Operator precedence sensitive layout

Consider the following piece of code as formatted by britanny:

dP =
  string "ne"
    $>  NE
    <|> string "nw"
    $>  NW
    <|> string "n"
    $>  N
    <|> string "se"
    $>  SE
    <|> string "sw"
    $>  SW
    <|> string "s"
    $>  S

This doesn't reflect the precedence of the operators involved very well and also makes the meaning less clear.

I would expect it to be formatted as such:

dP =
  string "ne" $>  NE
  <|> string "nw" $>  NW
  <|> string "n" $>  N
  <|> string "se" $>  SE
  <|> string "sw" $>  SW
  <|> string "s" $>  S

And I am sure there are other cases where long expressions with different operators could be layed out better if precedence was taken into account.

Is this something brittany can do?

Start by reproducing the Haskell example in the issue and locate brittany's layout handling for long expressions with mixed operators. Compare the current output with the expected grouping, then verify the formatter preserves operator precedence across additional mixed-precedence cases.