open-brainy/src/db/types.ts

/**
 * @module db/types
 * @description Shared types for the 8.0 generational-MVCC Db API: the
 * declarative transaction-operation union consumed by `brain.transact()`,
 * the options bags for `transact()` / `compactHistory()`, the persisted
 * record shapes of the generational record layer, and the narrow storage
 * contract (`GenerationStorage`) the record layer needs from an adapter.
 *
 * Persisted layout (all paths relative to the storage root):
 *
 * ```
 * _system/generation.json        — { generation, updatedAt }   (atomic tmp+rename)
 * _system/manifest.json          — { version, generation, committedAt, horizon }
 *                                   (atomic tmp+rename — the rename IS the commit point)
 * _system/tx-log.jsonl           — one JSON line per committed transact:
 *                                   { generation, timestamp, meta? }  (append-only)
 * _generations/<N>/tx.json       — the generation-N delta: touched noun/verb ids + meta
 * _generations/<N>/prev/<id>.json — immutable before-image of <id> as of commit N
 * ```
 *
 * Design note: the manifest holds the committed watermark + compaction
 * horizon rather than a full `id → latest record generation` map. The id
 * mapping is distributed across the per-generation `tx.json` deltas, which
 * makes each commit O(ids touched) instead of O(all ids) — see
 * `docs/ADR-001-generational-mvcc.md` for the full justification.
 */

import type { AddParams, UpdateParams, RelateParams, Entity, Relation } from '../types/brainy.types.js'

// ============================================================================
// Transaction operations (brain.transact input)
// ============================================================================

/**
 * @description Create an entity. Carries the same parameters as
 * `brain.add()`; supply `id` to choose the entity id (otherwise a UUID v4 is
 * generated and returned in the operation's planning, exactly as `add()`
 * does).
 */
export interface TxAddOperation<T = any> extends AddParams<T> {
  /** Discriminator. */
  op: 'add'
}

/**
 * @description Update an entity. Carries the same parameters as
 * `brain.update()` including per-entity CAS via `ifRev` — an `ifRev`
 * conflict rejects the whole transaction before anything is applied.
 */
export interface TxUpdateOperation<T = any> extends UpdateParams<T> {
  /** Discriminator. */
  op: 'update'
}

/**
 * @description Remove an entity (and, exactly like `brain.remove()`, every
 * relationship where it is source or target — the cascade is part of the
 * same atomic batch).
 */
export interface TxRemoveOperation {
  /** Discriminator. */
  op: 'remove'
  /** Id of the entity to remove. */
  id: string
}

/**
 * @description Create a relationship. Carries the same parameters as
 * `brain.relate()` (including `bidirectional`). Duplicate relationships
 * (same from/to/type) are deduplicated exactly as `relate()` does — the
 * operation becomes a no-op that resolves to the existing relationship id.
 */
export interface TxRelateOperation<T = any> extends RelateParams<T> {
  /** Discriminator. */
  op: 'relate'
}

/**
 * @description Delete a relationship by id (mirror of `brain.unrelate()`).
 */
export interface TxUnrelateOperation {
  /** Discriminator. */
  op: 'unrelate'
  /** Id of the relationship to delete. */
  id: string
}

/**
 * @description One declarative operation inside a `brain.transact()` batch.
 * The batch executes atomically: either every operation applies and the
 * store advances exactly one generation, or none apply and the store is
 * byte-identical to its pre-transaction state.
 */
export type TxOperation<T = any> =
  | TxAddOperation<T>
  | TxUpdateOperation<T>
  | TxRemoveOperation
  | TxRelateOperation<T>
  | TxUnrelateOperation

