Add regression tests for slow sentencepiece tokenizers. (#11737)

* add test_vocab_size for sentencepiece tok.

* add test_get_vocab for sentencepiece tok.

* add test_convert_token_and_id for sentencepiece tok.

* add test_tokenize_and_convert_tokens_to_string for all tok.

* improve test_tokenize_and_convert_tokens_to_string for sp. tok.

* add common tokenizer integration tests
- for albert
- for barthez

* add tokenizer integration tests to bert gen.

* add most tokenizer integration tests

* fix camembert tokenizer integration test

* add tokenizer integration test to marian

* add tokenizer integration test to reformer

* add typing and doc to tokenizer_integration_test_util

* fix tokenizer integration test of reformer

* improve test_sentencepiece_tokenize_and_convert_tokens_to_string

* empty commit to trigger CI

* fix tokenizer integration test of reformer

* remove code not needed anymore

* empty commit to trigger CI

* empty commit to trigger CI

tests/test_tokenization_camembert.py

@@ -18,7 +18,7 @@ import unittest
 
 from transformers import CamembertTokenizer, CamembertTokenizerFast
 from transformers.file_utils import is_torch_available
-from transformers.testing_utils import require_sentencepiece, require_tokenizers
+from transformers.testing_utils import require_sentencepiece, require_tokenizers, slow
 
 from .test_tokenization_common import TokenizerTesterMixin
 
@@ -45,6 +45,25 @@ class CamembertTokenizationTest(TokenizerTesterMixin, unittest.TestCase):
         tokenizer = CamembertTokenizer(SAMPLE_VOCAB)
         tokenizer.save_pretrained(self.tmpdirname)
 
+    def test_convert_token_and_id(self):
+        """Test ``_convert_token_to_id`` and ``_convert_id_to_token``."""
+        token = "<pad>"
+        token_id = 1
+
+        self.assertEqual(self.get_tokenizer()._convert_token_to_id(token), token_id)
+        self.assertEqual(self.get_tokenizer()._convert_id_to_token(token_id), token)
+
+    def test_get_vocab(self):
+        vocab_keys = list(self.get_tokenizer().get_vocab().keys())
+
+        self.assertEqual(vocab_keys[0], "<s>NOTUSED")
+        self.assertEqual(vocab_keys[1], "<pad>")
+        self.assertEqual(vocab_keys[-1], "<mask>")
+        self.assertEqual(len(vocab_keys), 1_004)
+
+    def test_vocab_size(self):
+        self.assertEqual(self.get_tokenizer().vocab_size, 1_005)
+
     def test_rust_and_python_bpe_tokenizers(self):
         tokenizer = CamembertTokenizer(SAMPLE_BPE_VOCAB)
         tokenizer.save_pretrained(self.tmpdirname)
@@ -88,3 +107,25 @@ class CamembertTokenizationTest(TokenizerTesterMixin, unittest.TestCase):
         ids = tokenizer.encode(sequence)
         rust_ids = rust_tokenizer.encode(sequence)
         self.assertListEqual(ids, rust_ids)
+
+    @slow
+    def test_tokenizer_integration(self):
+        # fmt: off
+        expected_encoding = {'input_ids': [[5, 54, 7196, 297, 30, 23, 776, 18, 11, 3215, 3705, 8252, 22, 3164, 1181, 2116, 29, 16, 813, 25, 791, 3314, 20, 3446, 38, 27575, 120, 6, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1], [5, 468, 17, 11, 9088, 20, 1517, 8, 22804, 18818, 10, 38, 629, 607, 607, 142, 19, 7196, 867, 56, 10326, 24, 2267, 20, 416, 5072, 15612, 233, 734, 7, 2399, 27, 16, 3015, 1649, 7, 24, 20, 4338, 2399, 27, 13, 3400, 14, 13, 6189, 8, 930, 9, 6]], 'attention_mask': [[1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0], [1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1, 1]]}  # noqa: E501
+        # fmt: on
+
+        # camembert is a french model. So we also use french texts.
+        sequences = [
+            "Le transformeur est un modèle d'apprentissage profond introduit en 2017, "
+            "utilisé principalement dans le domaine du traitement automatique des langues (TAL).",
+            "À l'instar des réseaux de neurones récurrents (RNN), les transformeurs sont conçus "
+            "pour gérer des données séquentielles, telles que le langage naturel, pour des tâches "
+            "telles que la traduction et la synthèse de texte.",
+        ]
+
+        self.tokenizer_integration_test_util(
+            expected_encoding=expected_encoding,
+            model_name="camembert-base",
+            revision="3a0641d9a1aeb7e848a74299e7e4c4bca216b4cf",
+            sequences=sequences,
+        )