youngkingdom/dify
1
0
Fork 0
mirror of https://github.com/langgenius/dify.git synced 2026-05-04 17:38:04 +08:00
Code Issues Packages Projects Releases Wiki Activity

merge main

Browse Source
This commit is contained in:
takatost
2024-03-28 14:38:21 +08:00
parent 63a4ddc251 20d16d7b31
commit 858ab8c8c4
32 changed files with 533 additions and 107 deletions
Download Patch File Download Diff File Expand all files Collapse all files

17
api/core/model_runtime/model_providers/__base/model_provider.py
View File

@ -1,4 +1,3 @@
import importlib
import os
from abc import ABC, abstractmethod
@ -7,6 +6,7 @@ import yaml
from core.model_runtime.entities.model_entities import AIModelEntity, ModelType
from core.model_runtime.entities.provider_entities import ProviderEntity
from core.model_runtime.model_providers.__base.ai_model import AIModel
from core.utils.module_import_helper import get_subclasses_from_module, import_module_from_source
class ModelProvider(ABC):
@ -104,17 +104,10 @@ class ModelProvider(ABC):
# Dynamic loading {model_type_name}.py file and find the subclass of AIModel
parent_module = '.'.join(self.__class__.__module__.split('.')[:-1])
spec = importlib.util.spec_from_file_location(f"{parent_module}.{model_type_name}.{model_type_name}", model_type_py_path)
mod = importlib.util.module_from_spec(spec)
spec.loader.exec_module(mod)
model_class = None
for name, obj in vars(mod).items():
if (isinstance(obj, type) and issubclass(obj, AIModel) and not obj.__abstractmethods__
and obj != AIModel and obj.__module__ == mod.__name__):
model_class = obj
break
mod = import_module_from_source(
f'{parent_module}.{model_type_name}.{model_type_name}', model_type_py_path)
model_class = next(filter(lambda x: x.__module__ == mod.__name__ and not x.__abstractmethods__,
get_subclasses_from_module(mod, AIModel)), None)
if not model_class:
raise Exception(f'Missing AIModel Class for model type {model_type} in {model_type_py_path}')

15
api/core/model_runtime/model_providers/model_provider_factory.py
View File

@ -1,4 +1,3 @@
import importlib
import logging
import os
from typing import Optional
@ -10,6 +9,7 @@ from core.model_runtime.entities.provider_entities import ProviderConfig, Provid
from core.model_runtime.model_providers.__base.model_provider import ModelProvider
from core.model_runtime.schema_validators.model_credential_schema_validator import ModelCredentialSchemaValidator
from core.model_runtime.schema_validators.provider_credential_schema_validator import ProviderCredentialSchemaValidator
from core.utils.module_import_helper import load_single_subclass_from_source
from core.utils.position_helper import get_position_map, sort_to_dict_by_position_map
logger = logging.getLogger(__name__)
@ -229,15 +229,10 @@ class ModelProviderFactory:
# Dynamic loading {model_provider_name}.py file and find the subclass of ModelProvider
py_path = os.path.join(model_provider_dir_path, model_provider_name + '.py')
spec = importlib.util.spec_from_file_location(f'core.model_runtime.model_providers.{model_provider_name}.{model_provider_name}', py_path)
mod = importlib.util.module_from_spec(spec)
spec.loader.exec_module(mod)
model_provider_class = None
for name, obj in vars(mod).items():
if isinstance(obj, type) and issubclass(obj, ModelProvider) and obj != ModelProvider:
model_provider_class = obj
break
model_provider_class = load_single_subclass_from_source(
module_name=f'core.model_runtime.model_providers.{model_provider_name}.{model_provider_name}',
script_path=py_path,
parent_type=ModelProvider)
if not model_provider_class:
logger.warning(f"Missing Model Provider Class that extends ModelProvider in {py_path}, Skip.")