/**
 * @description Options for `brain.transact()`.
 */
export interface TransactOptions {
  /**
   * Transaction metadata, reified Datomic-style: recorded in
   * `_system/tx-log.jsonl` alongside the generation number and commit
   * timestamp. Use it for audit fields (author, reason, request id) —
   * it replaces commit messages from the pre-8.0 versioning API.
   */
  meta?: Record<string, unknown>
  /**
   * Whole-store compare-and-swap. When provided, the transaction commits
   * only if the store's current generation equals this value; otherwise it
   * throws `GenerationConflictError` (with `expected`/`actual`) before any
   * record is staged.
   */
  ifAtGeneration?: number
}

/**
 * @description Per-operation results of a committed transaction, in input
 * order. `add` resolves to the created entity id, `relate` to the created
 * (or deduplicated existing) relationship id; `update`/`remove`/`unrelate`
 * resolve to the id they acted on.
 */
export interface TransactReceipt {
  /** The generation this transaction committed as. */
  generation: number
  /** Commit timestamp (ms since epoch) recorded in the tx-log. */
  timestamp: number
  /** Resolved id per input operation, in input order. */
  ids: string[]
}

// ============================================================================
// Compaction
// ============================================================================

/**
 * @description Options for `brain.compactHistory()` — explicit retention
 * **caps** (Model-B `retention` knob). Each supplied field is an upper bound:
 * the oldest unpinned generations are reclaimed while **ANY** supplied cap is
 * exceeded (predictable ops — the more you constrain, the more is reclaimed).
 * Live `Db` pins are ALWAYS exempt. With no options, every generation not
 * protected by a live pin is eligible.
 *
 * (8.0 rename: the pre-release `retainGenerations`/`retainMs` *floors* became
 * `maxGenerations`/`maxAge` *caps* — the `max*` naming signals the semantics.)
 */
export interface CompactHistoryOptions {
  /**
   * Keep at most this many of the most recent committed generations — reclaim
   * the oldest unpinned generations while the count exceeds it (`0` = reclaim
   * everything not protected by a live pin).
   */
  maxGenerations?: number
  /**
   * Keep only generations committed within the last `maxAge` milliseconds —
   * reclaim unpinned generations older than the window.
   */
  maxAge?: number
  /**
   * Keep total generational-history bytes at or below this — reclaim the oldest
   * unpinned generations while the total exceeds it. Byte accounting is the sum
   * of each surviving generation's serialized record set (`GenerationDelta.bytes`).
   */
  maxBytes?: number
}

/**
 * @description Result of `brain.compactHistory()`.
 */
export interface CompactHistoryResult {
  /** Number of generation record-sets reclaimed by this call. */
  removedGenerations: number
  /**
   * The new compaction horizon — the highest generation whose record-set has
   * been reclaimed. Generations BELOW the horizon are no longer reachable via
   * `asOf()` (their reads would need the reclaimed before-images); the
   * horizon itself stays reachable, resolved from the record-sets above it.
   */
  horizon: number
}

// ============================================================================
// Db surfaces
// ============================================================================

/**
 * @description Result of `db.since(priorDb)` — the ids touched by committed
 * generations in the interval `(priorDb.generation, db.generation]`. Under
 * Model-B EVERY write is its own generation, so single-operation writes
 * (`add`/`update`/`remove`/`relate` outside `transact()`) ARE reflected here,
 * exactly like `transact()` commits.
 */
export interface ChangedIds {
  /** Entity (noun) ids touched in the interval, sorted. */
  nouns: string[]
  /** Relationship (verb) ids touched in the interval, sorted. */
  verbs: string[]
  /** Exclusive lower bound of the interval. */
  fromGeneration: number
  /** Inclusive upper bound of the interval. */
  toGeneration: number
}

