Doesn't seem like there's much going on here lately, but I just stood up a cobe instance as a SlackBot and am very pleased with the efficiency and ease of use of this package. The goal was to bring an old IRC MegaHAL client back to life after migrating to Slack and it's gone swimmingly. Your implementation is actually loads more efficient than the old one, which was dying under the weight of its massive brain. Granted, we're not up to that size yet and we've lost the old logs, but we've got several years' worth of chatter from Slack and it's handling it really well.
I've only got one beef with it at the moment, and it's that the replies seem to be of a nearly uniform length. They seem like an average of 5-8 words. Our old bot used to regularly start rambling for a short paragraph, so it's not feeling quite right yet. Is it a relatively simple matter to add in some variability in reply length? I've looked through the code, but haven't had any luck finding what governs it.
Doesn't seem like there's much going on here lately, but I just stood up a cobe instance as a SlackBot and am very pleased with the efficiency and ease of use of this package. The goal was to bring an old IRC MegaHAL client back to life after migrating to Slack and it's gone swimmingly. Your implementation is actually loads more efficient than the old one, which was dying under the weight of its massive brain. Granted, we're not up to that size yet and we've lost the old logs, but we've got several years' worth of chatter from Slack and it's handling it really well.
I've only got one beef with it at the moment, and it's that the replies seem to be of a nearly uniform length. They seem like an average of 5-8 words. Our old bot used to regularly start rambling for a short paragraph, so it's not feeling quite right yet. Is it a relatively simple matter to add in some variability in reply length? I've looked through the code, but haven't had any luck finding what governs it.