Struct AddPipeline 

pub struct AddPipeline { /* private fields */ }

Ingest pipeline driven by the cognee-core task framework.

Routes add / add_with_params through cognee_core::pipeline::execute so the executor’s lifecycle hooks fire and TaskContext-aware tasks can publish run-scoped payload (LIB-06 / gap 08-07).

Backend handles required by cognee_core::TaskContextBuilder — thread_pool, graph_db, vector_db, database — must be attached via the chainable builders before calling add(). AddPipeline::new does not populate them; calling add() on an under-configured pipeline returns IngestionError::MissingBackend { ... }.

For composable pipeline-based execution (with concurrency, retry, etc.), use build_add_pipeline + cognee_core::execute directly.

Implementations

impl AddPipeline

pub fn new( storage: Arc<dyn StorageTrait>, database: Arc<dyn IngestDb>, ) -> AddPipeline

Create with the default MD5 hashing (Python-compatible).

Note: add() routes through cognee_core::pipeline::execute and requires the four executor-context handles attached via with_thread_pool, with_graph_db, with_vector_db, with_database. A missing handle surfaces as IngestionError::MissingBackend at add() time.

pub fn new_with_algorithm( storage: Arc<dyn StorageTrait>, database: Arc<dyn IngestDb>, hash_algorithm: HashAlgorithm, ) -> AddPipeline

Create with an explicit hash algorithm.

pub fn with_acl_db(self, acl_db: Arc<dyn AclDb>) -> AddPipeline

Enable ACL permission grants on newly created datasets.

When set, the pipeline grants all four permissions (read, write, delete, share) to the owner on each newly created dataset, matching Python’s create_authorized_dataset() behavior.

pub fn with_thread_pool(self, pool: Arc<dyn CpuPool>) -> AddPipeline

Attach the CPU pool used by cognee_core::TaskContext.

pub fn with_graph_db(self, graph: Arc<dyn GraphDBTrait>) -> AddPipeline

Attach the graph database backend used by cognee_core::TaskContext.

pub fn with_vector_db(self, vectors: Arc<dyn VectorDB>) -> AddPipeline

Attach the vector database backend used by cognee_core::TaskContext.

pub fn with_database(self, db: Arc<DatabaseConnection>) -> AddPipeline

Attach the relational DatabaseConnection used by cognee_core::TaskContext. This is the same SeaORM handle the SQL-backed IngestDb is built on.

pub fn with_pipeline_run_repo( self, repo: Arc<dyn PipelineRunRepository>, ) -> AddPipeline

Attach the PipelineRunRepository used to persist the four-state pipeline_runs trail (gap 08-07).

Embedded callers pass Arc::new(NoopPipelineRunRepository::new()) (or simply omit this call — the pipeline defaults to no-op). CLI and HTTP callers pass Arc::new(SeaOrmPipelineRunRepository::new(db)) so the rows surface in /api/v1/activity/pipeline-runs.

pub async fn add( &self, inputs: Vec<DataInput>, dataset_name: &str, owner_id: Uuid, tenant_id: Option<Uuid>, ) -> Result<Vec<Data>, Box<dyn Error>>

pub async fn add_with_params( &self, inputs: Vec<DataInput>, dataset_name: &str, owner_id: Uuid, tenant_id: Option<Uuid>, params: &AddParams, ) -> Result<Vec<Data>, Box<dyn Error>>

Like add, but accepts additional optional parameters.