Hi Bruce,
is it possible to print out a space before a substitute in the substitutes.txt file?? eg.
cuarenta _cuarenta
I did this above and it actually works, but when building CS the engine display a warning msg saying something like “subtitute contains _” is it ok??

I got to do this…

uno_mas _+1+mas     >>> “uno mas” means “one more”

cuz when user says “tengo 19 uno mas” the engine is detecting it as   +-1 +19-    here is the   prepare:

Original User Input: tengo 19 uno mas
Tokenized into: tengo 19 uno mas
Substituted (substitutes ) into: tengo 19 1 mas
Actual used input: tengo 19-1(19 uno) mas

Xref: 1:tengo   2:19-1 >3   3:mas
1:tengo   2:19-1   3:mas
badparse Tagged POS 3 words: tengo (MAINVERB Verb_present)  19-1 (Adjective_numb
er)  mas (Pronoun_object)
  MainSentence:  Verb: tengo   PRESENT

Concepts:

1: tengo (raw):  +~verb_present +~verb_bits +~verb +T~pai +T~pab +~verb_conjug
ate1 +~mainverb +tengo //
1:  (canonical):  //

2: 19-1 (raw):  +~adjective +~adjective_number +~number +T~pai +T~pab +~timeba
sedreference +19-1 //
2: 19-1 (canonical):  //  +-1 +19-

3: mas (raw):  +~pronoun +T~pai +T~pab +~pronoun_object +~pronoun_bits +~noun_
nodeterminer +~KINDERGARTEN
. +~sentenceend +mas +~adj-0a //
3: mas (canonical):  //

for most of the numbers I created interjections like
uno 1
dos 2
is there a better way to make the engine detect spanish numbers?

Hi Bruce, please I found a big problem too
is there some kind of limit, in the number of concepts, or number of items inside concepts?
I have 50 000 words dict,
1 concept with 400 words, 3 concepts with 200 words,
around 50 big concepts each one with around 40-60 words,
and 80 concepts each one with around 8-20 words

but recently I have some pattern rule that doesn’t detect my ~concept, as soon as I replace the ~concept with a [word1 word2] it works again,
you once said that concepts used momery or something like that,
am I overloading the use of memory??
why is that concept not working???

Thanks Advanced.

I solved my problem regarded to my last two posts,
I disabled cs_control #DO_NUMBER_MERGE | #DO_DATE_MERGE

Currently I’ve try to implement, this self-reflection script that was in the advanced manual

Could you please Bruce explain a little more about this

I have a topic, that run several other topics,
I would like have a way to know if in some of those topics, the both already said “Yo” [means I], I mean it has not already print I, but it has build it have it prepare to print.
so bot doesn’t repeat itself, eg.

bot should not say >>>> I’m 30, I live in a city, I work at…
instead it should say >>> I’m 30, live in a city, work at

Could you please Bruce tell me how can I evaluate that??

thanks advanced.

Off the top of my head, you could have the topics create a fact of what they want to say. Then in postprocessing you can review the facts, see if they start with I, strip I’s off later ones, and send the whole thing as data to be printed.

thanks Bruce, could you please post the most simple example of such a script?
currently I only use matchs, forward backward matching (u helped me to finally get these ones), $vars and $$tempvars. I know the manual explains but I can’t figure out a way of how exactly apply them. So perhaps you could hint a way, so I can get the grasp. Thanks advanced.

topic1:
u: (what is your job)  I work for a company

topic2:
u: ( do you have a pet) I have a cat

topic: postprocess topic
t:()  $$sentence1 = ^response(0)
    $$tmp = ^extract($$sentence 0 2)
    if ($$tmp == “I “)  # we can try to merge
    {
$$sentence2 = ^response(2)
$$tmp = ^extract($$sentence2 0 2)
if ($$tmp == "I ") # we can merge
{
$$rest = ^extract($$sentence2 2 999)
^reviseoutput(2 $$rest)
}     
    }

ROUGHLY it looks like this, not checked the names of routines or order of arguments. THe idea is to see if later responses stored all start with “I ” or some equivalent, stripping them to the remainder of the sentence and revising that sentence for output

mm pretty obscure to me xD, I will review the fact manual with this example, thanks
I understand that ^extract() is trying to retrieve items form a fact right? which contains 3 elements, I get that you use 2 0 or 999 in the elements kinda in a hurry, can this be done in a loop? like the control script loop that check all user input?

