Model Runtime (#1858)

Co-authored-by: StyleZhang <jasonapring2015@outlook.com> Co-authored-by: Garfield Dai <dai.hai@foxmail.com> Co-authored-by: chenhe <guchenhe@gmail.com> Co-authored-by: jyong <jyong@dify.ai> Co-authored-by: Joel <iamjoel007@gmail.com> Co-authored-by: Yeuoly <admin@srmxy.cn>
2025-12-16 14:26:52 +08:00 · 2024-01-02 23:42:00 +08:00
parent e91dd28a76
commit d069c668f8
807 changed files with 171310 additions and 23806 deletions
--- a/api/tests/integration_tests/model_runtime/openai_api_compatible/init.py
+++ b/api/tests/integration_tests/model_runtime/openai_api_compatible/init.py
--- a/api/tests/integration_tests/model_runtime/openai_api_compatible/test_llm.py
+++ b/api/tests/integration_tests/model_runtime/openai_api_compatible/test_llm.py
@@ -0,0 +1,181 @@
+import os
+from typing import Generator
+
+import pytest
+
+from core.model_runtime.entities.message_entities import AssistantPromptMessage, UserPromptMessage, \
+    SystemPromptMessage, PromptMessageTool
+from core.model_runtime.entities.llm_entities import LLMResult, LLMResultChunkDelta, \
+    LLMResultChunk
+from core.model_runtime.errors.validate import CredentialsValidateFailedError
+from core.model_runtime.model_providers.openai_api_compatible.llm.llm import OAIAPICompatLargeLanguageModel
+
+"""
+Using Together.ai's OpenAI-compatible API as testing endpoint
+"""
+
+def test_validate_credentials():
+    model = OAIAPICompatLargeLanguageModel()
+
+    with pytest.raises(CredentialsValidateFailedError):
+        model.validate_credentials(
+            model='mistralai/Mixtral-8x7B-Instruct-v0.1',
+            credentials={
+                'api_key': 'invalid_key',
+                'endpoint_url': 'https://api.together.xyz/v1/chat/completions',
+                'mode': 'chat'
+            }
+        )
+
+    model.validate_credentials(
+        model='mistralai/Mixtral-8x7B-Instruct-v0.1',
+        credentials={
+            'api_key': os.environ.get('TOGETHER_API_KEY'),
+            'endpoint_url': 'https://api.together.xyz/v1/chat/completions',
+            'mode': 'chat'
+        }
+    )
+
+def test_invoke_model():
+    model = OAIAPICompatLargeLanguageModel()
+
+    response = model.invoke(
+        model='mistralai/Mixtral-8x7B-Instruct-v0.1',
+        credentials={
+            'api_key': os.environ.get('TOGETHER_API_KEY'),
+            'endpoint_url': 'https://api.together.xyz/v1/completions',
+            'mode': 'completion'
+        },
+        prompt_messages=[
+            SystemPromptMessage(
+                content='You are a helpful AI assistant.',
+            ),
+            UserPromptMessage(
+                content='Who are you?'
+            )
+        ],
+        model_parameters={
+            'temperature': 1.0,
+            'top_k': 2,
+            'top_p': 0.5,
+        },
+        stop=['How'],
+        stream=False,
+        user="abc-123"
+    )
+
+    assert isinstance(response, LLMResult)
+    assert len(response.message.content) > 0
+
+def test_invoke_stream_model():
+    model = OAIAPICompatLargeLanguageModel()
+
+    response = model.invoke(
+        model='mistralai/Mixtral-8x7B-Instruct-v0.1',
+        credentials={
+            'api_key': os.environ.get('TOGETHER_API_KEY'),
+            'endpoint_url': 'https://api.together.xyz/v1/chat/completions',
+            'mode': 'chat'
+        },
+        prompt_messages=[
+            SystemPromptMessage(
+                content='You are a helpful AI assistant.',
+            ),
+            UserPromptMessage(
+                content='Who are you?'
