Documentation
¶
Index ¶
- type Driver
- func (s *Driver) Ancestry(ctx context.Context, hash string) ([]*merkle.Node, error)
- func (s *Driver) AncestryChain(ctx context.Context, hash string) (*storage.Chain, error)
- func (s *Driver) AncestryChains(ctx context.Context, hashes []string) (map[string]*storage.Chain, error)
- func (s *Driver) Close() error
- func (s *Driver) Count() int
- func (s *Driver) CountSessions(_ context.Context, opts storage.ListOpts) (storage.SessionStats, error)
- func (s *Driver) Depth(ctx context.Context, hash string) (int, error)
- func (s *Driver) Get(_ context.Context, hash string) (*merkle.Node, error)
- func (s *Driver) GetByParent(_ context.Context, parentHash *string) ([]*merkle.Node, error)
- func (s *Driver) Has(_ context.Context, hash string) (bool, error)
- func (s *Driver) Leaves(_ context.Context) ([]*merkle.Node, error)
- func (s *Driver) List(_ context.Context) ([]*merkle.Node, error)
- func (s *Driver) ListSessions(_ context.Context, opts storage.ListOpts) (*storage.Page[*merkle.Node], error)
- func (s *Driver) LoadDag(ctx context.Context, hash string) (*merkle.Dag, error)
- func (s *Driver) Migrate(_ context.Context) error
- func (s *Driver) Open(_ context.Context) error
- func (s *Driver) Put(_ context.Context, node *merkle.Node) (bool, error)
- func (s *Driver) Roots(ctx context.Context) ([]*merkle.Node, error)
- func (s *Driver) UpdateUsage(_ context.Context, hash string, usage *llm.Usage) error
Constants ¶
This section is empty.
Variables ¶
This section is empty.
Functions ¶
This section is empty.
Types ¶
type Driver ¶
type Driver struct {
// contains filtered or unexported fields
}
Driver implements Storer using an in-memory map.
func (*Driver) Ancestry ¶
Ancestry returns the path from a node back to its root (node first, root last). See AncestryChain for a variant that also signals when the walk stopped at a missing parent.
func (*Driver) AncestryChain ¶ added in v0.4.0
AncestryChain walks the parent chain starting at hash and returns a Chain describing whether the walk reached a real root, stopped at a parent that is not present in this store, or was guarded out of a cycle.
func (*Driver) AncestryChains ¶ added in v0.4.0
func (s *Driver) AncestryChains(ctx context.Context, hashes []string) (map[string]*storage.Chain, error)
AncestryChains walks each input hash's ancestry and returns a Chain per starting hash. The in-memory driver has O(1) Get, so the batched ent fast path offers no benefit here — this is a straightforward loop over AncestryChain.
func (*Driver) CountSessions ¶ added in v0.4.0
func (s *Driver) CountSessions(_ context.Context, opts storage.ListOpts) (storage.SessionStats, error)
CountSessions returns aggregate counts for the slice of data matching opts. Pagination fields on opts are ignored.
All numeric aggregates apply per-node, matching the SQL aggregate the Postgres driver runs for /v1/stats. completed_count uses leaf-only classification (assistant leaf with a terminal stop_reason) — the chain-context overrides in pkg/sessions.DetermineStatus are deliberately NOT applied here, so both drivers return identical numbers for the same store contents.
TotalDurationNs is wall-clock span (MAX − MIN of CreatedAt) over the matching set, NOT a sum of per-call Usage.TotalDurationNs. Per-call duration is now persisted on nodes (PCC-514); switching this aggregate to SUM is a separate decision since it changes the visible semantic.
func (*Driver) GetByParent ¶
GetByParent retrieves all nodes that have the provided parent. This is useful for determining where branching occurs.
func (*Driver) ListSessions ¶ added in v0.4.0
func (s *Driver) ListSessions(_ context.Context, opts storage.ListOpts) (*storage.Page[*merkle.Node], error)
ListSessions returns a page of leaf nodes (sessions), ordered by created_at descending then hash descending, optionally filtered by opts.
func (*Driver) LoadDag ¶ added in v0.5.0
LoadDag takes a node hash and returns the branch containing that node: its ancestry up to the root and all descendants reachable from that node.
func (*Driver) Put ¶
Put stores a node. Returns true if the node was newly inserted, false if it already existed (no-op due to content-addressing).
Put stores a copy of the node so that storage-managed metadata (currently CreatedAt) can be assigned without mutating the caller.