37
api/core/model_runtime/model_providers/wenxin/llm/ernie-3.5-4k-0205.yaml Normal file
View File

@ -0,0 +1,37 @@
model: ernie-3.5-8k
label:
en_US: Ernie-3.5-8K
model_type: llm
features:
- agent-thought
model_properties:
mode: chat
context_size: 4096
parameter_rules:
- name: temperature
use_template: temperature
min: 0.1
max: 1.0
default: 0.8
- name: top_p
use_template: top_p
- name: max_tokens
use_template: max_tokens
default: 1024
min: 2
max: 2048
- name: presence_penalty
use_template: presence_penalty
- name: frequency_penalty
use_template: frequency_penalty
- name: response_format
use_template: response_format
- name: disable_search
label:
zh_Hans: 禁用搜索
en_US: Disable Search
type: boolean
help:
zh_Hans: 禁用模型自行进行外部搜索。
en_US: Disable the model to perform external search.
required: false

37
api/core/model_runtime/model_providers/wenxin/llm/ernie-3.5-8k-0205.yaml Normal file
View File

@ -0,0 +1,37 @@
model: ernie-3.5-8k-0205
label:
en_US: Ernie-3.5-8K-0205
model_type: llm
features:
- agent-thought
model_properties:
mode: chat
context_size: 8192
parameter_rules:
- name: temperature
use_template: temperature
min: 0.1
max: 1.0
default: 0.8
- name: top_p
use_template: top_p
- name: max_tokens
use_template: max_tokens
default: 1024
min: 2
max: 2048
- name: presence_penalty
use_template: presence_penalty
- name: frequency_penalty
use_template: frequency_penalty
- name: response_format
use_template: response_format
- name: disable_search
label:
zh_Hans: 禁用搜索
en_US: Disable Search
type: boolean
help:
zh_Hans: 禁用模型自行进行外部搜索。
en_US: Disable the model to perform external search.
required: false

37
api/core/model_runtime/model_providers/wenxin/llm/ernie-3.5-8k-1222.yaml Normal file
View File

@ -0,0 +1,37 @@
model: ernie-3.5-8k-1222
label:
en_US: Ernie-3.5-8K-1222
model_type: llm
features:
- agent-thought
model_properties:
mode: chat
context_size: 8192
parameter_rules:
- name: temperature
use_template: temperature
min: 0.1
max: 1.0
default: 0.8
- name: top_p
use_template: top_p
- name: max_tokens
use_template: max_tokens
default: 1024
min: 2
max: 2048
- name: presence_penalty
use_template: presence_penalty
- name: frequency_penalty
use_template: frequency_penalty
- name: response_format
use_template: response_format
- name: disable_search
label:
zh_Hans: 禁用搜索
en_US: Disable Search
type: boolean
help:
zh_Hans: 禁用模型自行进行外部搜索。
en_US: Disable the model to perform external search.
required: false

37
api/core/model_runtime/model_providers/wenxin/llm/ernie-3.5-8k.yaml Normal file
View File

@ -0,0 +1,37 @@
model: ernie-3.5-8k
label:
en_US: Ernie-3.5-8K
model_type: llm
features:
- agent-thought
model_properties:
mode: chat
context_size: 8192
parameter_rules:
- name: temperature
use_template: temperature
min: 0.1
max: 1.0
default: 0.8
- name: top_p
use_template: top_p
- name: max_tokens
use_template: max_tokens
default: 1024
min: 2
max: 2048
- name: presence_penalty
use_template: presence_penalty
- name: frequency_penalty
use_template: frequency_penalty
- name: response_format
use_template: response_format
- name: disable_search
label:
zh_Hans: 禁用搜索
en_US: Disable Search
type: boolean
help:
zh_Hans: 禁用模型自行进行外部搜索。
en_US: Disable the model to perform external search.
required: false

37
api/core/model_runtime/model_providers/wenxin/llm/ernie-4.0-8k.yaml Normal file
View File

@ -0,0 +1,37 @@
model: ernie-4.0-8k
label:
en_US: Ernie-4.0-8K
model_type: llm
features:
- agent-thought
model_properties:
mode: chat
context_size: 8192
parameter_rules:
- name: temperature
use_template: temperature
min: 0.1
max: 1.0
default: 0.8
- name: top_p
use_template: top_p
- name: max_tokens
use_template: max_tokens
default: 1024
min: 2
max: 2048
- name: presence_penalty
use_template: presence_penalty
- name: frequency_penalty
use_template: frequency_penalty
- name: response_format
use_template: response_format
- name: disable_search
label:
zh_Hans: 禁用搜索
en_US: Disable Search
type: boolean
help:
zh_Hans: 禁用模型自行进行外部搜索。
en_US: Disable the model to perform external search.
required: false

