Skip to content

Latest commit

 

History

History
128 lines (97 loc) · 5.59 KB

README.md

File metadata and controls

128 lines (97 loc) · 5.59 KB

haskell-webservice

This project contains a web service written in Haskell using Happstack and SQLite. The web service provides an interface to (fictitious, randomly generated) weather readings taken at the North Pole during 2017. It has one endpoint, weather/date/,
the response to which will depend on the request type.

If the webservice is running on localhost on port 8000 and receives a GET or POST request to http://localhost:8000/weather/date/YYYY-mm-dd then it will return a JSON object with the data for that date. If weather data is held for that date, the response code will be 200 and the body will consist of JSON something like this:

[{"temperature":-23.164497,"date":"2017-01-01"}]

If no data exists for the date in question, the response code will be 404 and the body will consist of an empty JSON array.

We can also make PUT requests to insert or update a date/temperature pair in the database. In this case the URL needs to include the temperature. The easiest way to make a PUT request is by using curl:

$ curl -X PUT http://localhost:8000/date/1970-01-01/42
[]
$ curl http://localhost:8000/weather/date/1970-01-01
[{"temperature":42,"date":"1970-01-01"}]

If the date already exists, the temperature is updated. If not, a new record is added to the database. This call will return 200 and an empty JSON array if everything went OK.

Installation

This project relies on a local SQLite database, so you need to have SQLite installed to run the code. On Debian-based Linux systems (including Ubuntu) you can do this with the command sudo apt install sqlite3. These instructions cover other operating systems.

Set up the project in the usual way:

$ git clone https://github.com/jimburton/haskell-webservice
$ cd haskell-webservice
$ cabal configure

Now you can start setting up the application. There is a cabal target to install the database:

$ cabal run DB-setup

The step to build the database only needs to be done once but you can run it any time you want to restore the database to its original state. Now you can start the webservice:

$ cabal run webservice 
Listening for http:// on port 8000

Once the service is running you can interact with it in the browser by going to http://localhost:8000/weather/date/2017-01-01 or in a terminal by using something like curl:

$ curl http://localhost:8000/weather/date/2017-01-01
[{"date":"2017-01-01","temperature":-23.164497}]

While the service is running open a new terminal and run the tests with

cabal run test-webservice

Note that one test passes but three fail. You can get them all working by completing the exercise below.

Exercises

Your job is to extend the webservice in various ways. Before doing so you should read the first two chapters of the HappStack Book. You may also need to look at the docs for HappStack and sqlite-simple, the library we are using to interact with the database. As you work on the methods, keep running the tests and calling the webservice with curl or a browser to see what comes back.

  1. Add a new endpoint to the webservice, weather/range/d1/d2, where d1 and d2 are dates in the format YYYY-mm-dd. When this endpoint receives a GET request it should return all records in the database that fall between d1 and d2 (i.e. greater than or equal to d1 and less than or equal to d2) as an array of JSON objects.

    Start by changing the main function by adding a new route dirs "weather/range" ... to the list it contains. Copy one of the existing routes for the format. This route needs to call a new function that you will define in WeatherService.Service. Call it rangeHandler. Again, you can copy and adapt one of the existing functions such as dayHandler or dayPutHandler to get this working. Test your function in the browser and/or a terminal.

  2. Add a new endpoint to the webservice, weather/max/d1/d2, where d1 and d2 are dates in the format YYYY-mm-dd. When this endpoint receives a GET request it should return the details of the day with the maximum temperature between d1 and d2 (i.e. greater than or equal to d1 and less than or equal to d2) as an array containing a single JSON object.

  3. Add a new endpoint to the webservice, weather/above/t, where t is a signed floating point number. When this endpoint receives a GET request it should return all records in the database where the temperature is greater than or equal to t as an array of JSON objects. Other request methods should result in a 405 response code ("Method not allowed") and an empty JSON object.

  4. Convert the web service to use the web-routes library as described in chapter 7 of the HappStack Book.

Compare your work to mine in the solutions branch. There is a version of the webservice that uses web-routes in src/Main.hs and a version that doesn't in src/Main-NoRoutes.hs.