Merge branch 'improve_error_handling' into 'master'

Improve error handling

See merge request !12

texta_mlp/document.py

@@ -51,6 +51,7 @@ class Document:
             doc_path: str = "text",
             json_doc: dict = None,
             analyzers: list = [],
+            error: str = "",
     ):
 
         self.original_text = original_text
@@ -58,6 +59,7 @@ class Document:
         self.analyzers = analyzers
         self.dominant_language_code = dominant_language_code
         self.analysis_lang = analysis_lang
+        self.error = error
         self.json_doc = json_doc
 
         self.entity_mapper = entity_mapper
@@ -166,6 +168,8 @@ class Document:
             "detected": self.dominant_language_code,
             "analysis": self.analysis_lang
         }
+        if self.error:
+            container["error"] = self.error
 
         if "lemmas" in self.analyzers:
             container["lemmas"] = self.get_lemma()

texta_mlp/exceptions.py

@@ -5,3 +5,7 @@ class LanguageNotSupported(Exception):
 class BoundedListEmpty(Exception):
     """Raised when in Concatenator class the BOUNDS are not yet loaded, but concatenate() is tried""" 
     pass
+
+class StanzaPipelineFail(Exception):
+    """Raised when Stanza pipelines fail to load."""
+    pass

texta_mlp/mlp.py

@@ -202,10 +202,10 @@ class MLP:
         '''
         if lang not in self.supported_langs:
             analysis_lang = self.default_lang
-            sentences, entities = self._get_stanza_tokens(analysis_lang, processed_text) if processed_text else ([], [])
+            sentences, entities, e = self._get_stanza_tokens(analysis_lang, processed_text) if processed_text else ([], [], "")
         else:
             analysis_lang = lang
-            sentences, entities = self._get_stanza_tokens(analysis_lang, processed_text) if processed_text else ([], [])
+            sentences, entities, e = self._get_stanza_tokens(analysis_lang, processed_text) if processed_text else ([], [], "")
 
         document = Document(
             original_text=processed_text,
@@ -217,7 +217,8 @@ class MLP:
             json_doc=json_object,
             doc_path=doc_paths,
             entity_mapper=self.entity_mapper,
-            concat_resources=self.concat_resources
+            concat_resources=self.concat_resources,
+            error=e
         )
         return document
 
@@ -230,19 +231,26 @@ class MLP:
 
 
     def _get_stanza_tokens(self, lang: str, raw_text: str):
-        pipeline = self.stanza_pipelines[lang](raw_text)
-
         sentences = []
         entities = []
-        pip_pat = re.compile(r"(?<=\d)_(?=\d)")
-        for sentence in pipeline.sentences:
-            words = []
-            for word in sentence.words:
-                words.append(word)
-            sentences.append(words)
-            for entity in sentence.entities:
-                entities.append(entity)
-        return sentences, entities
+        e = ""
+        try:
+            pipeline = self.stanza_pipelines[lang](raw_text)
+
+            pip_pat = re.compile(r"(?<=\d)_(?=\d)")
+            for sentence in pipeline.sentences:
+                words = []
+                for word in sentence.words:
+                    words.append(word)
+                sentences.append(words)
+                for entity in sentence.entities:
+                    entities.append(entity)
+
+        except Exception as e:
+            self.logger.exception(e)
+            return sentences, entities, repr(e)
+
+        return sentences, entities, e
 
 
     def _get_stanza_ner(self, lang: str, raw_text: str):