+            )
+        ],
+        model_parameters={
+            'temperature': 1.0,
+            'top_k': 2,
+            'top_p': 0.5,
+        },
+        stop=['How'],
+        stream=True,
+        user="abc-123"
+    )
+
+    assert isinstance(response, Generator)
+
+    for chunk in response:
+        assert isinstance(chunk, LLMResultChunk)
+        assert isinstance(chunk.delta, LLMResultChunkDelta)
+        assert isinstance(chunk.delta.message, AssistantPromptMessage)
+
+# using OpenAI's ChatGPT-3.5 as testing endpoint
+def test_invoke_chat_model_with_tools():
+    model = OAIAPICompatLargeLanguageModel()
+
+    result = model.invoke(
+        model='gpt-3.5-turbo',
+        credentials={
+            'api_key': os.environ.get('OPENAI_API_KEY'),
+            'endpoint_url': 'https://api.openai.com/v1/chat/completions',
+            'mode': 'chat'
+        },
+        prompt_messages=[
+            SystemPromptMessage(
+                content='You are a helpful AI assistant.',
+            ),
+            UserPromptMessage(
+                content="what's the weather today in London?",
+            )
+        ],
+        tools=[
+            PromptMessageTool(
+                name='get_weather',
+                description='Determine weather in my location',
+                parameters={
+                    "type": "object",
+                    "properties": {
+                      "location": {
+                        "type": "string",
+                        "description": "The city and state e.g. San Francisco, CA"
+                      },
+                      "unit": {
+                        "type": "string",
+                        "enum": [
+                          "celsius",
+                          "fahrenheit"
+                        ]
+                      }
+                    },
+                    "required": [
+                      "location"
+                    ]
+                  }
+            ),
+        ],
+        model_parameters={
+            'temperature': 0.0,
+            'max_tokens': 1024
+        },
+        stream=False,
+        user="abc-123"
+    )
+
+    assert isinstance(result, LLMResult)
+    assert isinstance(result.message, AssistantPromptMessage)
+    assert len(result.message.tool_calls) > 0
+
+def test_get_num_tokens():
+    model = OAIAPICompatLargeLanguageModel()
+
+    num_tokens = model.get_num_tokens(
+        model='mistralai/Mixtral-8x7B-Instruct-v0.1',
+        credentials={
+            'api_key': os.environ.get('OPENAI_API_KEY'),
+            'endpoint_url': 'https://api.openai.com/v1/chat/completions'
+        },
+        prompt_messages=[
+            SystemPromptMessage(
+                content='You are a helpful AI assistant.',
+            ),
+            UserPromptMessage(
+                content='Hello World!'
+            )
+        ]
+    )
+
+    assert isinstance(num_tokens, int)
+    assert num_tokens == 21
--- a/api/tests/integration_tests/model_runtime/openai_api_compatible/test_text_embedding.py
+++ b/api/tests/integration_tests/model_runtime/openai_api_compatible/test_text_embedding.py
@@ -0,0 +1,79 @@
+import os
+
+import pytest
+
+from core.model_runtime.entities.text_embedding_entities import TextEmbeddingResult
+from core.model_runtime.errors.validate import CredentialsValidateFailedError
+from core.model_runtime.model_providers.openai_api_compatible.text_embedding.text_embedding import OAICompatEmbeddingModel
+
+"""
+Using OpenAI's API as testing endpoint
+"""
+
+def test_validate_credentials():
+    model = OAICompatEmbeddingModel()
+
+    with pytest.raises(CredentialsValidateFailedError):
+        model.validate_credentials(
+            model='text-embedding-ada-002',
+            credentials={
+                'api_key': 'invalid_key',
+                'endpoint_url': 'https://api.openai.com/v1/embeddings',
+                'context_size': 8184,
+                'max_chunks': 32
+                
+            }
+        )
+
+    model.validate_credentials(
+        model='text-embedding-ada-002',
+        credentials={
+            'api_key': os.environ.get('OPENAI_API_KEY'),
+            'endpoint_url': 'https://api.openai.com/v1/embeddings',
+            'context_size': 8184,
+            'max_chunks': 32
+        }
+    )
+
+
+def test_invoke_model():
+    model = OAICompatEmbeddingModel()
+
+    result = model.invoke(
+        model='text-embedding-ada-002',
+        credentials={
+            'api_key': os.environ.get('OPENAI_API_KEY'),
+            'endpoint_url': 'https://api.openai.com/v1/embeddings',
+            'context_size': 8184,
+            'max_chunks': 32
+        },
+        texts=[
+            "hello",
+            "world"
+        ],
+        user="abc-123"
+    )
+
+    assert isinstance(result, TextEmbeddingResult)
+    assert len(result.embeddings) == 2
+    assert result.usage.total_tokens == 2
+
+
+def test_get_num_tokens():
+    model = OAICompatEmbeddingModel()
+
+    num_tokens = model.get_num_tokens(
+        model='text-embedding-ada-002',
+        credentials={
+            'api_key': os.environ.get('OPENAI_API_KEY'),
+            'endpoint_url': 'https://api.openai.com/v1/embeddings',
+            'context_size': 8184,
+            'max_chunks': 32
+        },
+        texts=[
+            "hello",
+            "world"
+        ]
+    )
+
+    assert num_tokens == 2