hzoo/geturf

Fetches URF data.

GetUrf

For the Riot Games API Challenge. Fetch URF match data and send to a database.

Data Flow

1. Get 5 minute timestamps starting from 1427866500 (for NA).
2. Send timestamps to the endpoint api-challenge-v4.1 to get a list of URF match IDs.
3. For each match ID, get match data from the endpoint match-v2.2.
4. Transform result (remove unused fields, etc). 4. Create tables for matches, players, bans.
5. Save to the database (firebase -> local -> server db).
6. Figure out queries to send to site + Visualize (https://github.com/hzoo/urf).

Setup

• io.js: nvm install 1.6.3 && nvm use 1.6.3
• ES6 with Babel
• MySQL 5.6 (local or server): apt-get install mysql-server
  • Setup a database and name it what you will put in DB_NAME.
  • Add all the tables from ./data/create-tables.js

Create a json file at ./data/config.json.

{
    "LOL_REGION": "LOL-REGION-HERE", // ex: na
    "LOL_TOKEN": "LOL-API-TOKEN-HERE",
    "DB_HOST": "DB-HOST-HERE", // localhost unless you use a service
    "DB_USER": "DB-USER-HERE",
    "DB_PASSWORD": "DB-PASSWORD-HERE",
    "DB_NAME": "DB-DATABASE-NAME",
    "SSL_CA": "blah.pem",
    "SSL_CERT": "blah.pem",
    "SSL_KEY": "blah.pem"
}

run npm start

Structure

• npm start runs fetch/index.js
  • This sets up the database and lolapi wrapper
• fetch/fetch-urf-game-ids.js is run
  • This specificially fetches the URF matchIds by computing each 5 minute time interval (just add 300 to the timestamp) (starting with some initial value) and continuing to send arrays of matchIds to the database.
• fetch/fetch-urf-match-data.js is run
  • This fetches the match data and removes data not used and modifies it so the data can be put in the different database tables (matches, players (each game), and bans.
• This was able to continuously run until all the matches were fetched in a region and then I would change the LOL_REGION config.

Misc

• Running on Heroku (It's fetches different regions by changing the LOL_REGION config var).
• Database is using Cloud SQL (tested locally).
• It ended up being a lot easier/faster to just work with things I've used before or have some experience with rather than trying to use something new.