Module: Rubino::Memory

Defined in:
lib/rubino/memory.rb,
lib/rubino/memory/store.rb,
lib/rubino/memory/backend.rb,
lib/rubino/memory/flusher.rb,
lib/rubino/memory/backends.rb,
lib/rubino/memory/aux_retry.rb,
lib/rubino/memory/deduplicator.rb,
lib/rubino/memory/sqlite_graph.rb,
lib/rubino/memory/salience_gate.rb,
lib/rubino/memory/threat_scanner.rb,
lib/rubino/memory/backends/sqlite.rb,
lib/rubino/memory/sqlite_extraction.rb,
lib/rubino/memory/sqlite_extraction_prompt.rb

Overview

Helpers shared by the memory backends (the legacy Store and the FTS-backed Sqlite backend), which don’t share a class hierarchy. Pure module functions parameterized by the backend’s own dataset/column, so each backend keeps its distinct schema (‘:memories`/`:content` vs `:memory_facts`/`:text`) and its own budget config while the id-resolution and the budget compare+raise live in one place.

Defined Under Namespace

Modules: AuxRetry, Backends, SalienceGate, SqliteExtraction, SqliteExtractionPrompt, SqliteGraph Classes: Backend, Deduplicator, Flusher, Store, ThreatScanner

Class Method Summary collapse

Class Method Details

.enforce_budget!(group:, limit:, current:, requested:) ⇒ Object

The group-split char-budget check, shared by store/replace/update across backends: raise BudgetExceededError when adding requested chars to a group already holding current would exceed limit. Callers compute group/limit/current themselves (each backend meters a different column/dataset and reads a different config key) and call this only after confirming the limit is set, so a disabled budget never reaches here.

Raises:



39
40
41
42
43
44
45
 
# File 'lib/rubino/memory.rb', line 39

def enforce_budget!(group:, limit:, current:, requested:)
  return if current + requested <= limit

  raise Store::BudgetExceededError.new(
    group: group, limit: limit, current: current, requested: requested
  )
end

.resolve_row(dataset, id) ⇒ Object

Resolve a caller-supplied id to AT MOST ONE row of dataset. A blank id resolves to nothing — a bare-prefix LIKE on “” matches the ‘%` wildcard (EVERY row), so `memory delete “”` would wipe the whole store and report success (#416). An EXACT id wins; a non-empty prefix is accepted ONLY when unambiguous (matches exactly one row), so a short id from `memory list` still resolves but a 1-char prefix can never mass-select. limit(2) distinguishes “one” from “many”.



22
23
24
25
26
27
28
29
30
31
 
# File 'lib/rubino/memory.rb', line 22

def resolve_row(dataset, id)
  key = id.to_s
  return nil if key.strip.empty?

  exact = dataset.where(id: key).first
  return exact if exact

  matches = dataset.where(Sequel.like(:id, "#{key}%")).limit(2).all
  matches.size == 1 ? matches.first : nil
end