Truncate text in language detection logging

dataiku · Sep 30, 2020 · 28ef5a1 · 28ef5a1
1 parent b730ad2
commit 28ef5a1
Showing 1 changed file with 2 additions and 2 deletions.
diff --git a/python-lib/language_detector.py b/python-lib/language_detector.py
@@ -14,7 +14,7 @@
     LANGUAGE_REMAPPING_PYCLD3_LANGID,
 )
 
-from plugin_io_utils import generate_unique
+from plugin_io_utils import generate_unique, truncate_text_list
 
 
 class LanguageDetector:
@@ -66,7 +66,7 @@ def _cld3_detection(self, doc: AnyStr) -> (AnyStr, float):
 
     def _detection_filter(self, doc: AnyStr, lang_id: AnyStr, lang_probability: float) -> (AnyStr, float):
         if lang_probability < self.minimum_score or lang_id not in self.language_scope:
-            warning_msg = "Problem encountered for document: '{}'.\n".format(doc)
+            warning_msg = "Problem encountered for document: '{}'.\n".format(truncate_text_list([doc])[0])
             if lang_id not in self.language_scope:
                 warning_msg += "Detected language: '{}' not within language scope: {}.\n".format(
                     lang_id, self.language_scope