update anthropic version (uripeled2#25)

* update anthropic version

* fix import in test_anthropic_client.py

* fix import in test_anthropic_client.py

* Update antrofic version

llm_client/__init__.py

@@ -1,4 +1,4 @@
-__version__ = "0.6.1"
+__version__ = "0.6.2"
 
 from llm_client.base_llm_client import BaseLLMClient
 

llm_client/llm_api_client/anthropic_client.py

@@ -1,6 +1,6 @@
 from typing import Optional
 
-from anthropic import count_tokens
+from anthropic import AsyncAnthropic
 
 from llm_client.llm_api_client.base_llm_api_client import BaseLLMAPIClient, LLMAPIClientConfig
 from llm_client.consts import PROMPT_KEY
@@ -10,6 +10,7 @@
 COMPLETIONS_KEY = "completion"
 AUTH_HEADER = "x-api-key"
 ACCEPT_HEADER = "Accept"
+VERSION_HEADER = "anthropic-version"
 ACCEPT_VALUE = "application/json"
 MAX_TOKENS_KEY = "max_tokens_to_sample"
 
@@ -19,6 +20,9 @@ def __init__(self, config: LLMAPIClientConfig):
         super().__init__(config)
         if self._base_url is None:
             self._base_url = BASE_URL
+        self._anthropic = AsyncAnthropic()
+        if self._headers.get(VERSION_HEADER) is None:
+            self._headers[VERSION_HEADER] = self._anthropic.default_headers[VERSION_HEADER]
         self._headers[ACCEPT_HEADER] = ACCEPT_VALUE
         self._headers[AUTH_HEADER] = self._api_key
 
@@ -40,4 +44,4 @@ async def text_completion(self, prompt: str, model: Optional[str] = None, max_to
         return [response_json[COMPLETIONS_KEY]]
 
     async def get_tokens_count(self, text: str, **kwargs) -> int:
-        return count_tokens(text)
+        return await self._anthropic.count_tokens(text)

pyproject.toml

@@ -43,7 +43,7 @@ huggingface = [
     "transformers >= 4.0.0"
 ]
 anthropic = [
-    "anthropic >= 0.2.0"
+    "anthropic >= 0.3.2"
 ]
 google = [
     "google-generativeai >= 0.1.0"

tests/llm_api_client/anthropic_client/conftest.py

@@ -1,7 +1,8 @@
+from unittest.mock import patch, AsyncMock
+
 import pytest
 
-from llm_client import AnthropicClient
-from llm_client.llm_api_client.anthropic_client import BASE_URL, COMPLETE_PATH
+from llm_client.llm_api_client.anthropic_client import BASE_URL, COMPLETE_PATH, VERSION_HEADER, AnthropicClient
 from llm_client.llm_api_client.base_llm_api_client import LLMAPIClientConfig
 
 
@@ -23,3 +24,21 @@ def llm_client(config):
 @pytest.fixture
 def complete_url():
     return BASE_URL + COMPLETE_PATH
+
+
+@pytest.fixture
+def number_of_tokens():
+    return 10
+
+
+@pytest.fixture
+def anthropic_version():
+    return "2023-06-01"
+
+
+@pytest.fixture(autouse=True)
+def mock_anthropic(number_of_tokens, anthropic_version):
+    with patch("llm_client.llm_api_client.anthropic_client.AsyncAnthropic") as mock_anthropic:
+        mock_anthropic.return_value.count_tokens = AsyncMock(return_value=number_of_tokens)
+        mock_anthropic.return_value.default_headers = {VERSION_HEADER: anthropic_version}
+        yield mock_anthropic

tests/llm_api_client/anthropic_client/test_anthropic_client.py

@@ -1,11 +1,9 @@
-from unittest.mock import patch
-
 import pytest
 
-from llm_client import LLMAPIClientFactory, LLMAPIClientType, AnthropicClient
+from llm_client import LLMAPIClientFactory, LLMAPIClientType
 from llm_client.consts import PROMPT_KEY, MODEL_KEY
 from llm_client.llm_api_client.anthropic_client import AUTH_HEADER, COMPLETIONS_KEY, MAX_TOKENS_KEY, ACCEPT_HEADER, \
-    ACCEPT_VALUE
+    ACCEPT_VALUE, VERSION_HEADER, AnthropicClient
 
 
 @pytest.mark.asyncio
@@ -18,18 +16,41 @@ async def test_get_llm_api_client__with_anthropic(config):
 
