ONA code documentation
Copyright 2022 The OpenNARS authors.

Modules:
0: CONCEPT
1: CONFIG
2: CYCLE
3: DECISION
4: HASHTABLE
5: IMPLICATION
6: INFERENCE
7: INVERTEDATOMINDEX
8: MEMORY
9: NAL
10: NAR
11: NARSESE
12: OCCURRENCETIMEINDEX
13: PRIORITYQUEUE
14: RULETABLE
15: SHELL
16: STACK
17: STAMP
18: STATS
19: TABLE
20: TERM
21: TRUTH
22: USAGE
23: VARIABLE
24: GLOBALS

Module: CONCEPT

// Concept //

//A concept represents a container which allows the reasoner

//1. To remember events that already happened

//2. Learn under what circumstances they do occur.

//3. To revise the statements encoded in them

//Various design decisions:

//- Differently than in OpenNARS, only statements can form concepts.

//- Concepts store a belief event, predicted event, and eternal version

//- Concepts store a goal event but no eternal goal

//- Concepts hold a priority value to encode their attention

//- Concepts hold a usefulness value to encode their forgetting

//Data structure//

typedef struct {
long id;
Usage usage;
Term term;
Event belief; //the highest confident eternal belief
Event belief_spike;
Event predicted_belief;
Event goal_spike;
Table precondition_beliefs[OPERATIONS_MAX+1];
double priority;
long processID; //avoids duplicate processing
long processID2; //avoids duplicate processing
long processID3; //avoids duplicate processing
long lastSelectionTime; //for selective temporal compounding
} Concept;

Module: CONFIG

/*-------------------------*/
/* Anticipation parameters */
/*-------------------------*/

//Truth expectation needed for anticipation
#define ANTICIPATION_THRESHOLD_INITIAL 0.501

//Confidence of anticipation failures
#define ANTICIPATION_CONFIDENCE_INITIAL 0.01

//Anticipate for concrete yet unexperienced outcomes derived from generals
#define ANTICIPATE_FOR_NOT_EXISTING_SPECIFIC_TEMPORAL_IMPLICATION true
/*---------------------*/
/* Decision parameters */
/*---------------------*/

//Truth expectation to move on to next component goal in sequence
#define CONDITION_THRESHOLD_INITIAL 0.501

//Desire expectation needed for executions
#define DECISION_THRESHOLD_INITIAL 0.501

//Motor babbling chance
#define MOTOR_BABBLING_CHANCE_INITIAL 0.2

//Decisions above the following threshold will suppress babbling actions
#define MOTOR_BABBLING_SUPPRESSION_THRESHOLD 0.55

//Whether temporal non-procedural implications are allowed to derive subgoals
#define NOP_SUBGOALING true

//Subsumption confidence threshold above which a specific hypothesis inhibits more generals
#define SUBSUMPTION_CONFIDENCE_THRESHOLD 0.05

//Subsumption confidence threshold below which a specific hypothesis inhibits more generals
#define SUBSUMPTION_FREQUENCY_THRESHOLD 0.5

//How long goal events describing bad outcomes are considered in decision making
#define NEG_GOAL_AGE_MAX EVENT_BELIEF_DISTANCE
/*----------------------*/
/* Attention parameters */
/*----------------------*/

//Event selections per cycle for inference
#define BELIEF_EVENT_SELECTIONS 1

//Goal event selections per cycle for inference
#define GOAL_EVENT_SELECTIONS 1

//Event priority decay of events per cycle
#define EVENT_DURABILITY 0.9999

//Concept priority decay of events per cycle
#define CONCEPT_DURABILITY 0.9

//Minimum priority to accept events
#define MIN_PRIORITY 0

//Occurrence time distance in which case event belief is preferred over eternal
#define EVENT_BELIEF_DISTANCE 20

//Amount of belief concepts to select to be matched to the selected event
#define BELIEF_CONCEPT_MATCH_TARGET 80

//Adaptation speed of the concept priority threshold to meet the match target
#define CONCEPT_THRESHOLD_ADAPTATION 0.000001

//Usage boost for input
#define ETERNAL_INPUT_USAGE_BOOST 1000000

//Unification depth, 2^(n+1)-1, n=2 levels lead to value 7
#define UNIFICATION_DEPTH 31

//Questions concept activation priority
#define QUESTION_PRIMING_INITIAL 0.1
/*---------------------------------*/
/* Temporal compounding parameters */
/*---------------------------------*/

//Maximum length of sequences
#define MAX_SEQUENCE_LEN 2

//Maximum compound op length
#define MAX_COMPOUND_OP_LEN 1

//Max. occurrence time distance between precondition and consequence
#define PRECONDITION_CONSEQUENCE_DISTANCE EVENT_BELIEF_DISTANCE

