Making Conversation possible to create directly a full conversation (#…

…9434)

* Cleaning up conversation tests.

* Adding tests that don't require downloading models + conversation can be

fully created from static state.

* Making tests non flaky (by fixing generation length)

* Bumping isort version.

* Doc cleanup.

* Remove unused test in this PR.

* Torch import guard for TF.

* Missing torch guard.

* Small mistake in doc.

* Actual uses `_history` and `_index` cache.

+ remove dead enumerate
+ improve warning message.

* Update src/transformers/pipelines/conversational.py

Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>

* Update src/transformers/pipelines/conversational.py

Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>

* Update src/transformers/pipelines/conversational.py

Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>

* Adding comments and cleaner code to address history copy.

* Improving pipeline name in tests.

* Change tokenizer to a real one (still created at runtime with no

external dependency)

* Simplify DummyTok, reverse changes on tokenization.

* Removing DummyTok.

Co-authored-by: Sylvain Gugger <35901082+sgugger@users.noreply.github.com>

src/transformers/pipelines/conversational.py

@@ -33,6 +33,14 @@ class Conversation:
         conversation_id (:obj:`uuid.UUID`, `optional`):
             Unique identifier for the conversation. If not provided, a random UUID4 id will be assigned to the
             conversation.
+        past_user_inputs (:obj:`List[str]`, `optional`):
+            Eventual past history of the conversation of the user. You don't need to pass it manually if you use the
+            pipeline interactively but if you want to recreate history you need to set both :obj:`past_user_inputs` and
+            :obj:`generated_responses` with equal length lists of strings
+        generated_responses (:obj:`List[str]`, `optional`):
+            Eventual past history of the conversation of the model. You don't need to pass it manually if you use the
+            pipeline interactively but if you want to recreate history you need to set both :obj:`past_user_inputs` and
+            :obj:`generated_responses` with equal length lists of strings
 
     Usage::
 
@@ -47,14 +55,33 @@ class Conversation:
         conversation.add_user_input("Is it good?")
     """
 
-    def __init__(self, text: str = None, conversation_id: uuid.UUID = None):
+    def __init__(
+        self, text: str = None, conversation_id: uuid.UUID = None, past_user_inputs=None, generated_responses=None
+    ):
         if not conversation_id:
             conversation_id = uuid.uuid4()
+        if past_user_inputs is None:
+            past_user_inputs = []
+        if generated_responses is None:
+            generated_responses = []
+
         self.uuid: uuid.UUID = conversation_id
-        self.past_user_inputs: List[str] = []
-        self.generated_responses: List[str] = []
-        self.history: List[int] = []
+        self.past_user_inputs: List[str] = past_user_inputs
+        self.generated_responses: List[str] = generated_responses
         self.new_user_input: Optional[str] = text
+        self._index: int = 0
+        self._history: List[int] = []
+
+    def __eq__(self, other):
+        if not isinstance(other, Conversation):
+            return False
+        if self.uuid == other.uuid:
+            return True
+        return (
+            self.new_user_input == other.new_user_input
+            and self.past_user_inputs == other.past_user_inputs
+            and self.generated_responses == other.generated_responses
+        )
 
     def add_user_input(self, text: str, overwrite: bool = False):
         """
@@ -100,16 +127,6 @@ def append_response(self, response: str):
         """
         self.generated_responses.append(response)
 
-    def set_history(self, history: List[int]):
-        """
-        Updates the value of the history of the conversation. The history is represented by a list of :obj:`token_ids`.
-        The history is used by the model to generate responses based on the previous conversation turns.
-
-        Args:
-            history (:obj:`List[int]`): History of tokens provided and generated for this conversation.
-        """
-        self.history = history
-
     def __repr__(self):
         """
         Generates a string representation of the conversation.
@@ -167,12 +184,40 @@ def __init__(self, min_length_for_response=32, *args, **kwargs):
         super().__init__(*args, **kwargs)
 