/**
 * @description The result of {@link Brainy.diff}: ids that came into existence,
 * went away, or changed value between two states, each split by kind. Unlike the
 * raw touched-id set of {@link ChangedIds}, `diff` EARNS its name — it resolves
 * each touched id at BOTH endpoints and classifies by existence (added/removed)
 * and stable value comparison (modified). An id touched within the interval but
 * whose endpoint states are identical (touched-then-reverted) appears in NONE of
 * the three buckets. Orientation is `a → b`: `added` exists at `b` not `a`.
 */
export interface DiffResult {
  /** Ids absent at `a`, present at `b`. */
  added: { nouns: string[]; verbs: string[] }
  /** Ids present at `a`, absent at `b`. */
  removed: { nouns: string[]; verbs: string[] }
  /** Ids present at both endpoints with a changed stored value. */
  modified: { nouns: string[]; verbs: string[] }
  /** The resolved generation of endpoint `a`. */
  fromGeneration: number
  /** The resolved generation of endpoint `b`. */
  toGeneration: number
}

/**
 * @description One version of a single entity or relationship in its
 * {@link EntityHistory} — the materialized state at a committed generation that
 * changed it. `value` is `null` when the id did not exist at that version
 * (created-after / removed). Granularity is per-write (every `transact()` AND
 * single-op write is its own generation — Model-B).
 */
export interface HistoryVersion<T = any> {
  /** The committed generation that produced this version. */
  generation: number
  /** Commit timestamp of that generation (ms since epoch). */
  timestamp: number
  /** Materialized state at this version, or `null` if the id did not exist then. */
  value: Entity<T> | Relation<T> | null
}

/**
 * @description The result of {@link Brainy.history}: every distinct version of
 * ONE id over a generation range, oldest → newest. Consecutive identical states
 * are collapsed (one entry per distinct state). `kind` is auto-detected from the
 * id's presence in the noun vs verb spaces.
 */
export interface EntityHistory<T = any> {
  /** The id whose history this is. */
  id: string
  /** Whether `id` is an entity (`'noun'`) or relationship (`'verb'`). */
  kind: 'noun' | 'verb'
  /** Distinct versions, oldest first. */
  versions: HistoryVersion<T>[]
}

// ============================================================================
// Persisted record shapes
// ============================================================================

/**
 * @description An immutable before-image of one entity or relationship,
 * persisted under `_generations/<N>/prev/<id>.json` — the state of the id
 * immediately before commit `N` touched it. `metadata`/`vector` hold the
 * *raw stored objects* (exact bytes that live at the entity's canonical
 * storage paths); `null` means the corresponding file did not exist (for
 * before-images of created ids, both are `null`). Before-images are both
 * the crash-recovery undo log and the point-in-time read source: the state
 * of id X at pinned generation G is the before-image of the first committed
 * generation after G that touched X.
 */
export interface GenerationRecord {
  /** Whether this record images an entity (noun) or a relationship (verb). */
  kind: 'noun' | 'verb'
  /** Raw stored metadata object, or `null` if the metadata file was absent. */
  metadata: unknown | null
  /** Raw stored vector object, or `null` if the vector file was absent. */
  vector: unknown | null
}

