opensource
/
dify
mirror of https://github.com/langgenius/dify.git
2
0
Fork 0
Code Issues Packages Projects Releases Wiki Activity

feat: support azure openai temporary, must create deployment id same as openai model name, eg. gpt-3.5-turbo / text-embedding-ada-002 / ...

This commit is contained in:
John Wang 2023-05-19 09:34:16 +08:00
parent 3b3c604eb5
commit cbe0f6f3ad
11 changed files with 111 additions and 76 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

5
api/config.py
View File

@ -47,6 +47,7 @@ DEFAULTS = {
'PDF_PREVIEW': 'True',
'LOG_LEVEL': 'INFO',
'DISABLE_PROVIDER_CONFIG_VALIDATION': 'False',
'DEFAULT_LLM_PROVIDER': 'openai'
}
@ -181,6 +182,10 @@ class Config:
# You could disable it for compatibility with certain OpenAPI providers
self.DISABLE_PROVIDER_CONFIG_VALIDATION = get_bool_env('DISABLE_PROVIDER_CONFIG_VALIDATION')
# For temp use only
# set default LLM provider, default is 'openai', support `azure_openai`
self.DEFAULT_LLM_PROVIDER = get_env('DEFAULT_LLM_PROVIDER')
class CloudEditionConfig(Config):
def __init__(self):

