Bug Stomp

Me too. Huge amount of old transcripts I have downloaded it and found no way to remove it from the server.

I'm getting odd responses to the question "What is your favorite color" Generally the answer is "My favorite color is an olive" or "a pastel" or something similar. The problem is the prefixed article. Normally someone would say: "My favorite color is olive." or "My favorite color is orange." The article makes the response sound off. Also including adjectives like pastel gives an poor answer. "My favorite color is pastel blue" makes sense, "My favorite color is pastel" sounds incomplete.

Make your own response and keyphrase rank above 20.

I'm not sure whether this counts as a bug, but in the import language center, the field for the file name isn't large enough for me to see the whole file name. All I can see is the beginning of the path.

It would be nice, as a double check things went well, if the next screen that says "Upload of Import File Complete" in big green letters, could also mention the complete file name somewhere, before I continue to the trial run.

It already tells me the file size. File name shouldn't be too difficult an addition.

I have created a couple of empty plug-ins that I can't edit, delete or replace. I inadvertently tried to add the plug-in while not logged in. The system puts the plug-in's name into the plug-in list, so I can't reuse the name, but does not list me as creator (since I'm not logged in) so I can't edit or delete it. The system should check log-in status before taking any action. Meanwhile, there are a couple of (Eniola_something) plug-ins with no entries and no creator listed.

I am using a regular expression in a seek

Seek: ("^[A-Za-z]$") (re)

When I tried to edit the seek, the seek field in the Language Center turned into just "(" with the rest of the expression missing.

The original line in my file is

+ ("^[A-Za-z]+$") (re) [0] <?PF raw; rem (key1) as only "name" ?>

but when I imported it the '+' was missing from the Regular Expression (']+$' became ']$'), which I assume is why the parser did not recognize 'Samantha' as a valid entry, since it is longer than one character...

While editing a seek in the Language Center I got an error message:
"Keyphrases should not contain any punctuation.
Your entry has been adjusted to fit this."
(seek contained apostrophes "... My name's, my name's,...")
It apparently had no effect on the content of the seek.

Check to see if the apostrophe is still there some times it gets stripped out. Sometimes you have to use the slash apostrophe to get it to work.

I checked, the apostrophe was still there (whew). /' or \'?

well, the backslash went away...

Some forums where precise syntax is discussed have text boxes that are immune from html hijinks -- whatever you type into a text box is duplicated character for character with no changes.

Yeah, dream on girl...

In files you import, you have to duplicate backslashes anyway.

I have as Seek for the name the following line in the files:

+ my name is ([\\\\w\\\\s]+) (re) [0] <?PF raw; ?>

and the assignment of mem-name is in the answers. (Should work with both AIScript instructions in one line too, though.)

(\\w stands for "word character" (letters), and \\s for "whitespace" (space, tab, line feed/carriage return))

Maybe the double quotes make trouble?

When my bots start a conversation, I find their initiating line twice in my transcripts. (The second time with the name of the chat partner included.)

Likewise, when someone else starts a conversation, I don't find their initiating line in my transcripts at all.