my-project-public

#1

import importlib

#2

#3

#4

def load_class(class_type):

#5

module_path, class_name = class_type.rsplit(".", 1)

#6

module = importlib.import_module(module_path)

#7

return getattr(module, class_name)

class LlmFactory:

provider_to_class = {

#12

"anthropic": "embedchain.llm.anthropic.AnthropicLlm",

#13

"azure_openai": "embedchain.llm.azure_openai.AzureOpenAILlm",

#14

"cohere": "embedchain.llm.cohere.CohereLlm",

#15

"together": "embedchain.llm.together.TogetherLlm",

#16

"gpt4all": "embedchain.llm.gpt4all.GPT4ALLLlm",

#17

"ollama": "embedchain.llm.ollama.OllamaLlm",

#18

"huggingface": "embedchain.llm.huggingface.HuggingFaceLlm",

#19

"jina": "embedchain.llm.jina.JinaLlm",

#20

"llama2": "embedchain.llm.llama2.Llama2Llm",

#21

"openai": "embedchain.llm.openai.OpenAILlm",

#22

"vertexai": "embedchain.llm.vertex_ai.VertexAILlm",

#23

"google": "embedchain.llm.google.GoogleLlm",

#24

"aws_bedrock": "embedchain.llm.aws_bedrock.AWSBedrockLlm",

#25

"mistralai": "embedchain.llm.mistralai.MistralAILlm",

#26

"clarifai": "embedchain.llm.clarifai.ClarifaiLlm",

#27

"groq": "embedchain.llm.groq.GroqLlm",

#28

"nvidia": "embedchain.llm.nvidia.NvidiaLlm",

#29

"vllm": "embedchain.llm.vllm.VLLM",

#30

}

#31

provider_to_config_class = {

#32

"embedchain": "embedchain.config.llm.base.BaseLlmConfig",

#33

"openai": "embedchain.config.llm.base.BaseLlmConfig",

#34

"anthropic": "embedchain.config.llm.base.BaseLlmConfig",

}

@classmethod

def create(cls, provider_name, config_data):

#39

class_type = cls.provider_to_class.get(provider_name)

#40

# Default to embedchain base config if the provider is not in the config map

#41

config_name = "embedchain" if provider_name not in cls.provider_to_config_class else provider_name

#42

config_class_type = cls.provider_to_config_class.get(config_name)

#43

if class_type:

#44

llm_class = load_class(class_type)

#45

llm_config_class = load_class(config_class_type)

#46

return llm_class(config=llm_config_class(**config_data))

#47

else:

#48

raise ValueError(f"Unsupported Llm provider: {provider_name}")

#49

#50

#51

class EmbedderFactory:

#52

provider_to_class = {

#53

"azure_openai": "embedchain.embedder.azure_openai.AzureOpenAIEmbedder",

#54

"gpt4all": "embedchain.embedder.gpt4all.GPT4AllEmbedder",

#55

"huggingface": "embedchain.embedder.huggingface.HuggingFaceEmbedder",

#56

"openai": "embedchain.embedder.openai.OpenAIEmbedder",

#57

"vertexai": "embedchain.embedder.vertexai.VertexAIEmbedder",

#58

"google": "embedchain.embedder.google.GoogleAIEmbedder",

#59

"mistralai": "embedchain.embedder.mistralai.MistralAIEmbedder",

#60

"clarifai": "embedchain.embedder.clarifai.ClarifaiEmbedder",

#61

"nvidia": "embedchain.embedder.nvidia.NvidiaEmbedder",

#62

"cohere": "embedchain.embedder.cohere.CohereEmbedder",

#63

"ollama": "embedchain.embedder.ollama.OllamaEmbedder",

#64

"aws_bedrock": "embedchain.embedder.aws_bedrock.AWSBedrockEmbedder",

#65

}

#66

provider_to_config_class = {

#67

"azure_openai": "embedchain.config.embedder.base.BaseEmbedderConfig",

#68

"google": "embedchain.config.embedder.google.GoogleAIEmbedderConfig",

#69

"gpt4all": "embedchain.config.embedder.base.BaseEmbedderConfig",

#70

"huggingface": "embedchain.config.embedder.base.BaseEmbedderConfig",

#71

"clarifai": "embedchain.config.embedder.base.BaseEmbedderConfig",

#72

"openai": "embedchain.config.embedder.base.BaseEmbedderConfig",

#73

"ollama": "embedchain.config.embedder.ollama.OllamaEmbedderConfig",

#74

"aws_bedrock": "embedchain.config.embedder.aws_bedrock.AWSBedrockEmbedderConfig",

}

@classmethod

def create(cls, provider_name, config_data):

#79

class_type = cls.provider_to_class.get(provider_name)

#80

# Default to openai config if the provider is not in the config map

#81

config_name = "openai" if provider_name not in cls.provider_to_config_class else provider_name

#82

config_class_type = cls.provider_to_config_class.get(config_name)

#83

if class_type:

#84

embedder_class = load_class(class_type)

#85

embedder_config_class = load_class(config_class_type)

#86

return embedder_class(config=embedder_config_class(**config_data))

#87

else:

#88

raise ValueError(f"Unsupported Embedder provider: {provider_name}")

#89

#90

#91

class VectorDBFactory:

#92

provider_to_class = {

#93

"chroma": "embedchain.vectordb.chroma.ChromaDB",

#94

"elasticsearch": "embedchain.vectordb.elasticsearch.ElasticsearchDB",

#95

"opensearch": "embedchain.vectordb.opensearch.OpenSearchDB",

#96

"lancedb": "embedchain.vectordb.lancedb.LanceDB",

#97

"pinecone": "embedchain.vectordb.pinecone.PineconeDB",

#98

"qdrant": "embedchain.vectordb.qdrant.QdrantDB",

#99

"weaviate": "embedchain.vectordb.weaviate.WeaviateDB",

#100

"zilliz": "embedchain.vectordb.zilliz.ZillizVectorDB",

#101

}

#102

provider_to_config_class = {

#103

"chroma": "embedchain.config.vector_db.chroma.ChromaDbConfig",

#104

"elasticsearch": "embedchain.config.vector_db.elasticsearch.ElasticsearchDBConfig",

#105

"opensearch": "embedchain.config.vector_db.opensearch.OpenSearchDBConfig",

#106

"lancedb": "embedchain.config.vector_db.lancedb.LanceDBConfig",

#107

"pinecone": "embedchain.config.vector_db.pinecone.PineconeDBConfig",

#108

"qdrant": "embedchain.config.vector_db.qdrant.QdrantDBConfig",

#109

"weaviate": "embedchain.config.vector_db.weaviate.WeaviateDBConfig",

#110

"zilliz": "embedchain.config.vector_db.zilliz.ZillizDBConfig",

}

@classmethod

def create(cls, provider_name, config_data):

#115

class_type = cls.provider_to_class.get(provider_name)

#116

config_class_type = cls.provider_to_config_class.get(provider_name)

#117

if class_type:

#118

embedder_class = load_class(class_type)

#119

embedder_config_class = load_class(config_class_type)

#120

return embedder_class(config=embedder_config_class(**config_data))

#121

else:

#122

raise ValueError(f"Unsupported Embedder provider: {provider_name}")

#123

z6Mkq5mY3JWtxoxUobWcfNHm7AkRubgSWEZTkBVqZXJviFZ5/my-project-public