artheia 0.1.0

Host-side DSL for Adaptive-AUTOSAR-style nodes that talk to Classical AUTOSAR via the Theia CAN/FR gateway. Inspired by ARText.

artheia

A host-side DSL for modeling Adaptive-AUTOSAR-style components — atomic nodes with explicit TIPC addresses and typed ports, wired into compositions and clusters — plus the code generators that turn an .art model into C++ scaffolds, .proto wire types, deploy manifests, and netgraphs for the Theia runtime.

Inspired by ARText (the syntax aesthetic + the structural concepts: atomic components, ports, prototypes, compositions, connectors), but its own thing — not AUTOSAR-tool-compatible.

Install

pip install -e .            # or: pip install --find-links /opt/theia/wheels artheia

Gives you the artheia CLI, the artheia-lsp language server (for the editor integrations), and artheia-mcp (the generator surface as an MCP server).

The grammar in one breath

Three primitives, bottom to top:

Primitive	Is a	Owns
node	thread	one TIPC type/instance, typed ports, a behavior class
composition	process (one executable)	node prototypes + in-process wiring (connect)
cluster	distribution bundle	compositions + inter-process wiring; the deploy unit

Around them: message/enum (proto3-equivalent data), interface (senderReceiver for streams, clientServer for request/reply), and import across packages.

Example

package system.demo

// --- data ----------------------------------------------------------------
enum Mode { IDLE = 0  RUN = 1 }

message Tick   { uint32 seq }
message SetMode { Mode mode }

// --- contracts -----------------------------------------------------------
interface senderReceiver TickStream { }          // a stream of Tick
interface clientServer   ModeCtl {
    operation Set(in r:SetMode) returns Tick
}

// --- a node: one thread, one TIPC endpoint -------------------------------
node atomic Counter {
    tipc type=0x80020001 instance=0
    ports {
        receiver ticks_in requires TickStream     // inbound stream
        server   ctl      provides ModeCtl        // request/reply surface
    }
}

node atomic Driver {
    tipc type=0x80020002 instance=0
    ports {
        sender ticks_out provides TickStream
        client mode_call requires ModeCtl
    }
}

// --- a composition: the process, wiring prototypes together --------------
composition DemoProc {
    prototype Counter counter
    prototype Driver  driver

    connect driver.ticks_out to counter.ticks_in
    connect driver.mode_call to counter.ctl
}

// --- a cluster: the deploy/package unit ----------------------------------
cluster Demo {
    composition DemoProc proc
}

What you do with a model

artheia parse        system/system.art                 # resolve + validate the tree
artheia check-addresses system/system.art              # assert TIPC addresses unique
artheia gen-app --kind fc <component>.art --out apps    # C++ scaffold (lib/main/impl)
artheia gen-proto    <component>.art --out platform/proto
artheia gen-netgraph <component>.art --out netgraph.json
artheia gen-manifest <component>.art manifest/app.py    # deploy manifest module

Run artheia --help for the full command set (~30 generators incl. DBC/FIBEX import, config-schema migration, and the rig/manifest pipeline). The gen_server shape follows OTP — handle_call / handle_cast / handle_info per node.

Editor support

.art files get highlighting + LSP (diagnostics, goto-definition, completion) in VS Code and Emacs via the artheia-lsp server — see the umbrella repo's contrib/editors/.

Layout

artheia/
  grammar/*.tx        textX grammar
  model/              name resolution + validators
  generators/         the gen-* code generators + Jinja2 templates
  lsp/                the artheia-lsp language server (pygls)
  adapters/           the artheia-mcp server (FastMCP)
  cli.py              the `artheia` CLI (Click)

License

Apache-2.0 — see LICENSE.