feat: add Anthropic claude-3 models support (#2684)

api/core/model_runtime/model_providers/anthropic/anthropic.py

@@ -21,7 +21,7 @@ class AnthropicProvider(ModelProvider):
 
             # Use `claude-instant-1` model for validate,
             model_instance.validate_credentials(
-                model='claude-instant-1',
+                model='claude-instant-1.2',
                 credentials=credentials
             )
         except CredentialsValidateFailedError as ex:

api/core/model_runtime/model_providers/anthropic/anthropic.yaml

@@ -2,8 +2,8 @@ provider: anthropic
 label:
   en_US: Anthropic
 description:
-  en_US: Anthropic’s powerful models, such as Claude 2 and Claude Instant.
-  zh_Hans: Anthropic 的强大模型，例如 Claude 2 和 Claude Instant。
+  en_US: Anthropic’s powerful models, such as Claude 3.
+  zh_Hans: Anthropic 的强大模型，例如 Claude 3。
 icon_small:
   en_US: icon_s_en.svg
 icon_large:

api/core/model_runtime/model_providers/anthropic/llm/_position.yaml

+- claude-3-opus-20240229
+- claude-3-sonnet-20240229
+- claude-2.1
+- claude-instant-1.2
+- claude-2
+- claude-instant-1

api/core/model_runtime/model_providers/anthropic/llm/claude-2.yaml

@@ -34,3 +34,4 @@ pricing:
   output: '24.00'
   unit: '0.000001'
   currency: USD
+deprecated: true

api/core/model_runtime/model_providers/anthropic/llm/claude-3-opus-20240229.yaml

+model: claude-3-opus-20240229
+label:
+  en_US: claude-3-opus-20240229
+model_type: llm
+features:
+  - agent-thought
+  - vision
+model_properties:
+  mode: chat
+  context_size: 200000
+parameter_rules:
+  - name: temperature
+    use_template: temperature
+  - name: top_p
+    use_template: top_p
+  - name: top_k
+    label:
+      zh_Hans: 取样数量
+      en_US: Top k
+    type: int
+    help:
+      zh_Hans: 仅从每个后续标记的前 K 个选项中采样。
+      en_US: Only sample from the top K options for each subsequent token.
+    required: false
+  - name: max_tokens
+    use_template: max_tokens
+    required: true
+    default: 4096
+    min: 1
+    max: 4096
+  - name: response_format
+    use_template: response_format
+pricing:
+  input: '15.00'
+  output: '75.00'
+  unit: '0.000001'
+  currency: USD

api/core/model_runtime/model_providers/anthropic/llm/claude-3-sonnet-20240229.yaml

+model: claude-3-sonnet-20240229
+label:
+  en_US: claude-3-sonnet-20240229
+model_type: llm
+features:
+  - agent-thought
+  - vision
+model_properties:
+  mode: chat
+  context_size: 200000
+parameter_rules:
+  - name: temperature
+    use_template: temperature
+  - name: top_p
+    use_template: top_p
+  - name: top_k
+    label:
+      zh_Hans: 取样数量
+      en_US: Top k
+    type: int
+    help:
+      zh_Hans: 仅从每个后续标记的前 K 个选项中采样。
+      en_US: Only sample from the top K options for each subsequent token.
+    required: false
+  - name: max_tokens
+    use_template: max_tokens
+    required: true
+    default: 4096
+    min: 1
+    max: 4096
+  - name: response_format
+    use_template: response_format
+pricing:
+  input: '3.00'
+  output: '15.00'
+  unit: '0.000001'
+  currency: USD

