The Real Meta - Java Backend

Website: https://jumpingrock.github.io/League-Analyzer-Java-Backend/

Environment Details

Java 8
Maven 3.3.9
Spring 1.5.2

Included from spring

Actuator + docs
Cloud connectors + Cloud Security (need to make sure compatible with platform)
MongoDB
REST Docs
Security
Web (For embedded Tomcat and Spring MVC)
To add to our spring tool set: Swagger

External Maven Dependencies

com.github.rithms | riot-api-java Java Wrapper For API | Docs

Project Structure

com.meta.analyzer
    ├── Application.java		<= Startup Spring Boot
    ├── ApplicationProperties.java
    │ 
    ├── .riot.api.aggregator		<= Main controller for riot api calls 
    ├── .riot.api.grabber		<= Seperate method implementation for Riot Api calls
    │	 	├── MatchHistory	<= Get all info concering match history
    │	 	├── MatchItems		<= Get items based on overall match history, individual match, etc.
    
    com.meta.analyzer.rest.aws.*	<= Packages that use REST calls, Queries done through URLs
    com.meta.analyzer.jest.*		<= Packages that use JEST client, Queries created through ElasticSearch Java API

Flow of Application

REST Controllers -- getBuilds => Takes in summoner name and returns aggregated results from database. -- queueSummonerName => Take in summoner name. Add or update it to RDS to queue up aggregation.

Code Improvements and Refactoring Needed

Champion Match History in MatchHistory is a map write now, probably better to do 2D ArrayList.

In progress part of README

Your Most Popular Build

Input: Summoner Name

(DONE) SummonerName -> SUMMONER-V3 /lol/summoner/v3/summoners/by-name/{summonerName} -> account id

(DONE) account id -> MATCH-V3 /lol/match/v3/matchlists/by-account/{accountId} -> List of gameIds

(DONE) gameIds -> MATCH-V3 /lol/match/v3/matches/{matchId} -> Map <ChampionId, gameIds>

Map <ChampionId, gameIds> -> TODO -> Map <ChampionId, Items>

-> Aggregate ChampionId[item1%, item2%, ..., itemn%] -> Display list of champions and their item%s Output: Previously played champions and your item pick %

Putting to ES /* * Index/Type/Id | Summoners/summoner/(Id) <= Id has to be unique for queries * Entire Document: Summoner Name - ChampionId - MatchId -Items */

Extra Resources

AWS resources

Application for Riot API prod key

Early Goal: Display a summoner's preferred build on different champions. Early Goal Ex. Voyboy builds protobelt on several champions, some unpopular, like akali. Search Voyboy's name and we can display his preferred build on akali. Also players can find their own builds and compare what they do differently to pros. Extra Goal: Display builds people win with that contain unpopular items. Extra Goal Ex. An irelia player wins consistently with tear and phantom dancer. Display this build to users. With enough aggregated data we can show highest win % off meta builds for as many champions as possible. Extra Notes: I separated the app. into an Angular2 node express front end and Java/Database back end. I have several reasons why, but the point is there will be two project URLs. https://github.com/JumpingRock/LeagueAnalyzer-Angular2 https://github.com/JumpingRock/League-Analyzer-Java-Backend I am still early into development, so the projects will be skeletons of what I am up to, I have yet to choose a database even.

(Please recommend me one if you want!) Below is not relevant to the project, but I thought work mentioning: For the extra goal, Is there anyway I could work with someone to get a list of active ranked played below challenger/master? Otherwise my only idea is to store all the summoner names of people that use my application, and I can only display their builds. We worked with a list of users we could DISCOVER the hidden off-meta builds. Finally, any advice is welcome on architecture of the application! If you need clarification on anything please let me know!

Christopher-Hsieh/League-Analyzer-Java-Backend