Struct QuantRepository 

pub struct QuantRepository(/* private fields */);

Repository for Quant entities.

All standard CRUD, soft-delete, pagination, and bulk methods are provided automatically via Deref to backbone_orm::GenericCrudRepository.

Implementations

impl QuantRepository

pub fn new(pool: PgPool) -> Self

Create a new repository instance.

impl QuantRepository

Hand-written quant SQL — the reservation apex. Lives here (not in the move engine) per the module’s 4-layer rule: services orchestrate, repositories hold SQL.

pub async fn lock_or_init( &self, conn: &mut PgConnection, dims: QuantDims, ) -> Result<QuantRow, Error>

Lock the quant row for a dimension tuple FOR UPDATE, creating a zeroed row if absent — the reservation-engine analogue of BinRepository::lock_or_init.

This row lock is what serializes competing reservations: two _action_assign calls racing for the same stock queue on this row; the second reads the first’s committed reservation once the lock is granted (READ COMMITTED re-evaluates the row under lock), so reserved_quantity can never exceed what the on-hand actually allowed (R22).

Creation is serialized by a transaction-scoped advisory lock on the dimension key because the partial unique index (item, location, lot, package, owner) WHERE deleted_at IS NULL does NOT dedup NULL dimensions — without the advisory lock two first-writers could both insert. The lock auto-releases at commit/rollback of the caller’s transaction.

pub async fn lock_or_init_on( conn: &mut PgConnection, dims: QuantDims, ) -> Result<QuantRow, Error>

Self::lock_or_init for a caller that holds no repository: every path that may create a quant goes through this one, so they all take the same creation lock.

pub async fn adjust_reserved( &self, conn: &mut PgConnection, quant_id: Uuid, delta: Decimal, ) -> Result<QuantRow, Error>

The RESERVED step of _synchronize_quant (spec §3): move delta onto reserved_quantity — the authoritative counter. available_quantity (the stored T2 compute) is rewritten in the SAME statement as the derived shadow quantity - reserved, so it can never drift. The caller holds the row’s FOR UPDATE from Self::lock_or_init.

delta > 0 reserves; delta < 0 releases. The table CHECK (reserved_quantity >= 0) backstops release-over-reserve (a raw-SQL writer cannot underflow it — ADR-0015).

pub async fn apply_qty( &self, conn: &mut PgConnection, quant_id: Uuid, qty_delta: Decimal, reserved_delta: Decimal, ) -> Result<QuantRow, Error>

The AVAILABLE step of _synchronize_quant (spec §3): physically move stock — qty_delta onto quantity (negative = stock leaves this quant), plus a simultaneous reserved_delta (normally the release of the reservation this move consumed). The stored T2 compute is rewritten in lockstep. Stamp in_date on first arrival (FIFO ordering within the lot) when the quant transitions to a positive on-hand from non-positive.

pub async fn fetch_reservation_candidates( &self, conn: &mut PgConnection, item_id: Uuid, location_id: Uuid, ) -> Result<Vec<QuantRow>, Error>

Reservation candidates at one (item, location): the quants with free availability, oldest arrival first (in_date FIFO ordering — the default removal strategy; explicit fifo/lifo/fefo strategies layer on top when the tracking surface lands). Takes the caller’s transaction: _action_assign locks each candidate as it reserves it.

pub async fn recompute_reserved_from_mirror( &self, conn: &mut PgConnection, quant_id: Uuid, dims: QuantDims, ) -> Result<QuantRow, Error>

Re-derive one quant’s authoritative reserved_quantity from the MIRROR arm: the sum of the live move-line quantities matching this quant’s dimension tuple (state not done/cancel). This is the triangle’s self-heal — the reserved step of _synchronize_quant after a move’s lines flip out of the live set (done or cancel), including the partial-validate case where an operator shrinks a line below what it had reserved (the residual would otherwise stay held by no move). The mirror can only ever reflect reservations the engine minted, so the SUM is the truth by construction.