 
 @pytest.mark.asyncio
-async def test_text_completion__sanity(mock_aioresponse, llm_client, complete_url):
+async def test_text_completion__sanity(mock_aioresponse, llm_client, complete_url, anthropic_version):
+    mock_aioresponse.post(
+        complete_url,
+        payload={COMPLETIONS_KEY: "completion text"}
+    )
+
+    actual = await llm_client.text_completion(prompt="These are a few of my favorite", max_tokens=10)
+
+    assert actual == ["completion text"]
+    mock_aioresponse.assert_called_once_with(complete_url, method='POST',
+                                             headers={AUTH_HEADER: llm_client._api_key,
+                                                      ACCEPT_HEADER: ACCEPT_VALUE,
+                                                      VERSION_HEADER: anthropic_version},
+                                             json={PROMPT_KEY: 'These are a few of my favorite',
+                                                   MAX_TOKENS_KEY: 10, "temperature": 1,
+                                                   MODEL_KEY: llm_client._default_model},
+                                             raise_for_status=True)
+
+
+@pytest.mark.asyncio
+async def test_text_completion__with_version_header(mock_aioresponse, config, complete_url):
     mock_aioresponse.post(
         complete_url,
         payload={COMPLETIONS_KEY: "completion text"}
     )
+    config.headers[VERSION_HEADER] = "1.0.0"
+    llm_client = AnthropicClient(config)
 
     actual = await llm_client.text_completion(prompt="These are a few of my favorite", max_tokens=10)
 
     assert actual == ["completion text"]
     mock_aioresponse.assert_called_once_with(complete_url, method='POST',
                                              headers={AUTH_HEADER: llm_client._api_key,
-                                                      ACCEPT_HEADER: ACCEPT_VALUE},
+                                                      ACCEPT_HEADER: ACCEPT_VALUE,
+                                                      VERSION_HEADER: "1.0.0"},
                                              json={PROMPT_KEY: 'These are a few of my favorite',
                                                    MAX_TOKENS_KEY: 10, "temperature": 1,
                                                    MODEL_KEY: llm_client._default_model},
@@ -43,7 +64,7 @@ async def test_text_completion__without_max_tokens_raise_value_error(mock_aiores
 
 
 @pytest.mark.asyncio
-async def test_text_completion__override_model(mock_aioresponse, llm_client, complete_url):
+async def test_text_completion__override_model(mock_aioresponse, llm_client, complete_url, anthropic_version):
     new_model_name = "claude-instant"
     mock_aioresponse.post(
         complete_url,
@@ -56,15 +77,16 @@ async def test_text_completion__override_model(mock_aioresponse, llm_client, com
     assert actual == ["completion text"]
     mock_aioresponse.assert_called_once_with(complete_url, method='POST',
                                              headers={AUTH_HEADER: llm_client._api_key,
-                                                      ACCEPT_HEADER: ACCEPT_VALUE},
+                                                      ACCEPT_HEADER: ACCEPT_VALUE,
+                                                      VERSION_HEADER: anthropic_version},
                                              json={PROMPT_KEY: 'These are a few of my favorite',
                                                    MAX_TOKENS_KEY: 10, "temperature": 1,
                                                    MODEL_KEY: new_model_name},
                                              raise_for_status=True)
 
 
 @pytest.mark.asyncio
-async def test_text_completion__with_kwargs(mock_aioresponse, llm_client, complete_url):
+async def test_text_completion__with_kwargs(mock_aioresponse, llm_client, complete_url, anthropic_version):
     mock_aioresponse.post(
         complete_url,
         payload={COMPLETIONS_KEY: "completion text"}
@@ -75,7 +97,8 @@ async def test_text_completion__with_kwargs(mock_aioresponse, llm_client, comple
     assert actual == ["completion text"]
     mock_aioresponse.assert_called_once_with(complete_url, method='POST',
                                              headers={AUTH_HEADER: llm_client._api_key,
-                                                      ACCEPT_HEADER: ACCEPT_VALUE},
+                                                      ACCEPT_HEADER: ACCEPT_VALUE,
+                                                      VERSION_HEADER: anthropic_version},
                                              json={PROMPT_KEY: 'These are a few of my favorite',
                                                    MAX_TOKENS_KEY: 10,
                                                    MODEL_KEY: llm_client._default_model,
@@ -84,9 +107,8 @@ async def test_text_completion__with_kwargs(mock_aioresponse, llm_client, comple
 
 
 @pytest.mark.asyncio
-async def test_get_tokens_count__sanity(llm_client):
-    with patch("llm_client.llm_api_client.anthropic_client.count_tokens") as mock_count_tokens:
-        actual = await llm_client.get_tokens_count(text="These are a few of my favorite things!")
+async def test_get_tokens_count__sanity(llm_client, number_of_tokens, mock_anthropic):
+    actual = await llm_client.get_tokens_count(text="These are a few of my favorite things!")
 
-        assert actual == mock_count_tokens.return_value
-        mock_count_tokens.assert_called_once_with("These are a few of my favorite things!")
+    assert actual == 10
+    mock_anthropic.return_value.count_tokens.assert_awaited_once_with("These are a few of my favorite things!")