Skip to content

armgit / dify

Public Imported from GitHub
0

Build Agentic workflows, RAG pipelines, with rich AI model and tool support on one collaborative workspace. Deploy on cloud, VPC, or self-hosted, so teams move from prototype to production without rebuilding the stack.

1.6 KB
Raw Blame
"""Composition of the Agent App console use cases."""

from dataclasses import dataclass

from sqlalchemy.orm import Session, sessionmaker

from configs import dify_config
from repositories.app.agent_app_repository import AgentAppRepository
from repositories.app.agent_sandbox_repository import AgentSandboxRepository
from services.app.agent_app_feature_gateway import AgentAppFeatureValidator
from services.app.agent_app_sandbox_service import AgentAppSandboxService
from services.app.agent_app_service import AgentAppAccessService, AgentAppFeatureConfigService
from services.app.agent_sandbox_file_gateway import AgentSandboxFileGateway, create_sandbox_client
from services.file_request_service import FileRequestService


@dataclass(frozen=True, slots=True)
class AgentAppServices:
    access: AgentAppAccessService
    features: AgentAppFeatureConfigService
    sandbox: AgentAppSandboxService


def build_agent_app_services(*, database_client: sessionmaker[Session]) -> AgentAppServices:
    apps = AgentAppRepository(session_factory=database_client)
    return AgentAppServices(
        access=AgentAppAccessService(references=apps),
        features=AgentAppFeatureConfigService(features=apps, validator=AgentAppFeatureValidator()),
        sandbox=AgentAppSandboxService(
            bindings=AgentSandboxRepository(session_factory=database_client, apps=apps),
            files=AgentSandboxFileGateway(
                client_factory=create_sandbox_client,
                file_requests=FileRequestService(),
                files_url=dify_config.FILES_URL,
            ),
        ),
    )