Takes the caller’s transaction; the caller holds the quant’s FOR UPDATE from Self::lock_or_init. Returns the recomputed row.

impl QuantRepository

pub async fn fetch_on_hand( &self, pool: &PgPool, item_id: Uuid, location_id: Uuid, ) -> Result<QuantOnHandRow, Error>

SUM(quantity) / SUM(reserved_quantity) over one (item, location). Zeroed row when no quant exists (an unreceived item is unavailable, not an error).

pub async fn fetch_warehouse_on_hand( &self, pool: &PgPool, item_ids: &[Uuid], warehouse_id: Uuid, ) -> Result<Vec<QuantWarehouseRow>, Error>

One warehouse-scope availability row (the shop-pivot read): SUM(quantity) / SUM(reserved_quantity) over every quant whose location belongs to the pivot warehouse and holds stock (usage internal). available is NOT read off the stored available_quantity column — the caller computes on_hand − reserved per call, so the number is fresh off the reservation triangle regardless of any stored-compute staleness.

Batch over item_ids = ANY($1); items with no quant rows in the warehouse are simply absent from the result (the caller projects them as zeroed, unavailable rows — an unreceived item is unavailable, not an error). Pool-based and org-scoped: the read rides the ambient org scope the composing service set per request (ADR-0029).

Methods from Deref<Target = GenericCrudRepository<Quant, SoftDelete>>

pub fn pool(&self) -> &Pool<Postgres>

pub fn table_name(&self) -> &str

pub async fn create(&self, entity: &T) -> Result<T, Error>

where T: Serialize + Send + Sync,

Insert a new entity row and return the created row.

pub async fn bulk_create(&self, entities: &[T]) -> Result<Vec<T>, Error>

where T: Serialize + Send + Sync,

Insert multiple entity rows inside a single transaction.

pub async fn run_filtered_query( &self, pagination: PaginationParams, base_condition: Option<&str>, filters: &HashMap<String, String>, column_types: &HashMap<String, String>, search_fields: &[&str], ) -> Result<PaginatedResult<T>, Error>

where T: Send + Sync,

Execute a filtered / paginated query against this entity’s table.

base_condition — when Some, it is inserted as the __base_condition filter key which the ORM injects verbatim into the WHERE clause. Use this to add soft-delete guards without touching the caller-supplied filters.

NOTE: prefer the *_scoped entry points for anything a client can reach — they compose the company fence into this condition. This raw form applies no fence and is for internal/job callers that have already established scope.

pub async fn run_aggregate_query( &self, spec: &AggregateSpec, base_condition: Option<&str>, filters: &HashMap<String, String>, column_types: &HashMap<String, String>, search_fields: &[&str], ) -> Result<AggregateResult, Error>

where T: EntityRepoMeta + Send + Sync,

The aggregate twin of Self::run_filtered_query.

Takes the same base_condition so an aggregate is computed over exactly the row set the matching list would return — a total that counted soft-deleted rows, or another tenant’s, would be wrong in a way no caller could see.

pub async fn find_by_text_field( &self, field: &str, value: &str, ) -> Result<Option<T>, Error>

Find an active entity by a unique text field.

Filters metadata->>'deleted_at' IS NULL automatically.

pub async fn exists_by_text_field( &self, field: &str, value: &str, ) -> Result<bool, Error>

Check existence by a unique text field (active records only).

pub async fn find_by_uuid_field( &self, field: &str, value: Uuid, ) -> Result<Option<T>, Error>

Find an active entity by a unique UUID field.

pub async fn exists_by_uuid_field( &self, field: &str, value: Uuid, ) -> Result<bool, Error>

Check existence by a unique UUID field (active records only).

pub async fn list_paginated_filtered( &self, pagination: PaginationParams, filters: Option<&HashMap<String, String>>, ) -> Result<PaginatedResult<T>, Error>

where T: EntityRepoMeta + Send + Sync,

Paginate active entities with filter and search support.