         # We need at least an eos_token
-        assert self.tokenizer.eos_token_id is not None, "DialoguePipeline tokenizer should have an EOS token set"
+        assert self.tokenizer.eos_token_id is not None, "ConversationalPipeline tokenizer should have an EOS token set"
         if self.tokenizer.pad_token_id is None:
             self.tokenizer.pad_token = self.tokenizer.eos_token
 
         self.min_length_for_response = min_length_for_response
 
+    def _get_history(self, conversation):
+        """
+        Private function (subject to change) that simply tokenizes and concatenates past inputs. Also saves that
+        tokenization into the conversation state.
+
+        Args:
+            conversation (:class:`~transformers.Conversation`)
+
+        Returns:
+            :obj:`List[int]`: The list of tokens for the past input of that conversation.
+        """
+        # Make a copy to prevent messing cache up if there's an error
+        # within this function
+        history = conversation._history.copy()
+        index = conversation._index
+        new_index = index
+        for i, (past_user_input, generated_response) in enumerate(
+            zip(conversation.past_user_inputs[index:], conversation.generated_responses[index:])
+        ):
+            for el in (past_user_input, generated_response):
+                new_history = self._parse_and_tokenize([el])[0]
+                history.extend(new_history)
+            new_index = i + index + 1
+        conversation._index = new_index
+        conversation._history = history
+        # Hand back a copy to caller so they can't accidently modify our cache.
+        return history.copy()
+
     def __call__(
         self,
         conversations: Union[Conversation, List[Conversation]],
@@ -220,7 +265,7 @@ def __call__(
         with self.device_placement():
 
             inputs = self._parse_and_tokenize([conversation.new_user_input for conversation in conversations])
-            histories = [conversation.history for conversation in conversations]
+            histories = [self._get_history(conversation) for conversation in conversations]
             max_length = generate_kwargs.get("max_length", self.model.config.max_length)
             inputs = self._concat_inputs_history(inputs, histories, max_length)
 
@@ -266,7 +311,6 @@ def __call__(
                         clean_up_tokenization_spaces=clean_up_tokenization_spaces,
                     )
                 )
-                conversation.set_history(history[conversation_index])
                 output.append(conversation)
             if len(output) == 1:
                 return output[0]
@@ -333,7 +377,9 @@ def _concat_inputs_history(self, inputs: List[List[int]], histories: List[Option
                     if cutoff_eos_index == 0 or cutoff_eos_index == len(new_input) - 1:
                         break
                     else:
-                        new_input = new_input[cutoff_eos_index + 1 :]
+                        logger.warning(
+                            f"Cutting history off because it's too long ({len(new_input)} > {max_length - self.min_length_for_response}) for underlying model"
+                        )
             outputs.append(new_input)
         padded_outputs = self.tokenizer.pad(
             {"input_ids": outputs}, padding="longest", return_attention_mask=True, return_tensors=self.framework

tests/test_pipelines_conversational.py

@@ -14,15 +14,177 @@
 
 import unittest
 
-from transformers import AutoModelForSeq2SeqLM, AutoTokenizer, Conversation, ConversationalPipeline, pipeline
+from transformers import (
+    AutoModelForSeq2SeqLM,
+    AutoTokenizer,
+    Conversation,
+    ConversationalPipeline,
+    is_torch_available,
+    pipeline,
+)
 from transformers.testing_utils import require_torch, slow, torch_device
 
 from .test_pipelines_common import MonoInputPipelineCommonMixin
 
 
+if is_torch_available():
+    import torch
+
+    from transformers.models.gpt2 import GPT2Config, GPT2LMHeadModel
+
 DEFAULT_DEVICE_NUM = -1 if torch_device == "cpu" else 0
 
