mirror of
https://github.com/langgenius/dify-plugin-sdks.git
synced 2026-07-22 10:25:23 -04:00
immortal-wm
2fc9753673
* fix latency calc * feat: add timing context management to AIModel for latency tracking - Introduced a new TimingContextRaceConditionError to handle race conditions in multi-threaded environments. - Implemented a timing_context method in AIModel to track request timing and prevent concurrent usage. - Updated various model classes (LargeLanguageModel, ModerationModel, RerankModel, Speech2TextModel, TextEmbeddingModel, TTSModel) to utilize the new timing context for latency calculations. - Added tests to validate timing context behavior under concurrent and sequential usage scenarios. * feat: implement ModelFactory for dynamic model instance creation - Added ModelFactory class to generate stateless model instances based on provider configurations and model classes. - Updated PluginRegistration to utilize ModelFactory for managing model instances, enhancing code organization and maintainability. * feat: add unit tests for model registry and mock model provider - Introduced a new test file for validating model registry functionality. - Implemented mock classes for model provider and LLM to facilitate testing. - Added tests to ensure correct model instance retrieval from the registry. * apply ruff --------- Co-authored-by: Yeuoly <admin@srmxy.cn>
118 lines
3.7 KiB
Python
118 lines
3.7 KiB
Python
from abc import abstractmethod
|
|
|
|
from pydantic import ConfigDict
|
|
|
|
from dify_plugin.entities.model import EmbeddingInputType, ModelPropertyKey, ModelType
|
|
from dify_plugin.entities.model.text_embedding import TextEmbeddingResult
|
|
from dify_plugin.interfaces.model.ai_model import AIModel
|
|
|
|
|
|
class TextEmbeddingModel(AIModel):
|
|
"""
|
|
Model class for text embedding model.
|
|
"""
|
|
|
|
model_type: ModelType = ModelType.TEXT_EMBEDDING
|
|
|
|
# pydantic configs
|
|
model_config = ConfigDict(protected_namespaces=())
|
|
|
|
############################################################
|
|
# Methods that can be implemented by plugin #
|
|
############################################################
|
|
|
|
@abstractmethod
|
|
def _invoke(
|
|
self,
|
|
model: str,
|
|
credentials: dict,
|
|
texts: list[str],
|
|
user: str | None = None,
|
|
input_type: EmbeddingInputType = EmbeddingInputType.DOCUMENT,
|
|
) -> TextEmbeddingResult:
|
|
"""
|
|
Invoke large language model
|
|
|
|
:param model: model name
|
|
:param credentials: model credentials
|
|
:param texts: texts to embed
|
|
:param user: unique user id
|
|
:param input_type: embedding input type
|
|
:return: embeddings result
|
|
"""
|
|
raise NotImplementedError
|
|
|
|
@abstractmethod
|
|
def get_num_tokens(self, model: str, credentials: dict, texts: list[str]) -> list[int]:
|
|
"""
|
|
Get number of tokens for given prompt messages
|
|
|
|
:param model: model name
|
|
:param credentials: model credentials
|
|
:param texts: texts to embed
|
|
:return:
|
|
"""
|
|
raise NotImplementedError
|
|
|
|
############################################################
|
|
# For plugin implementation use only #
|
|
############################################################
|
|
|
|
def _get_context_size(self, model: str, credentials: dict) -> int:
|
|
"""
|
|
Get context size for given embedding model
|
|
|
|
:param model: model name
|
|
:param credentials: model credentials
|
|
:return: context size
|
|
"""
|
|
model_schema = self.get_model_schema(model, credentials)
|
|
|
|
if model_schema and ModelPropertyKey.CONTEXT_SIZE in model_schema.model_properties:
|
|
return model_schema.model_properties[ModelPropertyKey.CONTEXT_SIZE]
|
|
|
|
return 1000
|
|
|
|
def _get_max_chunks(self, model: str, credentials: dict) -> int:
|
|
"""
|
|
Get max chunks for given embedding model
|
|
|
|
:param model: model name
|
|
:param credentials: model credentials
|
|
:return: max chunks
|
|
"""
|
|
model_schema = self.get_model_schema(model, credentials)
|
|
|
|
if model_schema and ModelPropertyKey.MAX_CHUNKS in model_schema.model_properties:
|
|
return model_schema.model_properties[ModelPropertyKey.MAX_CHUNKS]
|
|
|
|
return 1
|
|
|
|
############################################################
|
|
# For executor use only #
|
|
############################################################
|
|
|
|
def invoke(
|
|
self,
|
|
model: str,
|
|
credentials: dict,
|
|
texts: list[str],
|
|
user: str | None = None,
|
|
input_type: EmbeddingInputType = EmbeddingInputType.DOCUMENT,
|
|
) -> TextEmbeddingResult:
|
|
"""
|
|
Invoke large language model
|
|
|
|
:param model: model name
|
|
:param credentials: model credentials
|
|
:param texts: texts to embed
|
|
:param user: unique user id
|
|
:param input_type: embedding input type
|
|
:return: embeddings result
|
|
"""
|
|
with self.timing_context():
|
|
try:
|
|
return self._invoke(model, credentials, texts, user, input_type)
|
|
except Exception as e:
|
|
raise self._transform_invoke_error(e) from e
|