Requires T: EntityRepoMeta for column type hints and search fields.

pub async fn aggregate_filtered( &self, spec: &AggregateSpec, filters: Option<&HashMap<String, String>>, ) -> Result<AggregateResult, Error>

where T: EntityRepoMeta + Send + Sync,

Group and reduce active entities, skipping the soft-deleted.

pub async fn list_paginated_filtered_scoped( &self, pagination: PaginationParams, filters: Option<&HashMap<String, String>>, company: Option<Uuid>, ) -> Result<PaginatedResult<T>, Error>

where T: EntityRepoMeta + Send + Sync,

Paginate active entities, fenced to company.

The company-scoped counterpart of Self::list_paginated_filtered and the one any client-reachable read should use. Fails closed: a fenced entity with no company returns MissingCompanyScope rather than every company’s rows.

The fence is ANDed into the same __base_condition the soft-delete guard uses, so it lands in SQL — which is what keeps total honest. A post-filter in the handler would fetch limit rows and return the survivors, leaving callers unable to tell “end of data” from “filtered”, and could not fix COUNT at all.

pub async fn list_deleted_filtered_scoped( &self, pagination: PaginationParams, filters: Option<&HashMap<String, String>>, company: Option<Uuid>, ) -> Result<PaginatedResult<T>, Error>

where T: EntityRepoMeta + Send + Sync,

Paginate soft-deleted entities, fenced to company.

/trash is the worse leak of the two: it serves rows whose owners were told the data is gone. Same fence, same fail-closed contract.

pub async fn list_deleted_filtered( &self, pagination: PaginationParams, filters: Option<&HashMap<String, String>>, ) -> Result<PaginatedResult<T>, Error>

where T: EntityRepoMeta + Send + Sync,

Paginate soft-deleted entities with filter support.

pub async fn find_by_id(&self, id: &str) -> Result<Option<T>, Error>

Find an active (non-deleted) entity by primary key.

pub async fn find_all(&self) -> Result<Vec<T>, Error>

Return all active (non-deleted) entities.

pub async fn update(&self, id: &str, entity: &T) -> Result<Option<T>, Error>

where T: Serialize + Send + Sync,

Full update — skips silently if the record is already soft-deleted.

pub async fn delete(&self, id: &str) -> Result<bool, Error>

Soft-delete an entity (sets metadata.deleted_at).

pub async fn count(&self) -> Result<u64, Error>

Count active (non-deleted) entities.

pub async fn exists(&self, id: &str) -> Result<bool, Error>

Return true if an active entity with the given ID exists.

pub async fn list_paginated( &self, pagination: PaginationParams, ) -> Result<PaginatedResult<T>, Error>

Paginate active entities (most-recent-first by ID).

pub async fn soft_delete(&self, id: &str) -> Result<bool, Error>

Set metadata.deleted_at to NOW() (soft delete).

pub async fn restore(&self, id: &str) -> Result<Option<T>, Error>

Remove deleted_at from metadata, restoring the entity.

pub async fn list_deleted( &self, pagination: PaginationParams, ) -> Result<PaginatedResult<T>, Error>

Paginate soft-deleted entities (trash view).

pub async fn empty_trash(&self) -> Result<u64, Error>

Permanently delete all soft-deleted rows (empty trash).

pub async fn find_deleted_by_id(&self, id: &str) -> Result<Option<T>, Error>

Find a soft-deleted entity by primary key.

pub async fn permanent_delete(&self, id: &str) -> Result<bool, Error>

Permanently delete a soft-deleted entity by primary key.

pub async fn count_active(&self) -> Result<u64, Error>

Count active (non-deleted) entities.

pub async fn count_deleted(&self) -> Result<u64, Error>

Count soft-deleted entities.