//Occurrence time distance to now to still correlate an outcome
#define CORRELATE_OUTCOME_RECENCY EVENT_BELIEF_DISTANCE

//Maximum time difference to form sequence between events
#define MAX_SEQUENCE_TIMEDIFF EVENT_BELIEF_DISTANCE

//Allow events which have not been selected to become preconditions
#define ALLOW_NOT_SELECTED_PRECONDITIONS_CONDITIONING false
/*------------------*/
/* Space parameters */
/*------------------*/

//Maximum amount of concepts
#define CONCEPTS_MAX 16384

//Amount of buckets for concept hashmap
#define CONCEPTS_HASHTABLE_BUCKETS CONCEPTS_MAX

//Maximum amount of belief events attention buffer holds
#define CYCLING_BELIEF_EVENTS_MAX 40

//Maximum amount of goal events attention buffer holds
#define CYCLING_GOAL_EVENTS_MAX 400

//Maximum amount of operations which can be registered
#define OPERATIONS_MAX 10

//Maximum amount of arguments an operation can babble
#define OPERATIONS_BABBLE_ARGS_MAX 10

//Maximum size of the stamp in terms of evidental base id's
#define STAMP_SIZE 10

//Maximum Implication table size
#define TABLE_SIZE 20

//Maximum compound term size
#define COMPOUND_TERM_SIZE_MAX 64

//Max. amount of atomic terms, must be <= 2^(sizeof(Atom)*8)
#define ATOMS_MAX 65536

//Amount of buckets for atoms hashmap
#define ATOMS_HASHTABLE_BUCKETS ATOMS_MAX

//The type of an atom
#define Atom unsigned short

//Maximum size of atomic terms in terms of characters
#define ATOMIC_TERM_LEN_MAX 32

//Maximum size of Narsese input in terms of characters
#define NARSESE_LEN_MAX 2148 //ATOMIC_TERM_LEN_MAX * COMPOUND_TERM_SIZE_MAX + 100 for punctuation event marker and TV

//Goal events queue derivation depth layers
#define CYCLING_GOAL_EVENTS_LAYERS 30

//Hashtable bucket size for atom counters in term
#define VAR_INTRO_HASHTABLE_BUCKETS COMPOUND_TERM_SIZE_MAX

//OccurrenceTimeIndex size (large enough to cover all events input and derived within EVENT_BELIEF_DISTANCE from currentTime)
#define OCCURRENCE_TIME_INDEX_SIZE 512
/*------------------*/
/* Truth parameters */
/*------------------*/

//Default frequency for input events
#define NAR_DEFAULT_FREQUENCY 1.0

//Default confidence for input events
#define NAR_DEFAULT_CONFIDENCE 0.9

//Default confidence for analytical premise
#define RELIANCE 0.9

//NAL evidental horizon
#define TRUTH_EVIDENTAL_HORIZON_INITIAL 1.0

//Time distance based projection decay of event truth
#define TRUTH_PROJECTION_DECAY_INITIAL 0.8

//Maximum value for confidence
#define MAX_CONFIDENCE 0.99

//Minimum confidence to accept events
#define MIN_CONFIDENCE 0.01
/*-----------------------*/
/* Derivation parameters */
/*-----------------------*/

//The NAL level of semantic inference
#define SEMANTIC_INFERENCE_NAL_LEVEL 7

//Filter for twice appearing atoms
#define ATOM_APPEARS_TWICE_FILTER true

//Filter for derivations which include nested implications or equivalences
#define NESTED_HOL_STATEMENT_FILTER true

//Filter for inheritance or similarity statement with dependent var
#define INH_OR_SIM_HAS_DEP_VAR_FILTER true

//We don't allow higher-order statements with A> or var2> components
#define HOL_STATEMENT_COMPONENT_HAS_INVALID_INH_OR_SIM_FILTER true

//Whether a higher-order statement is invalid if it contains a inh or sim without var
#define HOL_COMPONENT_NO_VAR_IS_INVALID_FILTER true

//Whether a higher-order statement is invalid if it contains a inh or sim without atomic term
#define HOL_COMPONENT_NO_ATOMIC_IS_INVALID_FILTER true

//Filter disjunction or conjunction in derivation if not right nested
#define JUNCTION_NOT_RIGHT_NESTED_FILTER true

//Variables introduced in set with more than 1 element filter
#define VARS_IN_MULTI_ELEMENT_SETS_FILTER true

//Filtering sub-statement terms with variables and atoms both like (&, $1, a)
#define TERMS_WITH_VARS_AND_ATOMS_FILTER true