Hello World

Writing an Ema apps involves two things at a minimum:

• a Route type type corresponding to the generated HTML file(s), as well as
• an EmaSite instance on that route type defining the site render pipeline.

The simplest Ema app is presented below:

import Ema

-- The Route type of our site
newtype Route = Route ()
  deriving newtype
    (Show, Eq, Ord, Generic, IsRoute)

-- Site pipeline (input & output)
instance EmaSite Route where
  siteInput _ _ =
    -- There is no input in a hello-world site
    pure $ pure ()
  siteOutput _ _ _ =
    -- The output of index.html is simply a hello-world message in HTML
    pure $ Ema.AssetGenerated Ema.Html "<b>Hello</b>, Ema"

main :: IO ()
main = 
  -- Hook everything up in main using runSite.
  void $ Ema.runSite @Route ()

Let’s walk through this code:

• The Route type represents the pages on our site. As there is only one page (index.html) in our hello-world site, we simply use ().
  • The unit type, (), already has an IsRoute instance, so we derive it via newtype.
  • In Add Routes, you will see how to write more elaborate route types and derive IsRoute for them. IsRoute is what tells Ema that a Haskell type is a route type (with URL encoders and decoders).
• The EmaSite typeclass defines the “site pipeline” – the input Model type and the output Asset:
  • siteOutput renders this route.
  • siteInput returns the model used in rendering the routes. In Add a Model we will use a custom model, and in Dynamic Model we will make it time-varying.
  • Ema.runSite takes a route type (via TypeApplications), and runs the Ema site.
• Running the resultant executable without arguments runs the Live Server, whereas running it with the gen subcommand will generate the static site (see CLI).