Movie application

Business case:

User can get list of movies that are in the cinema.

User can search for any movie that is out there by title and get detailed response about the movie.

Admin can use that to search for movies that could be in the cinema and add movie to the cinema repertoire.

Admin can create a catalog of movies (cinema repertoire) with showTime and price.

Admin can also change showTime and price of the movie.

Prerequisites - environment variables

1. You need an API key. Provide it as an environment variable. http://www.omdbapi.com/apikey.aspx
2. You need credentials for your mongodb instance.

export API_KEY="your_api_key"
export MONGO_HOST="your_host"
export MONGO_USERNAME="your_username"
export MONGO_PASSWORD="your_password"

2. When running application from IntelliJ Idea:
   Go to: /run/edit_configurations/your_configuration/enviroment_variables

API_KEY=your_api_key
MONGO_HOST="your_host"
MONGO_USERNAME="your_username"
MONGO_PASSWORD="your_password"

When running on Heroku and Travis export above environment variables.

Application is available online - wait around 20sek for instance to wake up
https://movie-recruitment-task.herokuapp.com/api/movies?title=fast

REST API documentation

Secured endpoint (online): https://movie-recruitment-task.herokuapp.com/api/swagger-ui.html

Localhost: http://localhost:8080/api/swagger-ui.html

Postman collection of requests: https://www.getpostman.com/collections/3100b0db16003a4a2956

Endpoints:

To make easier for mobile developer there is only one endpoint that contains most of the information needed. Payload is not big so I guess that's good idea.

Online: https://movie-recruitment-task.herokuapp.com/api/movies?title=The%20Fast%20and%20the%20Furious

Endpoint: GET: /movies?title="Fast and the Furious"

{
    "id": "tt0232500",
    "name": "The Fast and the Furious",
    "movieTime": "106 min",
    "description": "Los Angeles police officer Brian O'Conner must decide where his loyalty really lies when he becomes enamored with the street racing world he has been sent undercover to destroy.",
    "releaseDate": "22 Jun 2001",
    "externalRatings": [
        {
            "source": "Internet Movie Database",
            "value": "6.8/10"
        },
        {
            "source": "Rotten Tomatoes",
            "value": "53%"
        },
        {
            "source": "Metacritic",
            "value": "58/100"
        },
        {
            "source": "Imdb",
            "value": "6.8"
        }
    ],
    "internalReviews": [
        {
            "rating": "3",
            "review": "good film",
            "date": "2019-08-11T12:19:31.379Z"
        },
        {
            "rating": "4",
            "review": "bad movie",
            "date": "2019-08-11T12:19:31.380Z"
        }
    ]
}

Endpoint: GET: /catalogs - get all catalogs

[
    {
        "name": "catalog name",
        "movies": [
            {
                "title": "The Fast and the Furious",
                "imdbId": "tt0232500",
                "showTime": {
                    "time": [ 11, 54 ],
                    "date": [ 2019, 11, 25 ]
                },
                "price": {
                    "value": "123",
                    "currency": "PLN"
                }
            }
        ]
    }
]

Secured Endpoint: POST: /catalogs - create new catalog

{
    "name": "catalog name"
}

Secured Endpoint PUT: /catalogs - update or add movies to cinema repertoire

{
    "catalogName": "123",
    "movies": [
        {
            "imdbId": "tt0232500",
            "showTime": {
                "time": [
                    11,
                    54
                ],
                "date": [
                    2019,
                    11,
                    25
                ]
            },
            "price": {
                "value": "123",
                "currency": "PLN"
            }
        }
    ]
}

Build docker image

./gradlew jibDockerBuild

Things done:

Application and requirements

• Merged movie-api response imdbRating and Ratings as one list of ratings
• There is internalRating=(our customer rating and review) and externalRating=(returned from movie-api - only rating)
• You can create multiple catalogs so that it can be sold to other client with other repertoire.

Infrastructure

• Mongodb cluster is from cloud.mongodb.com (Replica set - 3 nodes)
• Two DBS on cluster: prod and test (prod for Heroku, test for local development and tests)
• Simple pipeline - Travis (build - runs tests) -> Heroku (waits for Travis to pass before deploy)
• Travis - there is problem with embedded mongo on travis. Travis and local integration tests are connected to cluster with DBS named test. Generally IT should be self-contained and im fan of in-memory tests, but that solution is good too (good thing is we are testing against real database), but we also depend on some external service which is not good in the same time). Another solution might be docker/docker-compose.
• Heroku - instance sleeps after 30min. Give a moment for instance to start.
• Different MongoDB for each environment. test, prod (test is used as local), (prod is used for Heroku).

Things done (but not clear):

• I assumed that second point from Challenge movie times is meant for movie time meaning Runtime of one movie (not for all Fast & Furious movies - seems pointless to return only runtime of the movie). So I merged second and third point into one.

Things that might/should be done:

• Commits should be done through PR
• API versioning - because that's a first iteration
• Standardize the format of returned movie-api ratings e.g. 5/10
• Create AverageRating for InternalRating - possible computed offline.
• Add pagination for internalReviews (maybe introduce completely new endpoint for it, because extending already existing functionality to paging would make this one endpoint a bit too complex).
• Logout from spring security
• Do more tests of edge cases