api/core/model_runtime/model_providers/anthropic/llm/claude-instant-1.2.yaml

+model: claude-instant-1.2
+label:
+  en_US: claude-instant-1.2
+model_type: llm
+features: [ ]
+model_properties:
+  mode: chat
+  context_size: 100000
+parameter_rules:
+  - name: temperature
+    use_template: temperature
+  - name: top_p
+    use_template: top_p
+  - name: top_k
+    label:
+      zh_Hans: 取样数量
+      en_US: Top k
+    type: int
+    help:
+      zh_Hans: 仅从每个后续标记的前 K 个选项中采样。
+      en_US: Only sample from the top K options for each subsequent token.
+    required: false
+  - name: max_tokens
+    use_template: max_tokens
+    required: true
+    default: 4096
+    min: 1
+    max: 4096
+  - name: response_format
+    use_template: response_format
+pricing:
+  input: '1.63'
+  output: '5.51'
+  unit: '0.000001'
+  currency: USD

api/core/model_runtime/model_providers/anthropic/llm/claude-instant-1.yaml

@@ -33,3 +33,4 @@ pricing:
   output: '5.51'
   unit: '0.000001'
   currency: USD
+deprecated: true

api/requirements.txt

@@ -35,7 +35,7 @@ docx2txt==0.8
 pypdfium2==4.16.0
 resend~=0.7.0
 pyjwt~=2.8.0
-anthropic~=0.7.7
+anthropic~=0.17.0
 newspaper3k==0.2.8
 google-api-python-client==2.90.0
 wikipedia==1.4.0

api/tests/integration_tests/model_runtime/__mock/anthropic.py

 import os
 from time import sleep
-from typing import Any, Generator, List, Literal, Union
+from typing import Any, Literal, Union, Iterable
+
+from anthropic.resources import Messages
+from anthropic.types.message_delta_event import Delta
 
 import anthropic
 import pytest
 from _pytest.monkeypatch import MonkeyPatch
-from anthropic import Anthropic
-from anthropic._types import NOT_GIVEN, Body, Headers, NotGiven, Query
-from anthropic.resources.completions import Completions
-from anthropic.types import Completion, completion_create_params
+from anthropic import Anthropic, Stream
+from anthropic.types import MessageParam, Message, MessageStreamEvent, \
+    ContentBlock, MessageStartEvent, Usage, TextDelta, MessageDeltaEvent, MessageStopEvent, ContentBlockDeltaEvent, \
+    MessageDeltaUsage
 
 MOCK = os.getenv('MOCK_SWITCH', 'false') == 'true'
 
+
 class MockAnthropicClass(object):
     @staticmethod
-    def mocked_anthropic_chat_create_sync(model: str) -> Completion:
-        return Completion(
-            completion='hello, I\'m a chatbot from anthropic',
+    def mocked_anthropic_chat_create_sync(model: str) -> Message:
+        return Message(
+            id='msg-123',
+            type='message',
+            role='assistant',
+            content=[ContentBlock(text='hello, I\'m a chatbot from anthropic', type='text')],
             model=model,
-            stop_reason='stop_sequence'
+            stop_reason='stop_sequence',
+            usage=Usage(
+                input_tokens=1,
+                output_tokens=1
+            )
         )
 
     @staticmethod
-    def mocked_anthropic_chat_create_stream(model: str) -> Generator[Completion, None, None]:
+    def mocked_anthropic_chat_create_stream(model: str) -> Stream[MessageStreamEvent]:
         full_response_text = "hello, I'm a chatbot from anthropic"
 
-        for i in range(0, len(full_response_text) + 1):
-            sleep(0.1)
-            if i == len(full_response_text):
-                yield Completion(
-                    completion='',
-                    model=model,
-                    stop_reason='stop_sequence'
-                )
-            else:
-                yield Completion(
-                    completion=full_response_text[i],
-                    model=model,
-                    stop_reason=''
+        yield MessageStartEvent(
+            type='message_start',
+            message=Message(
+                id='msg-123',
+                content=[],
+                role='assistant',
+                model=model,
+                stop_reason=None,
+                type='message',
+                usage=Usage(
+                    input_tokens=1,
+                    output_tokens=1
                 )
+            )
+        )
+
+        index = 0
+        for i in range(0, len(full_response_text)):
+            sleep(0.1)
+            yield ContentBlockDeltaEvent(
+                type='content_block_delta',
+                delta=TextDelta(text=full_response_text[i], type='text_delta'),
+                index=index
+            )
+
+            index += 1
+
+        yield MessageDeltaEvent(
+            type='message_delta',
+            delta=Delta(
+                stop_reason='stop_sequence'
+            ),
+            usage=MessageDeltaUsage(
+                output_tokens=1
+            )
+        )
+
+        yield MessageStopEvent(type='message_stop')
 