1
api/core/model_runtime/model_providers/wenxin/llm/ernie-bot-4.yaml
View File

@ -36,3 +36,4 @@ parameter_rules:
zh_Hans: 禁用模型自行进行外部搜索。
en_US: Disable the model to perform external search.
required: false
deprecated: true

1
api/core/model_runtime/model_providers/wenxin/llm/ernie-bot-8k.yaml
View File

@ -36,3 +36,4 @@ parameter_rules:
zh_Hans: 禁用模型自行进行外部搜索。
en_US: Disable the model to perform external search.
required: false
deprecated: true

1
api/core/model_runtime/model_providers/wenxin/llm/ernie-bot-turbo.yaml
View File

@ -27,3 +27,4 @@ parameter_rules:
use_template: frequency_penalty
- name: response_format
use_template: response_format
deprecated: true

1
api/core/model_runtime/model_providers/wenxin/llm/ernie-bot.yaml
View File

@ -36,3 +36,4 @@ parameter_rules:
required: false
- name: response_format
use_template: response_format
deprecated: true

30
api/core/model_runtime/model_providers/wenxin/llm/ernie-lite-8k-0308.yaml Normal file
View File

@ -0,0 +1,30 @@
model: ernie-lite-8k-0308
label:
en_US: ERNIE-Lite-8K-0308
model_type: llm
features:
- agent-thought
model_properties:
mode: chat
context_size: 8192
parameter_rules:
- name: temperature
use_template: temperature
min: 0.1
max: 1.0
default: 0.95
- name: top_p
use_template: top_p
min: 0
max: 1.0
default: 0.7
- name: max_tokens
use_template: max_tokens
default: 1024
min: 2
max: 1024
- name: presence_penalty
use_template: presence_penalty
default: 1.0
min: 1.0
max: 2.0

30
api/core/model_runtime/model_providers/wenxin/llm/ernie-lite-8k-0922.yaml Normal file
View File

@ -0,0 +1,30 @@
model: ernie-lite-8k-0922
label:
en_US: ERNIE-Lite-8K-0922
model_type: llm
features:
- agent-thought
model_properties:
mode: chat
context_size: 8192
parameter_rules:
- name: temperature
use_template: temperature
min: 0.1
max: 1.0
default: 0.95
- name: top_p
use_template: top_p
min: 0
max: 1.0
default: 0.7
- name: max_tokens
use_template: max_tokens
default: 1024
min: 2
max: 1024
- name: presence_penalty
use_template: presence_penalty
default: 1.0
min: 1.0
max: 2.0

30
api/core/model_runtime/model_providers/wenxin/llm/ernie-speed-128k.yaml Normal file
View File

@ -0,0 +1,30 @@
model: ernie-speed-128k
label:
en_US: ERNIE-Speed-128K
model_type: llm
features:
- agent-thought
model_properties:
mode: chat
context_size: 128000
parameter_rules:
- name: temperature
use_template: temperature
min: 0.1
max: 1.0
default: 0.95
- name: top_p
use_template: top_p
min: 0
max: 1.0
default: 0.7
- name: max_tokens
use_template: max_tokens
default: 1024
min: 2
max: 1024
- name: presence_penalty
use_template: presence_penalty
default: 1.0
min: 1.0
max: 2.0

30
api/core/model_runtime/model_providers/wenxin/llm/ernie-speed-8k.yaml Normal file
View File

@ -0,0 +1,30 @@
model: ernie-speed-8k
label:
en_US: ERNIE-Speed-8K
model_type: llm
features:
- agent-thought
model_properties:
mode: chat
context_size: 8192
parameter_rules:
- name: temperature
use_template: temperature
min: 0.1
max: 1.0
default: 0.95
- name: top_p
use_template: top_p
min: 0
max: 1.0
default: 0.7
- name: max_tokens
use_template: max_tokens
default: 1024
min: 2
max: 1024
- name: presence_penalty
use_template: presence_penalty
default: 1.0
min: 1.0
max: 2.0

