return original text if translation fails
+        # Try fallback translation with English as source
+        try:
+            fallback_translator = Translator(to_lang=target_lang, from_lang='en')
+            fallback_result = fallback_translator.translate(text)
+            if fallback_result and fallback_result.strip():
+                return fallback_result
+        except:
+            pass
+        return text  # Return original text if all translation attempts fail
 
 # Language code mapping for common languages
 LANGUAGE_CODES = {
@@ -129,11 +191,17 @@ def chat():
         processed_message = user_message
         if target_lang != "en":
             try:
-                # Try to translate user message to English for better AI understanding
-                english_translator = Translator(to_lang='en', from_lang='auto')
-                processed_message = english_translator.translate(user_message)
-                if not processed_message or not processed_message.strip():
-                    processed_message = user_message  # Fallback to original
+                # Detect source language of user message
+                detected_source = detect_language(user_message)
+                print(f"Detected user language: {detected_source}")
+                
+                # Translate to English only if source is not English
+                if detected_source != 'en':
+                    english_translator = Translator(to_lang='en', from_lang=detected_source)
+                    processed_message = english_translator.translate(user_message)
+                    if not processed_message or not processed_message.strip():
+                        processed_message = user_message  # Fallback to original
+                    print(f"Translated user message to English: {processed_message}")
             except Exception as e:
                 print(f"User message translation error: {str(e)}")
                 processed_message = user_message  # Use original message
@@ -180,7 +248,7 @@ def translate_endpoint():
         data = request.get_json()
         text = data.get("text")
         target_lang = data.get("target_lang", "en")
-        source_lang = data.get("source_lang", "auto")
+        source_lang = data.get("source_lang")  # Remove default 'auto'
         
         if not text:
             return jsonify({"error": "Text is required"}), 400
@@ -189,13 +257,17 @@ def translate_endpoint():
         target_lang = target_lang.lower()
         if target_lang in LANGUAGE_CODES:
             target_lang = LANGUAGE_CODES[target_lang]
+        
+        # Auto-detect source language if not provided
+        if not source_lang or source_lang.lower() == 'auto':
+            source_lang = detect_language(text)
             
         translated = translate_text(text, target_lang, source_lang)
         
         return jsonify({
             "original_text": text,
             "translated_text": translated,
-            "source_lang": source_lang,
+            "detected_source_lang": source_lang,
             "target_lang": target_lang
         })