/**
 * @description The per-generation delta persisted at
 * `_generations/<N>/tx.json`. For a `transact()` generation it is written and
 * fsynced *before* any operation of the transaction executes, so crash
 * recovery always knows exactly which ids to restore from before-images. For a
 * single-operation (Model-B) generation it is written at group-commit flush
 * time with `groupCommit: true` — see that flag.
 */
export interface GenerationDelta {
  /** The generation this delta belongs to. */
  generation: number
  /** Commit timestamp (ms since epoch). */
  timestamp: number
  /** Transaction metadata (mirrored into the tx-log on commit). */
  meta?: Record<string, unknown>
  /** Entity ids touched by this generation. */
  nouns: string[]
  /** Relationship ids touched by this generation. */
  verbs: string[]
  /**
   * Content-blob hashes referenced by this generation's before-image records
   * (a MULTISET — one entry per referencing record occurrence), captured at
   * persist time so compaction can release the exact history references it
   * reclaims without re-reading the records. Always present (possibly empty)
   * on deltas written under the temporal-blob contract; absent only on
   * pre-contract generations, for which compaction falls back to reading the
   * record-set itself.
   */
  blobHashes?: string[]
  /**
   * `true` for a Model-B single-operation generation persisted by the async
   * group-commit flush (`GenerationStore.flushPendingSingleOps`). It marks the
   * **drop-without-restore** crash-recovery contract: the live write already
   * mutated canonical storage and was acknowledged to the caller BEFORE the
   * flush, so an uncommitted (crashed-mid-flush) generation with this flag set
   * must be DISCARDED — its before-images must NEVER be restored, or recovery
   * would silently revert acknowledged user writes. Absent/`false` on a
   * `transact()` generation, whose before-images ARE the durable undo log and
   * are restored on rollback (the before-image + execute are one atomic unit).
   */
  groupCommit?: boolean
  /**
   * Serialized byte size of this generation's record set (the `prev/<id>.json`
   * before-images plus this delta), recorded at write time so `maxBytes` /
   * adaptive retention can bound total history without a storage size API. Read
   * back lazily by `GenerationStore.historyBytes()`. Absent on generations
   * written before byte accounting existed (treated as 0).
   */
  bytes?: number
}

/**
 * @description Shape of `_system/manifest.json`. Replaced via atomic
 * tmp+rename on every commit — the rename is the commit point: a generation
 * directory is committed if and only if its number is ≤ `generation`.
 */
export interface GenerationManifest {
  /** Schema version of the manifest file. */
  version: 1
  /** Committed-transaction watermark. */
  generation: number
  /** ISO timestamp of the most recent commit (or compaction). */
  committedAt: string
  /** Compaction horizon — record-sets for generations ≤ this are reclaimed. */
  horizon: number
}

/**
 * @description One line of `_system/tx-log.jsonl`.
 */
export interface TxLogEntry {
  /** The committed generation. */
  generation: number
  /** Commit timestamp (ms since epoch). */
  timestamp: number
  /** Transaction metadata, when supplied to `transact()`. */
  meta?: Record<string, unknown>
}

// ============================================================================
// Storage contract
// ============================================================================

/**
 * @description The narrow storage contract the generational record layer
 * needs. `BaseStorage` implements it structurally (both shipped adapters —
 * filesystem and memory — inherit the implementation), so any
 * `StorageAdapter` produced by the 8.0 storage factory satisfies it.
 *
 * Raw-object methods bypass branch scoping and operate on storage-root-
 * relative paths (the record layer owns `_system/` + `_generations/`);
 * entity-raw methods read/write the canonical entity files (branch-scoped,
 * write-cache coherent) so before/after images capture exactly the bytes
 * the live read paths see.
 */
export interface GenerationStorage {
  /** Read a raw object at a storage-root-relative path (`null` if absent). */
  readRawObject(path: string): Promise<any | null>
  /** Write a raw object at a storage-root-relative path (atomic tmp+rename on disk). */
  writeRawObject(path: string, data: any): Promise<void>
  /** Delete a raw object (no-op if absent). */
  deleteRawObject(path: string): Promise<void>
  /** List raw object paths under a prefix (normalized, `.gz`-stripped). */
  listRawObjects(prefix: string): Promise<string[]>
  /** Remove every object under a prefix (and the directory itself on disk). */
  removeRawPrefix(prefix: string): Promise<void>
  /** Durability barrier: fsync the given object paths (no-op in memory). */
  syncRawObjects(paths: string[]): Promise<void>

