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.
Python agentagentic-aiagentic-frameworkagentic-workflowaiautomationclaudedeepseekgenaigptllmlow-codemcpnextjsno-codeopenaipythonskillsworkflow
"""Dependencies for native Pydantic AI modules.
``layers`` is the sole source of module config and persisted state. Models are
validated copies and state changes must be dumped back explicitly. Services are
borrowed from the server lifespan; resources exist only inside one native run
and are closed by the capability that acquired them.
"""
from __future__ import annotations
from dataclasses import dataclass, field
from typing import TYPE_CHECKING, TypedDict
import httpx
from pydantic import JsonValue
from dify_agent.agent_stub.shell_env import ShellAgentStubTokenFactory
from dify_agent.runtime_backend import RuntimeBackendProfile, RuntimeLease
if TYPE_CHECKING:
from dify_agent.layers.knowledge_fs.session import KnowledgeFsSessionStore
from dify_agent.layers.shell.layer import ShellSession
class LayerData(TypedDict):
config: dict[str, JsonValue]
state: dict[str, JsonValue]
@dataclass
class Services:
"""Lifespan-owned infrastructure borrowed by every run."""
plugin_daemon_http_client: httpx.AsyncClient
dify_api_http_client: httpx.AsyncClient
plugin_daemon_url: str = "http://localhost:5002"
plugin_daemon_api_key: str = field(default="", repr=False)
inner_api_url: str = "http://localhost:5001"
inner_api_key: str = field(default="", repr=False)
runtime_backend_profile: RuntimeBackendProfile | None = None
shell_redact_patterns: list[str] = field(default_factory=list)
agent_stub_api_base_url: str | None = None
agent_stub_token_factory: ShellAgentStubTokenFactory | None = None
knowledge_session_store: KnowledgeFsSessionStore | None = None
@dataclass
class RunResources:
leases: dict[str, RuntimeLease] = field(default_factory=dict)
shells: dict[str, ShellSession] = field(default_factory=dict)
@dataclass
class Deps:
layers: dict[str, LayerData]
services: Services
run_id: str
resources: RunResources = field(default_factory=RunResources)