Home

Awesome

Generative AI Components (GenAIComps)

Build Enterprise-grade Generative AI Applications with Microservice Architecture

This initiative empowers the development of high-quality Generative AI applications for enterprises via microservices, simplifying the scaling and deployment process for production. It abstracts away infrastructure complexities, facilitating the seamless development and deployment of Enterprise AI services.

GenAIComps

GenAIComps provides a suite of microservices, leveraging a service composer to assemble a mega-service tailored for real-world Enterprise AI applications. All the microservices are containerized, allowing cloud native deployment. Checkout how the microservices are used in GenAIExamples.

Architecture

Installation

pip install opea-comps
git clone https://github.com/opea-project/GenAIComps
cd GenAIComps
pip install -e .

MicroService

Microservices are akin to building blocks, offering the fundamental services for constructing RAG (Retrieval-Augmented Generation) applications.

Each Microservice is designed to perform a specific function or task within the application architecture. By breaking down the system into smaller, self-contained services, Microservices promote modularity, flexibility, and scalability.

This modular approach allows developers to independently develop, deploy, and scale individual components of the application, making it easier to maintain and evolve over time. Additionally, Microservices facilitate fault isolation, as issues in one service are less likely to impact the entire system.

The initially supported Microservices are described in the below table. More Microservices are on the way.

MicroServiceFrameworkModelServingHWDescription
EmbeddingLangChain/LlamaIndexBAAI/bge-base-en-v1.5TEI-GaudiGaudi2Embedding on Gaudi2
EmbeddingLangChain/LlamaIndexBAAI/bge-base-en-v1.5TEIXeonEmbedding on Xeon CPU
RetrieverLangChain/LlamaIndexBAAI/bge-base-en-v1.5TEIXeonRetriever on Xeon CPU
RerankingLangChain/LlamaIndexBAAI/bge-reranker-baseTEI-GaudiGaudi2Reranking on Gaudi2
RerankingLangChain/LlamaIndexBBAAI/bge-reranker-baseTEIXeonReranking on Xeon CPU
ASRNAopenai/whisper-smallNAGaudi2Audio-Speech-Recognition on Gaudi2
ASRNAopenai/whisper-smallNAXeonAudio-Speech-RecognitionS on Xeon CPU
TTSNAmicrosoft/speecht5_ttsNAGaudi2Text-To-Speech on Gaudi2
TTSNAmicrosoft/speecht5_ttsNAXeonText-To-Speech on Xeon CPU
DataprepQdrantsentence-transformers/all-MiniLM-L6-v2NAGaudi2Dataprep on Gaudi2
DataprepQdrantsentence-transformers/all-MiniLM-L6-v2NAXeonDataprep on Xeon CPU
DataprepRedisBAAI/bge-base-en-v1.5NAGaudi2Dataprep on Gaudi2
DataprepRedisBAAI/bge-base-en-v1.5NAXeonDataprep on Xeon CPU
LLMLangChain/LlamaIndexIntel/neural-chat-7b-v3-3TGI GaudiGaudi2LLM on Gaudi2
LLMLangChain/LlamaIndexIntel/neural-chat-7b-v3-3TGIXeonLLM on Xeon CPU
LLMLangChain/LlamaIndexIntel/neural-chat-7b-v3-3Ray ServeGaudi2LLM on Gaudi2
LLMLangChain/LlamaIndexIntel/neural-chat-7b-v3-3Ray ServeXeonLLM on Xeon CPU
LLMLangChain/LlamaIndexIntel/neural-chat-7b-v3-3vLLMGaudi2LLM on Gaudi2
LLMLangChain/LlamaIndexIntel/neural-chat-7b-v3-3vLLMXeonLLM on Xeon CPU

A Microservices can be created by using the decorator register_microservice. Taking the embedding microservice as an example:

from langchain_community.embeddings import HuggingFaceHubEmbeddings

from comps import register_microservice, EmbedDoc, ServiceType, TextDoc


@register_microservice(
    name="opea_service@embedding_tgi_gaudi",
    service_type=ServiceType.EMBEDDING,
    endpoint="/v1/embeddings",
    host="0.0.0.0",
    port=6000,
    input_datatype=TextDoc,
    output_datatype=EmbedDoc,
)
def embedding(input: TextDoc) -> EmbedDoc:
    embed_vector = embeddings.embed_query(input.text)
    res = EmbedDoc(text=input.text, embedding=embed_vector)
    return res

MegaService

A Megaservice is a higher-level architectural construct composed of one or more Microservices, providing the capability to assemble end-to-end applications. Unlike individual Microservices, which focus on specific tasks or functions, a Megaservice orchestrates multiple Microservices to deliver a comprehensive solution.

Megaservices encapsulate complex business logic and workflow orchestration, coordinating the interactions between various Microservices to fulfill specific application requirements. This approach enables the creation of modular yet integrated applications, where each Microservice contributes to the overall functionality of the Megaservice.

Here is a simple example of building Megaservice:

from comps import MicroService, ServiceOrchestrator

EMBEDDING_SERVICE_HOST_IP = os.getenv("EMBEDDING_SERVICE_HOST_IP", "0.0.0.0")
EMBEDDING_SERVICE_PORT = os.getenv("EMBEDDING_SERVICE_PORT", 6000)
LLM_SERVICE_HOST_IP = os.getenv("LLM_SERVICE_HOST_IP", "0.0.0.0")
LLM_SERVICE_PORT = os.getenv("LLM_SERVICE_PORT", 9000)


class ExampleService:
    def __init__(self, host="0.0.0.0", port=8000):
        self.host = host
        self.port = port
        self.megaservice = ServiceOrchestrator()

    def add_remote_service(self):
        embedding = MicroService(
            name="embedding",
            host=EMBEDDING_SERVICE_HOST_IP,
            port=EMBEDDING_SERVICE_PORT,
            endpoint="/v1/embeddings",
            use_remote_service=True,
            service_type=ServiceType.EMBEDDING,
        )
        llm = MicroService(
            name="llm",
            host=LLM_SERVICE_HOST_IP,
            port=LLM_SERVICE_PORT,
            endpoint="/v1/chat/completions",
            use_remote_service=True,
            service_type=ServiceType.LLM,
        )
        self.megaservice.add(embedding).add(llm)
        self.megaservice.flow_to(embedding, llm)

Gateway

The Gateway serves as the interface for users to access the Megaservice, providing customized access based on user requirements. It acts as the entry point for incoming requests, routing them to the appropriate Microservices within the Megaservice architecture.

Gateways support API definition, API versioning, rate limiting, and request transformation, allowing for fine-grained control over how users interact with the underlying Microservices. By abstracting the complexity of the underlying infrastructure, Gateways provide a seamless and user-friendly experience for interacting with the Megaservice.

For example, the Gateway for ChatQnA can be built like this:

from comps import ChatQnAGateway

self.gateway = ChatQnAGateway(megaservice=self.megaservice, host="0.0.0.0", port=self.port)

Contributing to OPEA

Welcome to the OPEA open-source community! We are thrilled to have you here and excited about the potential contributions you can bring to the OPEA platform. Whether you are fixing bugs, adding new GenAI components, improving documentation, or sharing your unique use cases, your contributions are invaluable.

Together, we can make OPEA the go-to platform for enterprise AI solutions. Let's work together to push the boundaries of what's possible and create a future where AI is accessible, efficient, and impactful for everyone.

Please check the Contributing guidelines for a detailed guide on how to contribute a GenAI example and all the ways you can contribute!

Thank you for being a part of this journey. We can't wait to see what we can achieve together!

Additional Content