orka.memory.presets module

Memory Presets System - Minsky-Inspired Cognitive Architecture

This module implements a simplified memory configuration system based on Marvin Minsky’s cognitive theories from “The Society of Mind” and related work. It provides 6 preset memory types that cover the spectrum of cognitive memory needs.

Minsky-Inspired Memory Hierarchy

The 6 memory types are inspired by Minsky’s cognitive architecture concepts:

  1. Sensory Memory - Immediate, high-throughput processing

  2. Working Memory - Active processing and temporary storage

  3. Episodic Memory - Experience and interaction history

  4. Semantic Memory - Knowledge and learned facts

  5. Procedural Memory - Skills, patterns, and processes

  6. Meta Memory - System knowledge and self-reflection

Each preset includes: - Optimized decay rules for the memory type - Appropriate importance scoring - Vector search configuration - TTL and cleanup settings - Namespace organization

Usage Examples

Simple YAML Configuration:

```yaml orchestrator:

id: smart-assistant memory_preset: “episodic” # Just specify the preset! agents: […]

```

Advanced Configuration with Override:

```yaml orchestrator:

id: smart-assistant memory_preset: “semantic” memory_config:

# Preset provides base config, override specific values default_long_term_hours: 720 # Override: extend to 30 days

agents: […]

```

Agent-Specific Memory Types:

```yaml agents:

  • id: knowledge_agent type: memory memory_preset: “semantic” # For storing facts config:

    operation: write

    namespace: knowledge_base

  • id: interaction_logger type: memory memory_preset: “episodic” # For conversation history config:

    operation: write

    namespace: conversations

```

Performance Benefits

  • Simplified Configuration: 90% reduction in memory config complexity

  • Cognitive Alignment: Memory types match human cognitive patterns

  • Optimized Performance: Each preset is tuned for its use case

  • Backward Compatible: Existing configurations continue to work

orka.memory.presets.get_memory_preset(preset_name: str, operation: str | None = None) → Dict[str, Any][source]

Get a memory preset configuration by name, with optional operation-specific defaults.

Parameters:

  • preset_name – Name of the preset (sensory, working, episodic, semantic, procedural, meta)

  • operation – Memory operation type (‘read’ or ‘write’) for operation-specific defaults

Returns:

Dictionary containing the preset configuration, merged with operation-specific defaults

Raises:

ValueError – If preset_name is not found

orka.memory.presets.list_memory_presets() → Dict[str, str][source]

List all available memory presets with their descriptions.

Returns:

Dictionary mapping preset names to descriptions

orka.memory.presets.merge_preset_with_config(preset_name: str, custom_config: Dict[str, Any] | None = None, operation: str | None = None) → Dict[str, Any][source]

Merge a memory preset with custom configuration overrides, including operation-specific defaults.

Parameters:

  • preset_name – Name of the preset to use as base

  • custom_config – Custom configuration to override preset values

  • operation – Memory operation type (‘read’ or ‘write’) for operation-specific defaults

Returns:

Merged configuration dictionary with operation-specific defaults applied

orka.memory.presets.get_operation_defaults(preset_name: str, operation: str) → Dict[str, Any][source]

Get operation-specific defaults for a memory preset.

Parameters:

  • preset_name – Name of the preset

  • operation – Memory operation type (‘read’ or ‘write’)

Returns:

Dictionary containing operation-specific defaults

Raises:

ValueError – If preset_name is not found or operation is invalid

orka.memory.presets.validate_preset_config(config: Dict[str, Any]) → bool[source]

Validate that a memory configuration follows the expected structure.

Parameters:

config – Configuration dictionary to validate

Returns:

True if valid, False otherwise