├── api // api's only handle requests and responses
|   ├── auth.go
|   ├── error.go
|   ├── index.go
|   └── log_analytics.go
├── ctrls // the controlers handles all the logic and some cases db interactions
|   ├── analytics
|   |   └── log_analytics.go
|   ├── auth
|   |   ├── activate.go
|   |   ├── login.go
|   |   ├── recover.go
|   |   └── register.go
|   ├── user
|   └── test // local test files
├── conf // here we grab al the static values, like credentials, getting env variables, etc.
├── fn // the reason for separating fn in a folder is to avoid circuluar imports
├── public // here goes static and public files to be render
|   └── accountid.go
├── libs // to separte fn based on libraries
├── models // every model created represents a table in mysql db, so db interactions goes here
|   ├── account.go
|   ├── automigrate.go
|   ├── user_info.go // example: this model contains the user profile information
|   └── user.go // this model containse the email, password, token, etc
├── routes // defining routes by group
|   ├── account.go
|   ├── analytics.go
|   ├── auth.go
|   ├── index.go
|   └── user.go
├── test // global test files
├── main.go
└── main_test.go

main.go -> routes (index.go file first) -> api -> ctrls -> models

Run development

go run main.go

we need to have golang installed localy or using a docker container:

but for the docker container we need some more steps, but we have a script to do that

sh start-container.sh, will create the docker image, and run the docker container, with

some extra configurations needed.

Note: to be able to link the api with a db, we need to have a db up and running, at localhost or have a docker container with a db up and running and linked to the golang container, the code is already prepared for both scenarios, we linked the golang container to a container called db, you are going to see the configurations at the db/gorm.go file, and uncomment the startDB() fn in the main.go file. There is a file called models/automigrate.go which generates automatically the tables on the database, just be sure to have a database created and the automigrate will handle the rest.

The api is ready to run the test at src/ctrls/test it will run some golang functions, interacting with a mongodb cluster in the cloud.

There is 2 dockerfiles one for development and the Dockerfile.multi for production env, for the multi file it needs the latest version of docker to be able to create a production ready image.

Install this project on the next folder:

$GOPATH/src/easycast

---

Testing

For testing, i'll suggest to see the ginkgo and gomega libraries, its very similar to mocha and (chai or should) from nodejs.

to run the test you must be located in the path where the test_suite_test.go file is located, and run the ginkgo -v command.

if you want to test a particular function try

ginkgo -cover -v --focus=Bandwidth

Note: i suggest to run the src/ctrls/test/log_analytics_test.go file to see how it works, just be located at the path of the file and run the command: ginkgo -v.

---

Define API URLS

http://jsonapi.org/recommendations/

---

Error Handling

To handle errors in the project, we must first detect the type and level of the functions where the error occurred, the levels are the following:

• 1st Level Fn: this functions are the ones that handle the echo.Context or functions that depends from another function, so that this functions can handle the error, like sending back a message to the user.

e.g.

// for api fn

// in case of error
return sendError("level", msg, err)

• 2nd Level Fn: here are functions that are not dependent, so this functions must return the error, to the dependent functions to be able to handle the error.

e.g.

// functions must follow the following pattern
func superFunc(...params) (type of return, error) {}

// in case of error
return [valid value], err

// if err not exist
return [value], nil

API Response pattern

For endpoint responses, we must follow the next pattern ß

return c.JSON(http.StatusOK, echo.Map{
  [value | values]: ...values,
  "msg": "some message",
  "success": true | false
})

Documentation

To document the api we use RAML, and we follow the RAML specifications, and to generate the html we use a nodejs library called raml2html there is a sh file called bin/generate-doc.sh that executes the command to generate the documentation html file.