defmodule Gigex.Scraper.Songkick do
  @moduledoc """
  Scraper for songkick

  Songkick hosts concerts and festivals from all over the world.
  This scraper covers all the concerts and festivals in the Berliner area.

  NOTE: For the moment the list of events is very for each day.
  You could have a flood of events on the terminal.
  """

  @default_entries_limit 10
  @songkick "https://www.songkick.com"
  @songkick_berlin_url "#{@songkick}/metro-areas/28443-germany-berlin/"

  alias Gigex.Cache
  alias Gigex.Scraper.HTTP

  def get(opts \\ []) do
    case Cache.get(@songkick_berlin_url) do
      nil ->
        @songkick_berlin_url
        |> HTTP.get()
        |> gigs(opts)
        |> then(&Cache.put(@songkick_berlin_url, &1))

      content ->
        Enum.take(content, Keyword.get(opts, :limit, @default_entries_limit))
    end
  end

  defp gigs(content, opts) do
    limit = Keyword.get(opts, :limit, @default_entries_limit)

    content
    |> Floki.parse_document!()
    |> Floki.find(".event-listings-element")
    |> Enum.reduce_while(
      {[], 0},
      fn
        _event, {entries, ^limit} ->
          {:halt, Enum.reverse(entries)}

        event, {entries, acc} ->
          entry = %{
            name: extract_name(event),
            date: extract_date_from_event(event),
            location: extract_location(event),
            dotw: extract_day_of_the_week(event),
            link: extract_event_link(event),
            infos: extract_event_infos(event),
            datasource: "songkick"
          }

          {:cont, {[entry | entries], acc + 1}}
      end
    )
  end

  defp extract_date_from_event(event) do
    datetime =
      event
      |> Floki.children()
      |> Floki.attribute("datetime")
      |> Floki.text()

    case DateTime.from_iso8601(datetime) do
      {:ok, datetime, _} ->
        datetime
        |> DateTime.to_date()
        |> to_string()

      _ ->
        datetime
    end
  end

  defp extract_location(event) do
    event
    |> Floki.find(".location")
    |> Floki.text()
    |> String.trim()
    |> String.replace(~r/[\n|\s]+/, " ")
  end

  defp extract_name(event) do
    event
    |> Floki.find(".artists")
    |> Floki.text()
  end

  defp extract_day_of_the_week(event) do
    [human_date] = Floki.attribute(event, "title")

    human_date
    |> String.split()
    |> hd()
  end

  defp extract_event_link(event) do
    link =
      event
      |> Floki.find(".event-link")
      |> Floki.attribute("href")
      |> hd()

    "#{@songkick}#{link}"
  end

  defp extract_event_infos(event) do
    event_link = extract_event_link(event)

    event_link
    |> HTTP.get()
    |> Floki.parse_document!()
    |> Floki.find(".additional-details-container")
    |> Floki.text()
    # Trim surrounding blank space, `\n`, `\t`, etc.
    |> String.trim()
  end
end