Thanks Advanced.

Hi Bruce,
I was reading Finalizing A Bot Manual…
could you please tell me what is this used for???
TOPIC_NOKEYWORDS
it is only meant to work when testing pattern rules, right? what it is used for? could you please tell me where it have to be placed?
is this ok?
TOPIC: ~FOOD_NOKEYWORDS keep repeat []

also I would like to know if a sample input comment would work if its place like this…
#! Do you like swirl?  ?: ( swirl) I love raspberry swirl
I mean, before the pattern rule, but in the same line, just sepparated by a tab. could it work to, I say so, all my patterns are made that way, cuz its easier to me to read less lines.

thanks advanced.

:verify normally can do a range of tests on every rule. You can disable tests individually per rule, but if you know that the test makes no sense on any rule of the topic, it’s senseless to disable rule by rule
TOPIC_NOBLOCKING   TOPIC_NOPATTERNS TOPIC_NOSAMPLES   TOPIC_NOGAMBITS
tells the topic as a whole not to do that particular :verify behavior
TOPIC: ~FOOD_NOKEYWORDS
should be
TOPIC: ~FOOD TOPIC_NOKEYWORDS
#! Do you like swirl?  ?: ( swirl) I love raspberry swirl
No.  # means comment out the entire line thereafter.
#! Do you like swirl? 
?: ( swirl) I love raspberry swirl
is the correct formula.

Hi Bruce,

Is there a default file that contains the verb word “tener” [have]
cuz despite I have deleted the english DICT, I got this when doing :prepare

eb: > :prepare que tienes q hacer
TokenControl: DO_SUBSTITUTE_SYSTEM DO_PROPERNAME_MERGE DO_SPELLCHECK DO_INTERJEC
TION_SPLITTING DO_PARSE

Original User Input: que tienes q hacer
Tokenized into: que tienes q hacer
Substituted (interjections ) into: que tienes que hacer
Actual used input: que tienes que(q) hacer

Xref: 1:que   2:tienes   3:que   4:hacer
1:que   2:tienes   3:que   4:hacer
badparse Tagged POS 4 words: que (MAINSUBJECT Pronoun_subject)  tienes/tiene (MA
INVERB Verb_present)  que (OBJECT2 Pronoun_subject)  hacer (Verb_present)
  MainSentence: Subj: que   Verb: tienes   PRESENT

Concepts:

1: que (raw):  +~pronoun +T~pca +T~pba +~pronoun_subject +~pronoun_bits +~KIND
ERGARTEN +~mainsubject +que //
1: que (canonical):  //

2: tienes (raw):  +~verb_present +~verb_bits +~verb +T~pca +T~pba +~verb_conju
gate2 +~mainverb +tienes //
2: tiene (canonical):  +tiene +~solid_containers +~containers +~contains +~locka
ble +~functions +~closeable
. +~artifacts +~objects +~nounlist +~human_data +~liquid_containers //

3: que (raw):  +~pronoun +T~pca +T~pba +~pronoun_subject +~pronoun_bits +~KIND
ERGARTEN +~object2 +que //
3: que (canonical):  //

4: hacer (raw):  +~verb_present +~verb_bits +~verb +T~pca +T~pba +~must_be_sep
arate_phrasal_verb +~sentenceend //
4: hacer (canonical):  //

+~solid_containers +~containers +~contains +~lockable +~functions +~closeable??? where can I delete this info??
thanks advanced.

why is happening that, please anybody any idea??

I already solved it,
one question though,
what is the “TOPIC/BUILD0/patternWords0.txt” file used for???
thanks advanced.

when you use words in a pattern that are NOT in the existing dictionary, they are noted here as keywords that will need to be detected, so spell checking is not allowed to change them

I noticed CS has default ~malename n ~femalename concepts,
is there a file where I can modify them? I dindn’t find anything on TOPIC
thanks advanced.

No. They are built into the english dictionary data.