  /**
   * OPTIONAL transaction durability barrier. `commitTransaction` calls
   * {@link beginWriteBarrier} immediately before running the planned operations
   * and {@link flushWriteBarrier} immediately after — BEFORE the generation
   * counter and manifest are advanced. An adapter whose canonical writes are
   * not synchronously durable (the filesystem adapter's tmp+rename lands in the
   * page cache) MUST implement these so a transaction reported "committed" is
   * durable before its generation stamp: otherwise a hard kill can leave the
   * fsync'd counter ahead of the still-buffered entity bytes (phantom progress
   * for any generation-based consumer). Adapters whose writes are already
   * durable per-call (cloud object PUT) may leave these undefined — the store
   * treats them as no-ops. `beginWriteBarrier` also resets any tracking left by
   * an aborted prior transaction.
   */
  beginWriteBarrier?(): void
  /** @see beginWriteBarrier — fsync every canonical write since begin. */
  flushWriteBarrier?(): Promise<void>

  /** Read an entity's raw stored metadata+vector objects. */
  readNounRaw(id: string): Promise<{ metadata: any | null; vector: any | null }>
  /** Restore an entity's raw stored objects (`null` part ⇒ delete that file). */
  writeNounRaw(id: string, record: { metadata: any | null; vector: any | null }): Promise<void>
  /** Read a relationship's raw stored metadata+vector objects. */
  readVerbRaw(id: string): Promise<{ metadata: any | null; vector: any | null }>
  /** Restore a relationship's raw stored objects (`null` part ⇒ delete that file). */
  writeVerbRaw(id: string, record: { metadata: any | null; vector: any | null }): Promise<void>

  /** Append one line to `_system/tx-log.jsonl`. */
  appendTxLogLine(line: string): Promise<void>
  /** Read all lines of `_system/tx-log.jsonl` (empty array if absent). */
  readTxLogLines(): Promise<string[]>

  /**
   * OPTIONAL binary raw-byte primitives — the substrate for the generation
   * fact log's append-only CRC-framed segments. Feature-detected: a storage
   * layer that omits them hosts no fact log (dual-write is skipped; readers
   * fall back to canonical enumeration). Paths are used VERBATIM (no
   * suffixing). Append durability rides `syncRawObjects` at the commit
   * barrier, exactly like the staged history files.
   */
  appendRawBytes?(path: string, bytes: Uint8Array): Promise<void>
  /** Read a raw binary file whole; absent → null; a real fault throws. */
  readRawBytes?(path: string): Promise<Uint8Array | null>
  /** Replace a raw binary file atomically (tmp → fsync → rename). */
  writeRawBytes?(path: string, bytes: Uint8Array): Promise<void>
  /** Byte size of a raw binary file, or null when absent. */
  rawByteSize?(path: string): Promise<number | null>

  /**
   * OPTIONAL temporal-blob contract (implemented by blob-aware storage; the
   * generation store treats the hashes as opaque strings). Extract the
   * content-blob hashes a record-set references — a pure multiset extraction,
   * no side effects.
   */
  extractBlobHashesFromRecords?(records: GenerationRecord[]): string[]
  /**
   * OPTIONAL: record history references for the given hashes (one increment
   * per occurrence). Called BEFORE the referencing record-set is persisted —
   * a crash between the two can only over-count (a leak the scrub repairs),
   * never under-count (which would risk reclaiming bytes history still needs).
   */
  recordHistoryBlobReferences?(hashes: string[]): Promise<void>
  /**
   * OPTIONAL: release history references for the given hashes (one decrement
   * per occurrence) and physically reclaim any hash left with zero live AND
   * zero history references. Called AFTER the referencing record-set is
   * deleted by compaction — same over-count-only crash ordering.
   */
  releaseHistoryBlobReferences?(hashes: string[]): Promise<void>

  /**
   * Register the generation-bump hook invoked on every entity-visible
   * single-operation write (see `BaseStorage.setGenerationBumpHook`).
   */
  setGenerationBumpHook(hook: (() => void) | undefined): void

  /** Snapshot the entire store to a directory (hard-link farm on disk). */
  snapshotToDirectory(targetPath: string): Promise<void>
  /** Replace the entire store's contents from a snapshot directory. */
  restoreFromDirectory(sourcePath: string): Promise<void>
}