Bug Stomp

I am having trouble with certain one word keyphrases like ouch. I have used the regex ^ouch$ (re), and ouch (re) with a rank of 65. but it keeps going to "blab" and bringing up an xnone???

Are you using raw mode? If not, that may be the problem. If that doesn't fix it, you could try the blunderbuss approach:

[o][u][c][h] (re) (also in raw mode)

I tried [o][u][c][h] (re) raw rank of 65 and got this still blab.
Message: 'ouch' Time: 7.22
Message: (spell-corrected) 'ouch' Time: 7.23
Message: (preprocessed) 'ouch' Time: 7.23
Name Poss: Bildgesmythe (Name As Word: )
Begin Sentence Total Time: 7.24
This Sentence: ouch
This Phrase: ouch
Begin Interj Total Time: 7.24
Found: ouch
End Interj Total Time: 7.25
New Sentence: ''
Sentences:
Emotion: -1 Amp: 0 Hello? 0 Goodbye? 0 Yes? 0 No? 0 Haha? 0
ClusterTracker:
Message Info
Emotion: -1 Amp: 0 FoundHello: 0 FoundGoodbye: 0 FoundYes: 0 FoundNo: 0 FoundLaugh: 0
Sentences:
Emote?:

pre-GetKnown: 7.25
end-GetKnown: 7.25
Keyphrase Match:
Blank Message. Ending.

Emotional Analysis:

Favorites?:



What Noun?:


Considered BLAB (1).
Random Memory Number: 5 of 28
Memory Chance: 27 GossipTopic: 0 MemoryLevel: 1
No Match. (use xnone)

well, that is weird. I can't get it to recognize ([ou]+)[c]([h]+) (re), ([ou]+)[c]([h]+)([ !]+)$ (re), or even ([ouch]+) (re) in raw mode and rank 127! It ought to work, but I have no idea why it doesn't

I have upped the rank to 75 and tried
([o]+)([uch]+)([ !]+) (re)
^([ouch]+) (re)
all in raw mode still get
Considered BLAB (1)

Darn, because having it as a keyphrase would be much better than going to an xnone. I would at least like to offer them an aspirin or something.

Is anyone else having problems with ranks? For example, I had two keyphrases that use to work. 'I * think' rank 0, 'I do not' rank 15. I do not think used to come up under the I do not, now it's coming up under the I * think. The thought of having to change all the ranks gives me the chills.

Prob123, I have noticed a lot of goofiness, and I can't say it's just dropping seeks or ranks or whatnot. I would not go in and fix anything like that until we have the new server and can import. That's what I am waiting for, even though a few of my bots could really used some more work.

I've found that one-word responses are almost always treated as "blab" except for things in the "yes" equivalence class. So "hello!" will be treated as blab, but "Hello there!" will match a raw keyphrase for hello. Same with Shalom. Plain old "Shalom" is blab. "Shalom to you" will register with a "shalom" keyphrase.

I suspect the only fix is to get the Professor to tweak the AI Engine's blab-setting. *shrugs*

I was having a similar regex issue, and it was with a one word response as well - so thanks, prob and Ulrike, for your posts confirming my suspicions.

?? What is blab?

I guess it is a list of words that the AI engine just lables blab in debug and picks xnones for an answer, no matter what you may have for a response.