Struct WorkflowIr 

pub struct WorkflowIr {

    pub workflow_id: String,
    pub version: String,
    pub name: Option<String>,
    pub description: Option<String>,
    pub state_schema: String,
    pub start_node: String,
    pub nodes: HashMap<String, NodeDef>,
    pub edges: Vec<EdgeDef>,
    pub retry_policies: HashMap<String, RetryPolicy>,
    pub timeouts: TimeoutConfig,
    pub models: HashMap<String, ModelConfig>,
    pub tools: HashMap<String, ToolConfig>,
    pub mcp_servers: HashMap<String, McpServerConfig>,
    pub remote_agents: HashMap<String, RemoteAgentConfig>,
    pub labels: HashMap<String, String>,
    pub policy: Option<PolicySetIr>,
    pub token_budget: Option<TokenBudgetIr>,
    pub cost_budget_usd: Option<f64>,
    pub on_budget_exceeded: Option<String>,
    pub data_policy: Option<DataPolicyIr>,
}

The canonical Intermediate Representation for a JamJet workflow.

Both YAML and Python workflow definitions compile to this struct before being submitted to the runtime. The IR is serializable to JSON and YAML.

Fields

workflow_id: String

Unique workflow definition identifier.

version: String

Semantic version string (e.g., “1.0.0”).

name: Option<String>

Human-readable name.

description: Option<String>

Optional description.

state_schema: String

Reference to the Pydantic model or JSON Schema for workflow state.

start_node: String

The first node to execute.

nodes: HashMap<String, NodeDef>

All nodes in the workflow graph, keyed by node id.

edges: Vec<EdgeDef>

All edges (transitions) between nodes.

retry_policies: HashMap<String, RetryPolicy>

Named retry policies referenced by nodes.

timeouts: TimeoutConfig

Timeout configuration for this workflow.

models: HashMap<String, ModelConfig>

Named model configurations referenced by model nodes.

tools: HashMap<String, ToolConfig>

Named tool configurations referenced by tool nodes.

mcp_servers: HashMap<String, McpServerConfig>

Named MCP server configurations.

remote_agents: HashMap<String, RemoteAgentConfig>

Named remote A2A agents.

labels: HashMap<String, String>

Observability labels attached to all spans from this workflow.

policy: Option<PolicySetIr>

Workflow-level policy set (overrides global; node-level overrides this).

token_budget: Option<TokenBudgetIr>

Per-execution token budget enforcement.

cost_budget_usd: Option<f64>

Per-execution cost cap in USD. Execution fails / branches when exceeded.

on_budget_exceeded: Option<String>

Node to route to when any budget is exceeded (optional). If absent, the execution fails with WorkflowFailed.

data_policy: Option<DataPolicyIr>

Data handling policy (PII redaction + retention controls).

Implementations

impl WorkflowIr

pub fn from_json(s: &str) -> Result<Self, Error>

Parse a WorkflowIr from a JSON string.

pub fn from_yaml(s: &str) -> Result<Self, Error>

Parse a WorkflowIr from a YAML string.

pub fn to_json_pretty(&self) -> Result<String, Error>

Serialize to JSON (pretty-printed).

pub fn node(&self, id: &str) -> Option<&NodeDef>

Look up a node by id.

pub fn edges_from(&self, node_id: &str) -> Vec<&EdgeDef>

Return all edges originating from a given node.

pub fn successors(&self, node_id: &str) -> Vec<&str>

Return the successors of a node (all nodes it can transition to).

Trait Implementations

impl Clone for WorkflowIr

fn clone(&self) -> WorkflowIr

Returns a duplicate of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl Debug for WorkflowIr

fn fmt(&self, f: &mut Formatter<'_>) -> Result

Formats the value using the given formatter.

impl<'de> Deserialize<'de> for WorkflowIr

fn deserialize<__D>(__deserializer: __D) -> Result<Self, __D::Error>

where __D: Deserializer<'de>,

Deserialize this value from the given Serde deserializer.

impl Serialize for WorkflowIr

fn serialize<__S>(&self, __serializer: __S) -> Result<__S::Ok, __S::Error>

where __S: Serializer,

Serialize this value into the given Serde serializer.