[yotamarker]

https://www.yotamarker.com/t345-disoulv3-even-better-pet-waifu-convo-skills#975

I got the idea from my pet parrot. he would just love to chat and you could tell he understood the gist of things.
this skill takes the concept up a few notches tbh ngl tbph tbqh tbf tbpqh tbh tbh

but wait there is more, lets beef it up with a complimentary skill :

https://www.yotamarker.com/t344-diprefer-convo-beef-up-skill

[infurl]

I think it's awesome that you have a pet parrot. I always wanted to have a parrot pal.

Anyway I looked through your code just now and noticed that you are hard-coding lots of words into it for various purposes. Have you considered implementing some kind of centralized lexicon manager or database? It would encourage you to develop abstractions which in turn make your code more generalizable and powerful. Sure it's harder at first but if you persevere it gets easier.

[yotamarker]

I think it's awesome that you have a pet parrot. I always wanted to have a parrot pal.

Anyway I looked through your code just now and noticed that you are hard-coding lots of words into it for various purposes. Have you considered implementing some kind of centralized lexicon manager or database? It would encourage you to develop abstractions which in turn make your code more generalizable and powerful. Sure it's harder at first but if you persevere it gets easier.

no, he died over a decade ago. this is more of a tribute to him.
yes much of the code is hard coded, because I view some of the human skills to be hard coded.

for example : good morning. what is the time ?
I could mount it with a translator but the gists remain the same. I find most people to react the same to the same input or problems.
and I find most conversations with people to be like reruns. only the wordings change, and in some skills what changes are variables like the item, place,
and related dangers.
having said that, it would make her "feel" more human.

therefore, centralized lexicon manager or database is a based idea so I would love for you to elaborate more on it, and just pour what ever comes to your mind regarding this concept. also reading materials would also be nice.

[infurl]

therefore, centralized lexicon manager or database is a based idea so I would love for you to elaborate more on it, and just pour what ever comes to your mind regarding this concept. also reading materials would also be nice.

The basic idea is to separate out everything that your software "knows" about words and language into a shared system so you can reuse it easily. That system could be as simple as a library that you link and call when your code needs some information about a word e.g. that 'wonderful' denotes 'happiness'. Much better would be a database which provides a lot more power for accessing and reusing the information via SQL.

Code

CREATE TABLE words(spelling TEXT,category TEXT,emotion TEXT);
INSERT INTO words VALUES ('wonderful','adjective','happy');

Then your program could get a list of words that match just what it needs at the time. You can keep adding to the database independently of the software that uses it and your software gets smarter automatically.

Incidentally, you're not just limited to treating the data as data either. You can use it for generating code automatically too. All this takes more time to set up but in the long run it saves a lot of time, and let's face it, these are such huge projects that we're working on that anything you can do to make development more efficient is going to pay dividends.

As far as reading material is concerned, the state of the art is currently Combinatory Categorial Grammar (CCG).

https://yoavartzi.com/tutorial/

http://groups.inf.ed.ac.uk/ccg/index.html