Add custom detection function for language detection

.gitignore

@@ -7,3 +7,4 @@ __pycache__/
 .coverage
 .coverage.*
 *.pyc
+.vscode/

src/python/txtai/pipeline/text/translation.py

@@ -114,17 +114,21 @@ class Translation(HFModel):
         ids = [x.modelId for x in HfApi().list_models(author="Helsinki-NLP")] if self.findmodels else []
         return set(ids)
 
-    def detect(self, texts):
+    def detect(self, texts, custom_detect=None):
         """
         Detects the language for each element in texts.
 
         Args:
             texts: list of text
+            custom_detect: A function to use custom model to detect language
 
         Returns:
             list of languages
         """
 
+        if custom_detect is not None:
+            return custom_detect(texts)
+
         if not FASTTEXT:
             raise ImportError('Language detection is not available - install "pipeline" extra to enable')
 

test/python/testpipeline/testtranslation.py

@@ -13,6 +13,31 @@ class TestTranslation(unittest.TestCase):
     Translation tests.
     """
 
+    def testDetect(self):
+        """
+        Test language detection
+        """
+        translate = Translation()
+
+        test = ["This is a test language detection."]
+        language = translate.detect(test)
+
+        self.assertListEqual(language, ["en"])
+
+    def testDetectWithCustomFunc(self):
+        """
+        Test language detection with custom function
+        """
+        translate = Translation()
+
+        def dummy_func(text):
+            return ["en" for x in text]
+
+        test = ["This is a test language detection."]
+        language = translate.detect(test, dummy_func)
+
+        self.assertListEqual(language, ["en"])
+
     def testLongTranslation(self):
         """
         Test a translation longer than max tokenization length