Hi all,

I’m interested in doing something like this:

a: (_1?~noun)

if ($tags_level >= 0) {
^createfact(^join(user__ _1) is_a noun)
}

b: ($tags_level >= 1 _1?~noun_singular)
    ^createfact(^join(user__ _1) is_a noun_singular)
b: ($tags_level >= 1 _1?~noun_plural)
    ^createfact(^join(user__ _1) is_a noun_plural)

The goal is to save some of the POS tags and other semantic tags in user facts for later use.
But, I want to be able to control the semantic \ grammatical deepness as in the example above.

So, the first question is technically, how to write multiple conditions to determine if the rule fail or not. I don’t seems to be able to find a correct way of doing so.

The second is more theoretically, what’s the best way to divide the tags into levels, so for example, tags such as:
level 0: basic pos tags (noun, proper noun)
level 1: more specific pos categories (plural, singular)
level 2: sentence structure tags (subject, object)
level x: attributes such as: ~tool…
level y: attributes such as: ~enterable \ ~immovable
and so on.

I’m looking for a simple way to determine this, maybe wordnet has some tricks for it or the ChatScript mechanism already works with some groups of marks / tags I can use.

Thanks in advance,

Sam

well, not sure what your _1 in first rule comes from.
u: (_~noun)  would get you the first noun.

your b:  cannot have spaces in a comparison so must be $tags_level>=1   as a single token but could have been this
  b: ($tags_level>=1) refine()
    c: (_1?~noun_singular)  createfact
    c: (_1?~noun_plural) createfact

to do levels you could do variations of this:
u: (_*1) refine()
  a: (_0?~noun)  save a noun fact
    b: (_0?~noun_singular) save singular
    b: ...plural
  a: (_0?~verb)  save a verb fact
....

and another time not for part of speech but for things like phrases or mainsubject or whatever

So I understand it is not possible to enter multiple conditions in the same rule.

Another question about this pattern:
a: (_0?~verb)

how to do negative conditions (not member of verb)?

Thanks

I’m not sure what you mean… multiple conditions in the same rule are possible and common. give me an example of what you want to do.

as fo negation, it is
a: (_0!?~verb)
or
a: (!_0?~verb)

In general, I need a way to check for membership of input in a concept only if some variable holds specific value or is les than or greater than some value.

Another example of multiple conditions,
Assuming I want to distinguish between common noun and proper noun:

# noun AND not proper noun
a: (_0?~noun *AND* _0!?~noun_proper_singular)  # accepts *city* or *cities* but not *london*

# noun AND not proper noun
a: (_0?~noun_singular *OR* _0?~noun_plural)  # accepts *city* or *cities* but not *london*

I’m not sure how to implement OR and AND operators. I’m looking for the right way to extract information about the input.

In addition, is this way:
a: (_0?~noun)

Better than this way:
$input = _0
a: ($input?~noun)

in terms of performance?

So in summary, I think I need a reference to a list of conditional operations on concepts and input and some examples.

Thank you very much,

Sam

u:( $myvar<5 _~noun)  check for noun only if $myvar < 5

AND is automatic with multiple conditions…
a: (_0?noun _0!?~noun_proper_singular )
HOWEVER, in this case one does not need the first thing because a noun_singular is both a noun and not a proper noun.
a: (_0?~noun_singular)
OR is done with [ ]
a: ( [_0?~noun_singular _0?~noun_plural])

in terms of performance, there is only expense in assigning to a $var before testing.

and we have
u: ( << _~mainsubject _~mainverb _~mainobject >>)
for capturing a standard sentence in any order.

Thanks Bruce for the detailed reply. this example is what I needed.

I was wondering. for marks like ~mainsubject that being assigned to word only because of the context of the word.
When I assign the input to a variables like $word = _0. what is really the content of the variable? A string representing the word or object representing the word in the context it were found in.

This is why I’ve asked about performance difference between ($word?~concept) and (_0?~concept).

Thanks for your patient,

Sam

An _variable has both the original and canonical forms and where in the input sentence it was.
When you assign to a $variable, you pick which form (original or canonical) to save and you get that text word, no positional data.  The _variable is just 2 forms of text and an integer index. So in use, it is just the same as using a $var (just text) because you name whether it is the original or the canonical text you are using.  Ergo, storing in a $var before use just slows you down by the $var assignment cost.