 
+class SimpleConversationPipelineTests(unittest.TestCase):
+    def get_pipeline(self):
+        # When
+        config = GPT2Config(
+            vocab_size=263,
+            n_ctx=128,
+            max_length=128,
+            n_embd=64,
+            n_layer=1,
+            n_head=8,
+            bos_token_id=256,
+            eos_token_id=257,
+        )
+        model = GPT2LMHeadModel(config)
+        # Force model output to be L
+        V, D = model.lm_head.weight.shape
+        bias = torch.zeros(V, requires_grad=True)
+        bias[76] = 1
+
+        model.lm_head.bias = torch.nn.Parameter(bias)
+
+        # # Created with:
+        # import tempfile
+
+        # from tokenizers import Tokenizer, models
+        # from transformers.tokenization_utils_fast import PreTrainedTokenizerFast
+
+        # vocab = [(chr(i), i) for i in range(256)]
+        # tokenizer = Tokenizer(models.Unigram(vocab))
+        # with tempfile.NamedTemporaryFile() as f:
+        #     tokenizer.save(f.name)
+        #     real_tokenizer = PreTrainedTokenizerFast(tokenizer_file=f.name, eos_token="<eos>", bos_token="<bos>")
+
+        # real_tokenizer._tokenizer.save("dummy.json")
+        # Special tokens are automatically added at load time.
+        tokenizer = AutoTokenizer.from_pretrained("Narsil/small_conversational_test")
+        conversation_agent = pipeline(
+            task="conversational", device=DEFAULT_DEVICE_NUM, model=model, tokenizer=tokenizer
+        )
+        return conversation_agent
+
+    @require_torch
+    def test_integration_torch_conversation(self):
+        conversation_agent = self.get_pipeline()
+        conversation_1 = Conversation("Going to the movies tonight - any suggestions?")
+        conversation_2 = Conversation("What's the last book you have read?")
+        self.assertEqual(len(conversation_1.past_user_inputs), 0)
+        self.assertEqual(len(conversation_2.past_user_inputs), 0)
+
+        with self.assertLogs("transformers", level="WARNING") as log:
+            result = conversation_agent([conversation_1, conversation_2], max_length=48)
+            self.assertEqual(len(log.output), 2)
+            self.assertIn("You might consider trimming the early phase of the conversation", log.output[0])
+            self.assertIn("Setting `pad_token_id`", log.output[1])
+
+        # Two conversations in one pass
+        self.assertEqual(result, [conversation_1, conversation_2])
+        self.assertEqual(
+            result,
+            [
+                Conversation(
+                    None,
+                    past_user_inputs=["Going to the movies tonight - any suggestions?"],
+                    generated_responses=["L"],
+                ),
+                Conversation(
+                    None, past_user_inputs=["What's the last book you have read?"], generated_responses=["L"]
+                ),
+            ],
+        )
+
+        # One conversation with history
+        conversation_2.add_user_input("Why do you recommend it?")
+        with self.assertLogs("transformers", level="WARNING") as log:
+            result = conversation_agent(conversation_2, max_length=64)
+            self.assertEqual(len(log.output), 3)
+            self.assertIn("Cutting history off because it's too long", log.output[0])
+            self.assertIn("You might consider trimming the early phase of the conversation", log.output[1])
+            self.assertIn("Setting `pad_token_id`", log.output[2])
+
+        self.assertEqual(result, conversation_2)
+        self.assertEqual(
+            result,
+            Conversation(
+                None,
+                past_user_inputs=["What's the last book you have read?", "Why do you recommend it?"],
+                generated_responses=["L", "L"],
+            ),
+        )
+
+    @require_torch
+    def test_history_cache(self):
+        conversation_agent = self.get_pipeline()
+        conversation = Conversation(
+            "Why do you recommend it?",
+            past_user_inputs=["What's the last book you have read?"],
+            generated_responses=["b"],
+        )
+        with self.assertLogs("transformers", level="WARNING") as log:
+            _ = conversation_agent(conversation, max_length=60)
+            self.assertEqual(len(log.output), 3)
+            self.assertIn("Cutting history off because it's too long (63 > 28) for underlying model", log.output[0])
+            self.assertIn("63 is bigger than 0.9 * max_length: 60", log.output[1])
+            self.assertIn("Setting `pad_token_id`", log.output[2])
+        self.assertEqual(conversation._index, 1)
+        self.assertEqual(
+            conversation._history,
+            [
+                87,
+                104,
+                97,
+                116,
+                39,
+                115,
+                32,
+                116,
+                104,
+                101,
+                32,
+                108,
+                97,
+                115,
+                116,
+                32,
+                98,
+                111,
+                111,
+                107,
+                32,
+                121,
+                111,
+                117,
+                32,
+                104,
+                97,
+                118,
+                101,
+                32,
+                114,
+                101,
+                97,
+                100,
+                63,
+                259,  # EOS
+                98,  # b
+                259,  # EOS
+            ],
+        )
+
+
 class ConversationalPipelineTests(MonoInputPipelineCommonMixin, unittest.TestCase):
     pipeline_task = "conversational"
     small_models = []  # Models tested without the @slow decorator