Struct RunContext 

pub struct RunContext { /* private fields */ }

Thread-safe storage for task results (lock-free)

Uses Arc<str> keys for zero-cost cloning with same Arc used in events.

Added context storage for workflow context: block. Added inputs storage for workflow inputs: block.

Implementations

impl RunContext

pub fn new() -> Self

pub fn insert(&self, task_id: Arc<str>, result: TaskResult)

Insert a task result (accepts Arc<str> for zero-cost key reuse)

pub fn get(&self, task_id: &str) -> Option<TaskResult>

Get a task result

pub fn contains(&self, task_id: &str) -> bool

Check if task exists

pub fn is_completed_successfully(&self, task_id: &str) -> Option<bool>

Check if a task completed successfully without cloning the full TaskResult.

Returns None if the task doesn’t exist, Some(true) if succeeded, Some(false) if failed. Avoids O(T*D) TaskResult cloning in get_ready_tasks() hot path.

pub fn iter_results(&self) -> Vec<(Arc<str>, TaskResult)>

Iterate over all task results (cloned).

Returns (task_id, TaskResult) pairs for all stored results. Used by integrity checks at workflow end.

Note: for_each tasks store both individual iteration entries (task[0], task[1], …) and an aggregated parent entry (task). Media refs appear in both, so callers doing per-file checks may see duplicates. This is acceptable for warn-only checks.

pub fn is_success(&self, task_id: &str) -> bool

Check if task succeeded

pub fn is_failed(&self, task_id: &str) -> bool

Check if task failed (either directly or due to dependency failure)

pub fn is_dependency_failed(&self, task_id: &str) -> bool

Check if task failed due to a dependency failure

pub fn get_failed_dependency(&self, task_id: &str) -> Option<String>

Get the failed dependency name if task has DependencyFailed status

pub fn get_output(&self, task_id: &str) -> Option<Arc<Value>>

Get just the output Value for a task (for JSONPath resolution) Returns Arc<Value> for O(1) cloning instead of deep copy

pub fn set_media(&self, task_id: &Arc<str>, media: Vec<MediaRef>)

Stage media refs for a task (called from run_invoke).

pub fn take_media(&self, task_id: &Arc<str>) -> Vec<MediaRef>

Take staged media refs for a task (called from runner after building TaskResult). Returns empty vec if no media was staged.

pub fn media_budget(&self) -> &Arc<MediaBudget>

Get the shared per-run media budget.

pub fn set_workspace_root(&self, root: PathBuf)

Set the workspace root (called by Runner at workflow start).

pub fn workspace_root(&self) -> PathBuf

Get the workspace root path (cloned).

pub fn resolve_path(&self, path: &str) -> Option<Value>

Resolve a dot-separated path (e.g., “weather.summary”)

Uses jsonpath module internally for unified path resolution. Supports both simple dot notation and array indices.

Media paths are intercepted before standard output resolution:

• "task_id.media" → full media array as JSON
• "task_id.media[0].hash" → specific media ref field
• "task_id.media[0].path" → specific media ref field

pub fn set_context(&self, context: LoadedContext)

Set workflow context

Called by Runner at workflow start after loading context files.

pub fn get_context_file(&self, alias: &str) -> Option<Value>

Get a context file by alias

Returns the loaded value for {{context.files.alias}} bindings.

pub fn get_context_session(&self) -> Option<Value>

Get session data

Returns the loaded session for {{context.session.key}} bindings.

pub fn has_context(&self) -> bool

Check if context is loaded

pub fn resolve_context_path(&self, path: &str) -> Option<Value>

Resolve a context path

Supports:

• context.files.alias → file content
• context.files.alias.field → nested field
• context.session → session data
• context.session.field → session field

pub fn set_inputs(&self, inputs: FxHashMap<String, Value>)

Set workflow inputs

Called by Runner at workflow start with input definitions. Each input is a JSON object with type, default, description, etc.

pub fn get_input_default(&self, name: &str) -> Option<Value>

Get an input’s value by name

Supports two formats:

• Full form: { type: string, default: "value" } → extracts default field
• Shorthand: "value" or 123 or true → uses value directly

Returns None if input doesn’t exist.

pub fn has_inputs(&self) -> bool

Check if inputs are loaded

pub fn resolve_input_path(&self, path: &str) -> Option<Value>

Resolve an input path

Supports:

• inputs.param → default value of parameter
• inputs.param.field → nested field in default value (if object)

Trait Implementations

impl Clone for RunContext

fn clone(&self) -> RunContext

Returns a duplicate of the value.

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

Performs copy-assignment from source.

impl Default for RunContext

fn default() -> Self

Returns the “default value” for a type.