microservice-survey

A surveys microservice for AMT course 2017-2018

Getting Started

Clone the whole project.

Prerequisites

You first have to install docker and have the last version of IntelliJ IDEA Ultimate (2017.2.5 here).

Deployment

First method (only to launch the API Surveys, no scripts):

Open the docker terminal and go to the topology-amt folder. There is a file called docker-compose.yml.

Enter the line command : docker-compose up --build.

It will build and run all the images with the good configuration (PHPMyAdmin and MySQL). Then open your web browser and and go to the following url if you want to administrate the Data Base: http://192.168.99.100:6060/.

The username is root and the pasword adminpw.
Attention: at this moment, you just have a database that is running and not the server with the API.

To launch this server, you have to launch IntelliJ IDEA, open the project in the folder spring-server that contains the survey API and run it. You will be able to manually test your own requests with this API on the following interface: http://localhost:8080/api/swagger-ui.html#/questions-api-controller.

Then if you want to execute some tests on the api-server, you can open the survey-specs folder that contains all the tests and run it with IntelliJ IDEA by doing a right click on the project and select Run then All tests.

Warnings:

With this method, you need to manually launch the API Users (https://github.com/PestaKit/microservice-users) and be aware of the configuration dependencies (IP addresses and port mapping) what can be a little bit tricky. You will probably change some configurations to get access to the server with the mentioned link. This is due to the fact we started to implement our API without the API Users dependency. Our API Surveys depends now on the API Users for authentication. If there is no authentication process, you won't be able to use the API Surveys (authorization problem). Be sure the API Users is running when you launch the API Surveys.

Second method (to launch full topology with scripts and docker-compose):

With this second method, you are sure to launch to full topology to use both APIs (API Users and Surveys). You will use the script and docker-compose we made during our integration test phase.

That is why you need to put in the same folder: a folder containing our full API Surveys, an other folder with the API Users and finally a folder that contains the integration project you can get here (https://github.com/PestaKit/integration).

This integration project contains a script build-images.sh. Run it and it will build the images you need. Then go to the docker-topology folder and enter the line command : docker-compose up --build. It will create the full topology (be aware of running the script build-images.sh before). Then you can access to the specs of each API with the following links:

https://192.168.99.100:2201/api https://192.168.99.100:2200/api

You are know ready to play with these APIs. But don't forget to check next chapters if you need more details on the endpoints and how to use them.

Endpoints

You can check the different endpoints with Swagger (if you choosed the second method of deployment): https://192.168.99.100:2201/api

Authors

Julien Brêchet, Adrien Marco, Ali Miladi and Dany Tchente

Acknowledgments

Starting template: Olivier Liechti, Miguel Santamaria

API Utilisation warnings and global operation

Here are some rules about the utilisation of the API if you want to use it to develop your web application:

• First off all, you have to be authenticated to use our API. So please, check the API Users (https://github.com/PestaKit/microservice-users) to know how to create a user and get the correct token you will need to show while executing the different actions this API gives to you.

• When you create some questions, there are independent and unique. Then you can link it to a survey that is unique too. So if you want, for example, to send the same survey each week to a group of people, it means you have to create each week a new and unique survey which contains the same linked questions as the previous week.

• If you are interested to check the answers to the questions of a survey, you have to know that the answer is an independent object that is in relation with: the id of the user who give an answer, the id of the survey that contains the question, the id of the answered question, the choice he selected and the time. So if you want to collect the answers for a particular user between different surveys that contain the same questions (like the given example), you have to search for all answers which contain the same user id and question id because you know the surveys id are not the same.

• Do not forget that, every time a user give an answer to a question, an object (type: answer) can be immediately created to save the answer. Because of the time attribute of an answer, a user can change of mind when he is completing a survey. You will be able to know if more than one answer was given to a question (for the same survey and the same user). Then you can, for example, just consider the latest answer by consulting the time attribute.

• The choices belong to a question and are not linked even if we store it in the database. The choice is hard coded in the question. For example, if a choice is called banana, you will find in the database as many choices banana (with different Id's) as we have questions with the banana choice.

• A survey does not implement the notion of open/close. It means that a survey is accessible on every time. We can't open or close a survey.

• Each choice of a response has its position. This position starts with the index 1 and must be incremental.

• When you create two identical answers, only one is considered.

Cucumber tests

If we want to test our microservice Survey with Cucumber, we need to comment a few lines to ignore the integration of the microservice User. We need to do that because we test only functionalities of our service alone.

We need to comment @PreAuthorize("hasRole('User')") annotations in our 3 API controllers which are SurveysApiController, QuestionsApiController and AnswersApiController. We need to comment importations of springframework.security and pestakit.users too.

Once we've done those changes, we can make our Cucumber tests without Build errors.