pub async fn bulk_soft_delete(&self, ids: &[String]) -> Result<u64, Error>

Soft-delete many active rows atomically.

pub async fn bulk_restore(&self, ids: &[String]) -> Result<Vec<T>, Error>

Restore many soft-deleted rows atomically, returning the restored rows.

pub async fn bulk_permanent_delete(&self, ids: &[String]) -> Result<u64, Error>

Permanently delete many soft-deleted rows atomically.

pub async fn restore_all(&self) -> Result<Vec<T>, Error>

Restore every soft-deleted row, returning the restored rows. A single UPDATE ... RETURNING is atomic, and the returned rows let the service layer emit a Restored event per entity.

pub async fn bulk_update(&self, entities: &[T]) -> Result<Vec<T>, Error>

where T: Serialize + Send + Sync,

Update many active rows atomically. Every entity must reference an existing active (non-soft-deleted) row or the whole batch is rolled back.

Trait Implementations

impl CrudRepository<Quant> for QuantRepository

fn fetch_related_json<'life0, 'life1, 'life2, 'async_trait>( &'life0 self, table: &'life1 str, ids: &'life2 [String], ) -> Pin<Box<dyn Future<Output = Vec<Value>> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait, 'life1: 'async_trait, 'life2: 'async_trait,

Hydrate ?include= relations: fetch rows from table by id list, as JSON (raw row_to_json). table comes from EntityRepoMeta::relations() (generator-emitted), never client input. Default: no expansion. The Postgres-backed generated repos override this via impl_crud_repository!.

fn create<'life0, 'async_trait>( &'life0 self, entity: Quant, ) -> Pin<Box<dyn Future<Output = Result<Quant, RepositoryError>> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait,

Create a new entity

fn find_by_id<'life0, 'life1, 'async_trait>( &'life0 self, id: &'life1 str, ) -> Pin<Box<dyn Future<Output = Result<Option<Quant>, RepositoryError>> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait, 'life1: 'async_trait,

Find entity by ID (excluding soft-deleted)

fn find_by_id_including_deleted<'life0, 'life1, 'async_trait>( &'life0 self, id: &'life1 str, ) -> Pin<Box<dyn Future<Output = Result<Option<Quant>, RepositoryError>> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait, 'life1: 'async_trait,

Find entity by ID (including soft-deleted, for trash operations)

fn update<'life0, 'async_trait>( &'life0 self, entity: Quant, ) -> Pin<Box<dyn Future<Output = Result<Quant, RepositoryError>> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait,

Update an existing entity

fn soft_delete<'life0, 'life1, 'async_trait>( &'life0 self, id: &'life1 str, ) -> Pin<Box<dyn Future<Output = Result<bool, RepositoryError>> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait, 'life1: 'async_trait,

Soft delete an entity

fn restore<'life0, 'life1, 'async_trait>( &'life0 self, id: &'life1 str, ) -> Pin<Box<dyn Future<Output = Result<Option<Quant>, RepositoryError>> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait, 'life1: 'async_trait,

Restore a soft-deleted entity

fn hard_delete<'life0, 'life1, 'async_trait>( &'life0 self, id: &'life1 str, ) -> Pin<Box<dyn Future<Output = Result<bool, RepositoryError>> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait, 'life1: 'async_trait,

Permanently delete an entity

fn list<'life0, 'async_trait>( &'life0 self, page: u32, limit: u32, ) -> Pin<Box<dyn Future<Output = Result<(Vec<Quant>, u64), RepositoryError>> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait,

List entities with pagination (excluding soft-deleted)

fn list_filtered<'life0, 'async_trait>( &'life0 self, page: u32, limit: u32, filters: HashMap<String, String>, ) -> Pin<Box<dyn Future<Output = Result<(Vec<Quant>, u64), RepositoryError>> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait,