72
api/core/embedding/openai_embedding.py
View File

@ -11,9 +11,10 @@ from core.llm.error_handle_wraps import handle_llm_exceptions, handle_llm_except
@retry(reraise=True, wait=wait_random_exponential(min=1, max=20), stop=stop_after_attempt(6))
def get_embedding(
text: str,
engine: Optional[str] = None,
openai_api_key: Optional[str] = None,
text: str,
engine: Optional[str] = None,
api_key: Optional[str] = None,
**kwargs
) -> List[float]:
"""Get embedding.
@ -25,11 +26,12 @@ def get_embedding(
"""
text = text.replace("\n", " ")
return openai.Embedding.create(input=[text], engine=engine, api_key=openai_api_key)["data"][0]["embedding"]
return openai.Embedding.create(input=[text], engine=engine, api_key=api_key, **kwargs)["data"][0]["embedding"]
@retry(reraise=True, wait=wait_random_exponential(min=1, max=20), stop=stop_after_attempt(6))
async def aget_embedding(text: str, engine: Optional[str] = None, openai_api_key: Optional[str] = None) -> List[float]:
async def aget_embedding(text: str, engine: Optional[str] = None, api_key: Optional[str] = None, **kwargs) -> List[
float]:
"""Asynchronously get embedding.
NOTE: Copied from OpenAI's embedding utils:
@ -42,16 +44,17 @@ async def aget_embedding(text: str, engine: Optional[str] = None, openai_api_key
# replace newlines, which can negatively affect performance.
text = text.replace("\n", " ")
return (await openai.Embedding.acreate(input=[text], engine=engine, api_key=openai_api_key))["data"][0][
return (await openai.Embedding.acreate(input=[text], engine=engine, api_key=api_key, **kwargs))["data"][0][
"embedding"
]
@retry(reraise=True, wait=wait_random_exponential(min=1, max=20), stop=stop_after_attempt(6))
def get_embeddings(
list_of_text: List[str],
engine: Optional[str] = None,
openai_api_key: Optional[str] = None
list_of_text: List[str],
engine: Optional[str] = None,
api_key: Optional[str] = None,
**kwargs
) -> List[List[float]]:
"""Get embeddings.
@ -67,14 +70,14 @@ def get_embeddings(
# replace newlines, which can negatively affect performance.
list_of_text = [text.replace("\n", " ") for text in list_of_text]
data = openai.Embedding.create(input=list_of_text, engine=engine, api_key=openai_api_key).data
data = openai.Embedding.create(input=list_of_text, engine=engine, api_key=api_key, **kwargs).data
data = sorted(data, key=lambda x: x["index"]) # maintain the same order as input.
return [d["embedding"] for d in data]
@retry(reraise=True, wait=wait_random_exponential(min=1, max=20), stop=stop_after_attempt(6))
async def aget_embeddings(
list_of_text: List[str], engine: Optional[str] = None, openai_api_key: Optional[str] = None
list_of_text: List[str], engine: Optional[str] = None, api_key: Optional[str] = None, **kwargs
) -> List[List[float]]:
"""Asynchronously get embeddings.
@ -90,7 +93,7 @@ async def aget_embeddings(
# replace newlines, which can negatively affect performance.
list_of_text = [text.replace("\n", " ") for text in list_of_text]
data = (await openai.Embedding.acreate(input=list_of_text, engine=engine, api_key=openai_api_key)).data
data = (await openai.Embedding.acreate(input=list_of_text, engine=engine, api_key=api_key, **kwargs)).data
data = sorted(data, key=lambda x: x["index"]) # maintain the same order as input.
return [d["embedding"] for d in data]
@ -98,19 +101,30 @@ async def aget_embeddings(
class OpenAIEmbedding(BaseEmbedding):
def __init__(
self,
mode: str = OpenAIEmbeddingMode.TEXT_SEARCH_MODE,
model: str = OpenAIEmbeddingModelType.TEXT_EMBED_ADA_002,
deployment_name: Optional[str] = None,
openai_api_key: Optional[str] = None,
**kwargs: Any,
self,
mode: str = OpenAIEmbeddingMode.TEXT_SEARCH_MODE,
model: str = OpenAIEmbeddingModelType.TEXT_EMBED_ADA_002,
deployment_name: Optional[str] = None,
openai_api_key: Optional[str] = None,
**kwargs: Any,
) -> None:
"""Init params."""
super().__init__(**kwargs)
new_kwargs = {}
if 'embed_batch_size' in kwargs:
new_kwargs['embed_batch_size'] = kwargs['embed_batch_size']
if 'tokenizer' in kwargs:
new_kwargs['tokenizer'] = kwargs['tokenizer']
super().__init__(**new_kwargs)
self.mode = OpenAIEmbeddingMode(mode)
self.model = OpenAIEmbeddingModelType(model)
self.deployment_name = deployment_name
self.openai_api_key = openai_api_key
self.openai_api_type = kwargs.get('openai_api_type')
self.openai_api_version = kwargs.get('openai_api_version')
self.openai_api_base = kwargs.get('openai_api_base')
@handle_llm_exceptions
def _get_query_embedding(self, query: str) -> List[float]:
@ -122,7 +136,9 @@ class OpenAIEmbedding(BaseEmbedding):
if key not in _QUERY_MODE_MODEL_DICT:
raise ValueError(f"Invalid mode, model combination: {key}")
engine = _QUERY_MODE_MODEL_DICT[key]
return get_embedding(query, engine=engine, openai_api_key=self.openai_api_key)
return get_embedding(query, engine=engine, api_key=self.openai_api_key,
api_type=self.openai_api_type, api_version=self.openai_api_version,
api_base=self.openai_api_base)
def _get_text_embedding(self, text: str) -> List[float]:
"""Get text embedding."""
@ -133,7 +149,9 @@ class OpenAIEmbedding(BaseEmbedding):
if key not in _TEXT_MODE_MODEL_DICT:
raise ValueError(f"Invalid mode, model combination: {key}")
engine = _TEXT_MODE_MODEL_DICT[key]
return get_embedding(text, engine=engine, openai_api_key=self.openai_api_key)
return get_embedding(text, engine=engine, api_key=self.openai_api_key,
api_type=self.openai_api_type, api_version=self.openai_api_version,
api_base=self.openai_api_base)
async def _aget_text_embedding(self, text: str) -> List[float]:
"""Asynchronously get text embedding."""
@ -144,7 +162,9 @@ class OpenAIEmbedding(BaseEmbedding):
if key not in _TEXT_MODE_MODEL_DICT:
raise ValueError(f"Invalid mode, model combination: {key}")
engine = _TEXT_MODE_MODEL_DICT[key]
return await aget_embedding(text, engine=engine, openai_api_key=self.openai_api_key)
return await aget_embedding(text, engine=engine, api_key=self.openai_api_key,
api_type=self.openai_api_type, api_version=self.openai_api_version,
api_base=self.openai_api_base)
def _get_text_embeddings(self, texts: List[str]) -> List[List[float]]:
"""Get text embeddings.
@ -160,7 +180,9 @@ class OpenAIEmbedding(BaseEmbedding):
if key not in _TEXT_MODE_MODEL_DICT:
raise ValueError(f"Invalid mode, model combination: {key}")
engine = _TEXT_MODE_MODEL_DICT[key]
embeddings = get_embeddings(texts, engine=engine, openai_api_key=self.openai_api_key)
embeddings = get_embeddings(texts, engine=engine, api_key=self.openai_api_key,
api_type=self.openai_api_type, api_version=self.openai_api_version,
api_base=self.openai_api_base)
return embeddings
async def _aget_text_embeddings(self, texts: List[str]) -> List[List[float]]:
@ -172,5 +194,7 @@ class OpenAIEmbedding(BaseEmbedding):
if key not in _TEXT_MODE_MODEL_DICT:
raise ValueError(f"Invalid mode, model combination: {key}")
engine = _TEXT_MODE_MODEL_DICT[key]
embeddings = await aget_embeddings(texts, engine=engine, openai_api_key=self.openai_api_key)
embeddings = await aget_embeddings(texts, engine=engine, api_key=self.openai_api_key,
api_type=self.openai_api_type, api_version=self.openai_api_version,
api_base=self.openai_api_base)
return embeddings

