Use correct tokenizer for Bedrock/Anthropic LLMs (#11561)

**Description**

This PR implements the usage of the correct tokenizer in Bedrock LLMs,
if using anthropic models.

**Issue:** #11560

**Dependencies:** optional dependency on `anthropic` python library.

**Twitter handle:** jtolgyesi


---------

Co-authored-by: Bagatur <baskaryan@gmail.com>

libs/langchain/langchain/chat_models/bedrock.py

@@ -9,6 +9,10 @@ from langchain.llms.bedrock import BedrockBase
 from langchain.pydantic_v1 import Extra
 from langchain.schema.messages import AIMessage, AIMessageChunk, BaseMessage
 from langchain.schema.output import ChatGeneration, ChatGenerationChunk, ChatResult
+from langchain.utilities.anthropic import (
+    get_num_tokens_anthropic,
+    get_token_ids_anthropic,
+)
 
 
 class ChatPromptAdapter:
@@ -86,3 +90,15 @@ class BedrockChat(BaseChatModel, BedrockBase):
 
         message = AIMessage(content=completion)
         return ChatResult(generations=[ChatGeneration(message=message)])
+
+    def get_num_tokens(self, text: str) -> int:
+        if self._model_is_anthropic:
+            return get_num_tokens_anthropic(text)
+        else:
+            return super().get_num_tokens(text)
+
+    def get_token_ids(self, text: str) -> List[int]:
+        if self._model_is_anthropic:
+            return get_token_ids_anthropic(text)
+        else:
+            return super().get_token_ids(text)

libs/langchain/langchain/llms/bedrock.py

@@ -8,6 +8,10 @@ from langchain.llms.base import LLM
 from langchain.llms.utils import enforce_stop_tokens
 from langchain.pydantic_v1 import BaseModel, Extra, root_validator
 from langchain.schema.output import GenerationChunk
+from langchain.utilities.anthropic import (
+    get_num_tokens_anthropic,
+    get_token_ids_anthropic,
+)
 
 HUMAN_PROMPT = "\n\nHuman:"
 ASSISTANT_PROMPT = "\n\nAssistant:"
@@ -222,6 +226,10 @@ class BedrockBase(BaseModel, ABC):
     def _get_provider(self) -> str:
         return self.model_id.split(".")[0]
 
+    @property
+    def _model_is_anthropic(self) -> bool:
+        return self._get_provider() == "anthropic"
+
     def _prepare_input_and_invoke(
         self,
         prompt: str,
@@ -318,7 +326,7 @@ class Bedrock(LLM, BedrockBase):
             from bedrock_langchain.bedrock_llm import BedrockLLM
 
             llm = BedrockLLM(
-                credentials_profile_name="default", 
+                credentials_profile_name="default",
                 model_id="amazon.titan-text-express-v1",
                 streaming=True
             )
@@ -393,3 +401,15 @@ class Bedrock(LLM, BedrockBase):
             return completion
 
         return self._prepare_input_and_invoke(prompt=prompt, stop=stop, **kwargs)
+
+    def get_num_tokens(self, text: str) -> int:
+        if self._model_is_anthropic:
+            return get_num_tokens_anthropic(text)
+        else:
+            return super().get_num_tokens(text)
+
+    def get_token_ids(self, text: str) -> List[int]:
+        if self._model_is_anthropic:
+            return get_token_ids_anthropic(text)
+        else:
+            return super().get_token_ids(text)

libs/langchain/langchain/utilities/anthropic.py

@@ -0,0 +1,25 @@
+from typing import Any, List
+
+
+def _get_anthropic_client() -> Any:
+    try:
+        import anthropic
+    except ImportError:
+        raise ImportError(
+            "Could not import anthropic python package. "
+            "This is needed in order to accurately tokenize the text "
+            "for anthropic models. Please install it with `pip install anthropic`."
+        )
+    return anthropic.Anthropic()
+
+
+def get_num_tokens_anthropic(text: str) -> int:
+    client = _get_anthropic_client()
+    return client.count_tokens(text=text)
+
+
+def get_token_ids_anthropic(text: str) -> List[int]:
+    client = _get_anthropic_client()
+    tokenizer = client.get_tokenizer()
+    encoded_text = tokenizer.encode(text)
+    return encoded_text.ids