Struct TaskContext 

pub struct TaskContext {
    pub thread_pool: Arc<dyn CpuPool>,
    pub database: Arc<DatabaseConnection>,
    pub graph_db: Arc<dyn GraphDBTrait>,
    pub vector_db: Arc<dyn VectorDB>,
    pub cancellation: CancellationToken,
    pub progress: ProgressToken,
    pub pipeline_ctx: Option<PipelineContext>,
    pub exec_status: Arc<dyn ExecStatusManager>,
    pub pipeline_watcher: Option<Arc<dyn PipelineWatcher>>,
}

Runtime dependencies and control tokens for a single pipeline task.

Build via TaskContextBuilder.

Fields

thread_pool: Arc<dyn CpuPool>

CPU-bound work executor (wraps a Rayon pool by default).

database: Arc<DatabaseConnection>

Relational / metadata database connection.

graph_db: Arc<dyn GraphDBTrait>

Graph database.

vector_db: Arc<dyn VectorDB>

Vector database.

cancellation: CancellationToken

Token the task checks to detect cancellation requests.

progress: ProgressToken

Token the task uses to report progress.

pipeline_ctx: Option<PipelineContext>

Pipeline run identity and current data item context.

exec_status: Arc<dyn ExecStatusManager>

Per-item incremental status tracker (deduplication / resume).

pipeline_watcher: Option<Arc<dyn PipelineWatcher>>

Optional pipeline watcher injected by the registry.

When set, the pipeline executor routes lifecycle events here in addition to (or instead of) any watcher passed directly to execute(). Set by PipelineRunRegistry so library functions can publish events without knowing about the registry.

Implementations

impl TaskContext

pub fn pipeline(&self) -> &PipelineContext

Convenience accessor for the pipeline context.

Panics if the context was not set — only call this from tasks that are known to run inside a pipeline executor.

pub fn with_progress(self: &Arc<Self>, progress: ProgressToken) -> Arc<Self>

Create a new Arc<TaskContext> with a different progress token. All other fields are shallow-cloned.

pub fn with_current_data(self: &Arc<Self>, data: Arc<dyn Value>) -> Arc<Self>

Create a new Arc<TaskContext> with current_data set on the pipeline context. All Arc fields are shallow-cloned (cheap reference bumps).

Returns the original Arc unchanged if no pipeline_ctx is present.

pub fn with_user_email(self: &Arc<Self>, email: String) -> Arc<Self>

Create a new Arc<TaskContext> with user_email set on the pipeline context. All Arc fields are shallow-cloned.

Returns the original Arc unchanged if no pipeline_ctx is present.

pub fn with_run_id(self: &Arc<Self>, run_id: Uuid) -> Arc<Self>

Create a new Arc<TaskContext> with run_id set on the pipeline context. All other fields are shallow-cloned.

Returns the original Arc unchanged if no pipeline_ctx is present.

pub async fn publish_payload_field(&self, key: &str, value: Value)

Publish a run-scoped payload field. Tasks running inside crate::pipeline::execute call this to attach metadata that downstream observers read via the registry’s payload accumulator.

Silently no-ops if no pipeline_watcher is attached or if pipeline_ctx.run_id was never set (i.e. the task is not running inside execute()).