test.top

#### This file is used to try Ghost features. 
#### It tries to cover every synatx and basic usage.

# Ghost Rules have the following format
#	(pattern && action) ==> goal
# That is if a pattern is matched action is then executed and the goal is said to be achieved.
# A goal also has an urge to express the level of need to achieve that goal.
# Rules are organized into topics.
# Ghost action selector runs in a loop to select a single rule and execute it.
# Rule is selected based on the following criteria:
#	whether a pattern is matched or not
#	the strength of the goal
#	the urge of the goal
#	importance of the rule (rules just selected are considered less important)
#	whether a rule is in the current topic (priority is given)


### Urge
# Set initial urges for the goal.
# By increasing the urge of a goal, one can increase the chance of rules containing the goal
# to be selected by the action selector.
urge: (please_user=1 introduction=1 curiosity=0.7 stay_awake=1)

### Concept
### Sequence enclosed in parenthesis
# Concepts used to group similar words and phrases together
concept: ~HI [hi hello hey (how be you)]


# Rules are organized into topics.
# The topic definition here applies to all rules below until
# the next topic definition.
# The ff topic has 'keep' feature. It means all rules in this topic can be
# considered by the action selector many times as oppossed to other rules
# which are never reconsidered once they get executed. 
topic: ~GREETING keep [~HI]



# Top level goal shared by all rules under it until the next top level goal.
goal: (stay_awake=0.7)


### Ghost rule with a kind, a label, a pattern, an action and a rule-level goal.

#goal: (introduction=0.6)
?: (what be you) I am a robot

#goal: (introduction=0.7)
?: (what be you name) My name is Dr. Roboto



### Randomized output
u: (~HI) [hi] [hello] [hey] [how are you] there


### Gambits
##goal: (curiosity=0.7)
# t: what is your name?

### Match variable (start with _)
### User variables (start with $)

#goal: (please_user=0.6)
u: (me name be _*) $name= '_0 Hi '_0


?: (do you eat _[meat fish chicken]) No, I hate _0
?: (do you use _*1 or _*1) I neither use _0 nor _1


### Multi-goaled rule
#goal: (please_user=0.8 introduction=0.5)
u: (what be me name $name) Your name is $name


### Rejoinders
u: (you hate human) no. you are very interesting. Do you hate robots?
	a: (yes) I am sure you are going to like me though
	a: (no) We will get along fine

### Choices
?: (you [swim ride fish ]) I do.


### Indefinite wildcards
?: (when * you * home) I go home tomorrow


### Precise wildcards
?: (when *1 you *1 school) I went to school yestearday


### Range restricted wildcards
?: (you *~2 go *~2 office) I often go to the office


### Unordered matching
s: ( << I birds love >> ) I love birds too


### Sentence boundaries
u: (what is an elephant > ) An elephant is a pachyderm
u: ( roses < I like ) I like roses too

### Negation
u: (!hate I *1 meat) Is meat good for you?



### Use of functions in contexts
# The functions are defined in functions.scm and return either (stv 0 1) or (stv 1 1) ==> (falase or true)
# u: ( * ^func-true) This always runs
u: ( * ^func-false) This never runs

### Functions with arguments
u: (are you happy ^is-happy(.7)) Yes I am happy
u: (are you happy ^neg-is-happy(.7)) No I am not happy

### Use of functions in actions
# The functions are defined in functions.scm and return list of word nodes if need be
u: (who killed the _*) I think ^func-findkiller killed the '_0



### Sample rule for face detection
u: (inner msg face is detected) I see your face!


### Sample rule for demonstrating returning and looping C++ detector functions
#goal: (testing_goal=0.5)
 u:(returning detector function) Random returning number is ^func-returning-result

#goal: (testing_goal=0.5)
u: (looping detector function) Random looping number is ^func-looping-result


### Ordered Goal
# Goal of next rules decrease in a geometric sequence with a factor 0.5 
## ordered-goal: (curiosity=0.8)

# u: (what be _*) please define '_0

# u: (why be _*) why not

# and so on.. as there are more rules the geometric sum converges to 0.8