These are chat archives for yasp-dota/yasp

17th
Nov 2015
Matt Doak
@ClickInspire
Nov 17 2015 07:07

Hey @howardchung, I'm trying to get this set up on Compute Engine with Ubuntu 14.04 but I keep getting caught up at npm run build it seems to be stuck on the steam and dota2 packages saying that it can't find the file to replace within node_modules. Maybe I'm just confused about how to set it up in the first place, but if I'm trying to get skill working alone I assume the process is:

  1. SSH to the new Compute Engine
  2. Clone YASP
  3. Run through the Quickstart process
  4. Run node skill

Any tips?

I had guessed that pm2 was strictly for the web component which I'm not so concerned about running. Just want to see if I can try running some prediction models on the high level match data.
I would love to share my findings to see if it's something that might help YASP.
Although you may already be working on similar things with the data set yourselves.
Howard Chung
@howardchung
Nov 17 2015 08:43
I'm now using pm2 to run all multi-process worker types, but this doesn't include skill.
Not sure what the issue is, maybe it's a permissions thing (don't do things as root)
Also note that skill only adds skill data for pre-existing matches so if your database is empty it won't do anything
Matt Doak
@ClickInspire
Nov 17 2015 16:29
Ah interesting. So reading through it again, it sounds like none of the services actually pull all of the matches from the API--only pulling matches for the players that are registered (or matches that are directly requested) is that correct? Just looking at scanner you only keep match data for those that are registered (https://github.com/yasp-dota/yasp/blob/master/scanner.js#L109) although I would think that if I forced the insertMatch for every match found it should gather the full set of matches from the API.
Howard Chung
@howardchung
Nov 17 2015 17:53
Yes, that's correct