-    def mocked_anthropic(self: Completions, *,
-        max_tokens_to_sample: int,
-        model: Union[str, Literal["claude-2.1", "claude-instant-1"]],
-        prompt: str,
-        stream: Literal[True],
-        **kwargs: Any
-    ) -> Union[Completion, Generator[Completion, None, None]]:
+    def mocked_anthropic(self: Messages, *,
+                         max_tokens: int,
+                         messages: Iterable[MessageParam],
+                         model: str,
+                         stream: Literal[True],
+                         **kwargs: Any
+                         ) -> Union[Message, Stream[MessageStreamEvent]]:
         if len(self._client.api_key) < 18:
             raise anthropic.AuthenticationError('Invalid API key')
 
@@ -55,12 +90,13 @@ class MockAnthropicClass(object):
         else:
             return MockAnthropicClass.mocked_anthropic_chat_create_sync(model=model)
 
+
 @pytest.fixture
 def setup_anthropic_mock(request, monkeypatch: MonkeyPatch):
     if MOCK:
-        monkeypatch.setattr(Completions, 'create', MockAnthropicClass.mocked_anthropic)
+        monkeypatch.setattr(Messages, 'create', MockAnthropicClass.mocked_anthropic)
 
     yield
 
     if MOCK:
-        monkeypatch.undo()
\ No newline at end of file
+        monkeypatch.undo()

api/tests/integration_tests/model_runtime/anthropic/test_llm.py

@@ -15,14 +15,14 @@ def test_validate_credentials(setup_anthropic_mock):
 
     with pytest.raises(CredentialsValidateFailedError):
         model.validate_credentials(
-            model='claude-instant-1',
+            model='claude-instant-1.2',
             credentials={
                 'anthropic_api_key': 'invalid_key'
             }
         )
 
     model.validate_credentials(
-        model='claude-instant-1',
+        model='claude-instant-1.2',
         credentials={
             'anthropic_api_key': os.environ.get('ANTHROPIC_API_KEY')
         }
@@ -33,7 +33,7 @@ def test_invoke_model(setup_anthropic_mock):
     model = AnthropicLargeLanguageModel()
 
     response = model.invoke(
-        model='claude-instant-1',
+        model='claude-instant-1.2',
         credentials={
             'anthropic_api_key': os.environ.get('ANTHROPIC_API_KEY'),
             'anthropic_api_url': os.environ.get('ANTHROPIC_API_URL')
@@ -49,7 +49,7 @@ def test_invoke_model(setup_anthropic_mock):
         model_parameters={
             'temperature': 0.0,
             'top_p': 1.0,
-            'max_tokens_to_sample': 10
+            'max_tokens': 10
         },
         stop=['How'],
         stream=False,
@@ -64,7 +64,7 @@ def test_invoke_stream_model(setup_anthropic_mock):
     model = AnthropicLargeLanguageModel()
 
     response = model.invoke(
-        model='claude-instant-1',
+        model='claude-instant-1.2',
         credentials={
             'anthropic_api_key': os.environ.get('ANTHROPIC_API_KEY')
         },
@@ -78,7 +78,7 @@ def test_invoke_stream_model(setup_anthropic_mock):
         ],
         model_parameters={
             'temperature': 0.0,
-            'max_tokens_to_sample': 100
+            'max_tokens': 100
         },
         stream=True,
         user="abc-123"
@@ -97,7 +97,7 @@ def test_get_num_tokens():
     model = AnthropicLargeLanguageModel()
 
     num_tokens = model.get_num_tokens(
-        model='claude-instant-1',
+        model='claude-instant-1.2',
         credentials={
             'anthropic_api_key': os.environ.get('ANTHROPIC_API_KEY')
         },