25
api/core/model_runtime/model_providers/wenxin/llm/ernie-speed-appbuilder.yaml Normal file
View File

@ -0,0 +1,25 @@
model: ernie-speed-appbuilder
label:
en_US: ERNIE-Speed-AppBuilder
model_type: llm
features:
- agent-thought
model_properties:
mode: chat
context_size: 8192
parameter_rules:
- name: temperature
use_template: temperature
min: 0.1
max: 1.0
default: 0.95
- name: top_p
use_template: top_p
min: 0
max: 1.0
default: 0.7
- name: presence_penalty
use_template: presence_penalty
default: 1.0
min: 1.0
max: 2.0

24
api/core/model_runtime/model_providers/wenxin/llm/ernie_bot.py
View File

@ -121,15 +121,29 @@ class ErnieMessage:
class ErnieBotModel:
api_bases = {
'ernie-bot': 'https://aip.baidubce.com/rpc/2.0/ai_custom/v1/wenxinworkshop/chat/completions',
'ernie-bot': 'https://aip.baidubce.com/rpc/2.0/ai_custom/v1/wenxinworkshop/chat/ernie-3.5-4k-0205',
'ernie-bot-4': 'https://aip.baidubce.com/rpc/2.0/ai_custom/v1/wenxinworkshop/chat/completions_pro',
'ernie-bot-8k': 'https://aip.baidubce.com/rpc/2.0/ai_custom/v1/wenxinworkshop/chat/ernie_bot_8k',
'ernie-bot-8k': 'https://aip.baidubce.com/rpc/2.0/ai_custom/v1/wenxinworkshop/chat/completions',
'ernie-bot-turbo': 'https://aip.baidubce.com/rpc/2.0/ai_custom/v1/wenxinworkshop/chat/eb-instant',
'ernie-3.5-8k': 'https://aip.baidubce.com/rpc/2.0/ai_custom/v1/wenxinworkshop/chat/completions',
'ernie-3.5-8k-0205': 'https://aip.baidubce.com/rpc/2.0/ai_custom/v1/wenxinworkshop/chat/ernie-3.5-8k-0205',
'ernie-3.5-8k-1222': 'https://aip.baidubce.com/rpc/2.0/ai_custom/v1/wenxinworkshop/chat/ernie-3.5-8k-1222',
'ernie-3.5-4k-0205': 'https://aip.baidubce.com/rpc/2.0/ai_custom/v1/wenxinworkshop/chat/ernie-3.5-4k-0205',
'ernie-4.0-8k': 'https://aip.baidubce.com/rpc/2.0/ai_custom/v1/wenxinworkshop/chat/completions_pro',
'ernie-speed-8k': 'https://aip.baidubce.com/rpc/2.0/ai_custom/v1/wenxinworkshop/chat/ernie_speed',
'ernie-speed-128k': 'https://aip.baidubce.com/rpc/2.0/ai_custom/v1/wenxinworkshop/chat/ernie-speed-128k',
'ernie-speed-appbuilder': 'https://aip.baidubce.com/rpc/2.0/ai_custom/v1/wenxinworkshop/chat/ai_apaas',
'ernie-lite-8k-0922': 'https://aip.baidubce.com/rpc/2.0/ai_custom/v1/wenxinworkshop/chat/eb-instant',
'ernie-lite-8k-0308': 'https://aip.baidubce.com/rpc/2.0/ai_custom/v1/wenxinworkshop/chat/ernie-lite-8k',
}
function_calling_supports = [
'ernie-bot',
'ernie-bot-8k',
'ernie-3.5-8k',
'ernie-3.5-8k-0205',
'ernie-3.5-8k-1222',
'ernie-3.5-4k-0205'
]
api_key: str = ''
@ -285,6 +299,12 @@ class ErnieBotModel:
**parameters
}
if 'max_tokens' in parameters and type(parameters['max_tokens']) == int:
body['max_output_tokens'] = parameters['max_tokens']
if 'presence_penalty' in parameters and type(parameters['presence_penalty']) == float:
body['penalty_score'] = parameters['presence_penalty']
if system_message:
body['system'] = system_message