25
api/core/llm/llm_builder.py
View File

@ -1,10 +1,13 @@
from typing import Union, Optional
from flask import current_app
from langchain.callbacks import CallbackManager
from langchain.llms.fake import FakeListLLM
from core.constant import llm_constant
from core.llm.provider.llm_provider_service import LLMProviderService
from core.llm.streamable_azure_chat_open_ai import StreamableAzureChatOpenAI
from core.llm.streamable_azure_open_ai import StreamableAzureOpenAI
from core.llm.streamable_chat_open_ai import StreamableChatOpenAI
from core.llm.streamable_open_ai import StreamableOpenAI
@ -31,12 +34,19 @@ class LLMBuilder:
if model_name == 'fake':
return FakeListLLM(responses=[])
provider = current_app.config.get('DEFAULT_LLM_PROVIDER')
mode = cls.get_mode_by_model(model_name)
if mode == 'chat':
# llm_cls = StreamableAzureChatOpenAI
llm_cls = StreamableChatOpenAI
if provider == 'openai':
llm_cls = StreamableChatOpenAI
else:
llm_cls = StreamableAzureChatOpenAI
elif mode == 'completion':
llm_cls = StreamableOpenAI
if provider == 'openai':
llm_cls = StreamableOpenAI
else:
llm_cls = StreamableAzureOpenAI
else:
raise ValueError(f"model name {model_name} is not supported.")
@ -93,11 +103,12 @@ class LLMBuilder:
"""
if not model_name:
raise Exception('model name not found')
#
# if model_name not in llm_constant.models:
# raise Exception('model {} not found'.format(model_name))
if model_name not in llm_constant.models:
raise Exception('model {} not found'.format(model_name))
model_provider = llm_constant.models[model_name]
# model_provider = llm_constant.models[model_name]
model_provider = current_app.config.get('DEFAULT_LLM_PROVIDER')
provider_service = LLMProviderService(tenant_id=tenant_id, provider_name=model_provider)
return provider_service.get_credentials(model_name)

3
api/core/llm/provider/azure_provider.py
View File

@ -36,8 +36,7 @@ class AzureProvider(BaseProvider):
"""
Returns the API credentials for Azure OpenAI as a dictionary.
"""
encrypted_config = self.get_provider_api_key(model_id=model_id)
config = json.loads(encrypted_config)
config = self.get_provider_api_key(model_id=model_id)
config['openai_api_type'] = 'azure'
config['deployment_name'] = model_id
return config

2
api/core/llm/provider/base.py
View File

@ -14,7 +14,7 @@ class BaseProvider(ABC):
def __init__(self, tenant_id: str):
self.tenant_id = tenant_id
def get_provider_api_key(self, model_id: Optional[str] = None, prefer_custom: bool = True) -> str:
def get_provider_api_key(self, model_id: Optional[str] = None, prefer_custom: bool = True) -> Union[str | dict]:
"""
Returns the decrypted API key for the given tenant_id and provider_name.
If the provider is of type SYSTEM and the quota is exceeded, raises a QuotaExceededError.

20
api/core/llm/streamable_azure_open_ai.py Normal file
View File

@ -0,0 +1,20 @@
from langchain.llms import AzureOpenAI
from langchain.schema import LLMResult
from typing import Optional, List
from core.llm.error_handle_wraps import handle_llm_exceptions, handle_llm_exceptions_async
class StreamableAzureOpenAI(AzureOpenAI):
@handle_llm_exceptions
def generate(
self, prompts: List[str], stop: Optional[List[str]] = None
) -> LLMResult:
return super().generate(prompts, stop)
@handle_llm_exceptions_async
async def agenerate(
self, prompts: List[str], stop: Optional[List[str]] = None
) -> LLMResult:
return await super().agenerate(prompts, stop)

30
web/app/components/header/account-setting/provider-page/azure-provider/index.tsx
View File

