Ecosyste.ms: Awesome

An open API service indexing awesome lists of open source software.

Awesome Lists | Featured Topics | Projects

https://github.com/mplewis/battlepets


https://github.com/mplewis/battlepets

Last synced: 2 months ago
JSON representation

Awesome Lists containing this project

README 

        
# BattlePets



## Matt Lewis, for Wunder Capital's code challenge



BattlePets is an online battle arena for adorable animals called pets.



## Pets



Pets have four attributes:



* Strength

* Agility

* Wit

* Perception



Trainers can recruit pets from the wilderness. Pets come from the wilderness with names they pick themselves and stats they've had since birth.



## Training



Once recruited, the pets' stats can be trained using the following [regimens](config/training.yml):



* Sprinting: Agility ++++, Perception +

* Lifting: Strength ++++

* Hunting: Perception ++, Wit +++



For example, training in Sprinting will increase Agility greatly and increase Perception slightly.



## Contests



Once trained, pets can battle in the arena in three different types of [contests](config/contest.yml). Each contest tests different sets of attributes:



* Power: Strength •••, Agility •

* Evasion: Wit •••, Perception ••, Agility •

* Speed: Agility •••, Perception •



For example, if two pets enter a Power contest, Strength is the attribute that has the most influence on who wins, but Agility also counts toward the final score.



## Variability



Pets aren't perfect - they're living digital organisms. Some days they train more effectively, and some days they get nothing done at all. Their contest performance varies too - a pet with 6 strength may lose to a pet with 4 strength if it does poorly and the competitor does really well.



## Experience



Pets gain experience as they compete. Every contest win is worth 100 experience and every loss is worth 25 experience. Experience doesn't translate into ability or stats – it simply indicates how much a pet has competed.



# Asynchronous Matches



A match of a contest takes a while. Each pet has to be evaluated on its abilities, and the judges are lazy. When a match is kicked off, it won't complete right away - the pets are tested one at a time while the judges evaluate them. Each evaluation can take anywhere from 0 to 5 seconds.



A contest is made up of performances from different pets. A contest is complete (`contest.complete == true`) when all of its performances are complete (`performance.complete == true`).



For convenience, a `progress` attribute is present on Match object that indicates the current count of complete and incomplete performances.



Async match evaluation lives in [app/jobs/pet_evaluator.rb](app/jobs/pet_evaluator.rb).



# Get Started



This app runs on Postgres. It runs fine in dev mode. It's running on Heroku, but the worker process isn't yet online because it costs money.



## Development



```sh

# Install gems

bundle install



# Set up your DB in the usual Rails way

rake db:create db:migrate db:seed



# After seeding the DB with pets, trainings, and matches, we need to start an

# async worker to run the match evaluations



# DO ONE AND ONLY ONE OF THESE:

# Start an asynchronous worker to handle contest scoring

bin/delayed_job start

# OR, start a few of them to run jobs in parallel

bin/delayed_job start -n 10

# OR, start just one in debug mode

rake jobs:work



# Start the Passenger server on port 3000

bundle exec passenger start -p 3000 --max-pool-size 3

```



Finally, visit http://localhost:3000. You should see an empty leaderboard.



Check out the REST API below to see how to work with the app. You'll probably want to start by creating a few pets and training them. Once that's done, you can start a contest!



## Production



If you put this app on Heroku, you'll need to set some environment variables for production mode. Here's an easy way to do that:



```sh

heroku config:set SECRET_TOKEN=`rake db:secret` SECRET_KEY_BASE=`rake db:secret`

```



# REST API



See [api_docs.py](api_docs.py) for runnable examples of the important REST calls:



```

Leaderboard - top ten pets, sorted by win count

    GET https://battlepets-mplewis.herokuapp.com/



Create New Pet

    POST https://battlepets-mplewis.herokuapp.com/pets



List All Pets

    GET https://battlepets-mplewis.herokuapp.com/pets



Get Info on Pet

    GET https://battlepets-mplewis.herokuapp.com/pets/1



Train Pet

    POST https://battlepets-mplewis.herokuapp.com/pets/1/train

    Params: training type



Start a Match

    POST https://battlepets-mplewis.herokuapp.com/matches

    Params: match type, pets competing



Get Match Results

    GET https://battlepets-mplewis.herokuapp.com/matches/1

```



# Testing



Test coverage is minimal, but the goal is to cover the important stuff – business logic that can make or break the app. Here's what's covered and what needs to be:



- [x] Pets can be trained and their stats increase

- [ ] Contest scoring allows variability in pet performance

- [ ] Pet training increases stats with some variability

- [ ] Pet training should never decrease pet stats



# Tech Debt



* The JSON views are kind of a mess. I'd like it if Match only showed winner and performance info when all performances are complete.

* It would be really nice to make the JSON views consistent. The Active Model Serializer works pretty well but misses some edge cases in associations.

* I'd like to break out the leaderboard into an Active Model Serializer for consistency.

* This app is SO SLOW. There are about a million places to optimize DB accesses.

* It's kind of gross to create a Pet's attributes in the model before_create callback.

* There's no checking for null attributes anywhere or enforcing that attributes are non-null. If you make an attribute null with PUT, you'll probably crash the app.

* PetTrainer and MatchRunner are inconsistent re: which class is responsible for reading the YAML. This should be refactored so they both look the same.