feat: language detection

app.py

@@ -92,7 +92,7 @@ def check_errors(text):
         gr.Warning(f'La longitud del texto ({len(text)} caracteres) sobrepasa el máximo permitido.')
         error = True
     if not CheckLanguageIsSpanish().detect_spanish(text):
-        gr.Warning('El texto tiene que estar en Español.')
+        gr.Warning('El texto está en Ingles o posee oraciones en dicho idioma.')
         error = True
     return error
 

utils.py

@@ -1,37 +1,25 @@
-# import spacy
-# from spacy.language import Language
-# from spacy_langdetect import LanguageDetector
+from lingua import Language, LanguageDetectorBuilder
 
-# class CheckLanguageIsSpanish:
-#     def __init__(self):
-
-#         self.nlp_es = spacy.load("es_core_news_sm")
-#         Language.factory("language_detector", func=self.get_lang_detector)
-#         self.nlp_es.add_pipe('language_detector', last=True)
-
-#     def get_lang_detector(self, nlp, name):
-#         return LanguageDetector()
+class CheckLanguageIsSpanish:
+    def __init__(self):
+        self.languages = [Language.SPANISH, Language.ENGLISH]
+        self.detector = LanguageDetectorBuilder.from_languages(*self.languages).build()
 
-#     def detect_spanish(self, text):
-#         doc = self.nlp_es(text)
-#         lang_prediction = doc._.language
-#         confidence = lang_prediction['score'] * 100
-#         if confidence > 95:
-#             print(f"espanio perri {confidence}")
-#             return True
-#         else:
-#             print(f"no espanio perri {confidence}")
-#             return False
-    
-from langdetect import detect_langs
+    def detect_english(self, text):
+        spanish = True
 
-class CheckLanguageIsSpanish:
-    def detect_spanish(self, text, confidence_threshold=0.85):
-        lang_results = detect_langs(text)
-        top_language = lang_results[0].lang
-        top_confidence = lang_results[0].prob
+        # Compares confidence values between languages
+        confidence_values = self.detector.compute_language_confidence_values(text)
+        confidence_dict = {confidence.language.name: confidence.value for confidence in confidence_values}
+        if confidence_dict["ENGLISH"] > confidence_dict["SPANISH"]:
+            spanish = False
 
-        if top_language == 'es' and top_confidence >= confidence_threshold:
-            return True
-        else:
-            return False
+        # Checks if there is at least one sentence with 5 words or more in English
+        languages = self.detector.detect_multiple_languages_of(text)
+        for result in languages:
+            if result.language.name == "ENGLISH":
+                if result.word_count >= 5:
+                    spanish = False
+                    break
+        
+        return spanish