@ -20,7 +20,7 @@ const AzureProvider = ({
const [token, setToken] = useState(provider.token as ProviderAzureToken || {})
const handleFocus = () => {
if (token === provider.token) {
token.azure_api_key = ''
token.openai_api_key = ''
setToken({...token})
onTokenChange({...token})
}
@ -35,31 +35,17 @@ const AzureProvider = ({
<div className='px-4 py-3'>
<ProviderInput
className='mb-4'
name={t('common.provider.azure.resourceName')}
placeholder={t('common.provider.azure.resourceNamePlaceholder')}
value={token.azure_api_base}
onChange={(v) => handleChange('azure_api_base', v)}
/>
<ProviderInput
className='mb-4'
name={t('common.provider.azure.deploymentId')}
placeholder={t('common.provider.azure.deploymentIdPlaceholder')}
value={token.azure_api_type}
onChange={v => handleChange('azure_api_type', v)}
/>
<ProviderInput
className='mb-4'
name={t('common.provider.azure.apiVersion')}
placeholder={t('common.provider.azure.apiVersionPlaceholder')}
value={token.azure_api_version}
onChange={v => handleChange('azure_api_version', v)}
name={t('common.provider.azure.apiBase')}
placeholder={t('common.provider.azure.apiBasePlaceholder')}
value={token.openai_api_base}
onChange={(v) => handleChange('openai_api_base', v)}
/>
<ProviderValidateTokenInput
className='mb-4'
name={t('common.provider.azure.apiKey')}
placeholder={t('common.provider.azure.apiKeyPlaceholder')}
value={token.azure_api_key}
onChange={v => handleChange('azure_api_key', v)}
value={token.openai_api_key}
onChange={v => handleChange('openai_api_key', v)}
onFocus={handleFocus}
onValidatedStatus={onValidatedStatus}
providerName={provider.provider_name}
@ -72,4 +58,4 @@ const AzureProvider = ({
)
}
export default AzureProvider
export default AzureProvider

6
web/app/components/header/account-setting/provider-page/provider-item/index.tsx
View File

@ -33,12 +33,12 @@ const ProviderItem = ({
const { notify } = useContext(ToastContext)
const [token, setToken] = useState<ProviderAzureToken | string>(
provider.provider_name === 'azure_openai'
? { azure_api_base: '', azure_api_type: '', azure_api_version: '', azure_api_key: '' }
? { openai_api_base: '', openai_api_key: '' }
: ''
)
const id = `${provider.provider_name}-${provider.provider_type}`
const isOpen = id === activeId
const providerKey = provider.provider_name === 'azure_openai' ? (provider.token as ProviderAzureToken)?.azure_api_key : provider.token
const providerKey = provider.provider_name === 'azure_openai' ? (provider.token as ProviderAzureToken)?.openai_api_key : provider.token
const comingSoon = false
const isValid = provider.is_valid
@ -135,4 +135,4 @@ const ProviderItem = ({
)
}
export default ProviderItem
export default ProviderItem

8
web/i18n/lang/common.en.ts
View File

@ -148,12 +148,8 @@ const translation = {
editKey: 'Edit',
invalidApiKey: 'Invalid API key',
azure: {
resourceName: 'Resource Name',
resourceNamePlaceholder: 'The name of your Azure OpenAI Resource.',
deploymentId: 'Deployment ID',
deploymentIdPlaceholder: 'The deployment name you chose when you deployed the model.',
apiVersion: 'API Version',
apiVersionPlaceholder: 'The API version to use for this operation.',
apiBase: 'API Base',
apiBasePlaceholder: 'The API Base URL of your Azure OpenAI Resource.',
apiKey: 'API Key',
apiKeyPlaceholder: 'Enter your API key here',
helpTip: 'Learn Azure OpenAI Service',

10
web/i18n/lang/common.zh.ts
View File

@ -149,14 +149,10 @@ const translation = {
editKey: '编辑',
invalidApiKey: '无效的 API 密钥',
azure: {
resourceName: 'Resource Name',
resourceNamePlaceholder: 'The name of your Azure OpenAI Resource.',
deploymentId: 'Deployment ID',
deploymentIdPlaceholder: 'The deployment name you chose when you deployed the model.',
apiVersion: 'API Version',
apiVersionPlaceholder: 'The API version to use for this operation.',
apiBase: 'API Base',
apiBasePlaceholder: '输入您的 Azure OpenAI API Base 地址',
apiKey: 'API Key',
apiKeyPlaceholder: 'Enter your API key here',
apiKeyPlaceholder: '输入你的 API 密钥',
helpTip: '了解 Azure OpenAI Service',
},
openaiHosted: {

6
web/models/common.ts
View File

@ -55,10 +55,8 @@ export type Member = Pick<UserProfileResponse, 'id' | 'name' | 'email' | 'last_l
}
export type ProviderAzureToken = {
azure_api_base: string
azure_api_key: string
azure_api_type: string
azure_api_version: string
openai_api_base: string
openai_api_key: string
}
export type Provider = {
provider_name: string