List entities with pagination and filters (excluding soft-deleted)

fn list_filtered_with_info<'life0, 'async_trait>( &'life0 self, page: u32, limit: u32, filters: HashMap<String, String>, ) -> Pin<Box<dyn Future<Output = Result<(Vec<Quant>, PaginationInfo), RepositoryError>> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait,

list_filtered, carrying the pagination info (cursor positions included) so the HTTP layer can surface keyset paging. Default: the tuple form with the cursor fields absent; the generated Postgres repositories override this with the real keyset walk.

fn table_name(&self) -> Option<&str>

Group and reduce entities under the same filters as list_filtered.

fn aggregate_filtered<'life0, 'life1, 'async_trait>( &'life0 self, spec: &'life1 AggregateSpec, filters: HashMap<String, String>, ) -> Pin<Box<dyn Future<Output = Result<AggregateResult, RepositoryError>> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait, 'life1: 'async_trait,

fn list_deleted<'life0, 'async_trait>( &'life0 self, page: u32, limit: u32, ) -> Pin<Box<dyn Future<Output = Result<(Vec<Quant>, u64), RepositoryError>> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait,

List soft-deleted entities with pagination

fn count<'life0, 'async_trait>( &'life0 self, ) -> Pin<Box<dyn Future<Output = Result<u64, RepositoryError>> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait,

Count all entities (excluding soft-deleted)

fn count_deleted<'life0, 'async_trait>( &'life0 self, ) -> Pin<Box<dyn Future<Output = Result<u64, RepositoryError>> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait,

Count soft-deleted entities

fn bulk_create<'life0, 'async_trait>( &'life0 self, entities: Vec<Quant>, ) -> Pin<Box<dyn Future<Output = Result<Vec<Quant>, RepositoryError>> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait,

Bulk create entities

fn empty_trash<'life0, 'async_trait>( &'life0 self, ) -> Pin<Box<dyn Future<Output = Result<u64, RepositoryError>> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait,

Permanently delete all soft-deleted entities

fn bulk_soft_delete<'life0, 'life1, 'async_trait>( &'life0 self, ids: &'life1 [String], ) -> Pin<Box<dyn Future<Output = Result<u64, RepositoryError>> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait, 'life1: 'async_trait,

Soft-delete many entities by id. Returns the number affected.

fn bulk_restore<'life0, 'life1, 'async_trait>( &'life0 self, ids: &'life1 [String], ) -> Pin<Box<dyn Future<Output = Result<Vec<Quant>, RepositoryError>> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait, 'life1: 'async_trait,

Restore many soft-deleted entities by id. Returns the restored entities.

fn bulk_hard_delete<'life0, 'life1, 'async_trait>( &'life0 self, ids: &'life1 [String], ) -> Pin<Box<dyn Future<Output = Result<u64, RepositoryError>> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait, 'life1: 'async_trait,

Permanently delete many entities by id. Returns the number affected.

fn restore_all<'life0, 'async_trait>( &'life0 self, ) -> Pin<Box<dyn Future<Output = Result<Vec<Quant>, RepositoryError>> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait,

Restore every soft-deleted entity. Returns the restored entities.

fn bulk_update<'life0, 'async_trait>( &'life0 self, entities: Vec<Quant>, ) -> Pin<Box<dyn Future<Output = Result<Vec<Quant>, RepositoryError>> + Send + 'async_trait>>

where Self: 'async_trait, 'life0: 'async_trait,

Update many entities atomically. Returns the updated entities.

fn exists<'life0, 'life1, 'async_trait>( &'life0 self, id: &'life1 str, ) -> Pin<Box<dyn Future<Output = Result<bool, RepositoryError>> + Send + 'async_trait>>

where 'life0: 'async_trait, 'life1: 'async_trait, Self: 'async_trait,

Check if an entity exists by ID

impl Deref for QuantRepository

type Target = GenericCrudRepository<Quant>

The resulting type after dereferencing.

fn deref(&self) -> &Self::Target

Dereferences the value.