[Enhancement] Optionally use the YouTube API for improved fast indexing (#282)

* Started adding youtube API for fast indexing

* Hooked youtube API into fast indexing

* Added youtube_api_key to settings

* Added youtube api key to settings UI

* Added tests

* Refactored the youtube api module

* More refactor

* Changed editing mode name from basic to standard

* [WIP] started on copy changes

* Updated copy

lib/pinchflat/fast_indexing/fast_indexing_helpers.ex

@@ -13,12 +13,13 @@ defmodule Pinchflat.FastIndexing.FastIndexingHelpers do
   alias Pinchflat.Media
   alias Pinchflat.Sources.Source
   alias Pinchflat.FastIndexing.YoutubeRss
+  alias Pinchflat.FastIndexing.YoutubeApi
   alias Pinchflat.Downloading.DownloadingHelpers
 
   alias Pinchflat.YtDlp.Media, as: YtDlpMedia
 
   @doc """
-  Fetches new media IDs from a source's YouTube RSS feed, indexes them, and kicks off downloading
+  Fetches new media IDs for a source from YT's API or RSS, indexes them, and kicks off downloading
   tasks for any pending media items. See comments in `FastIndexingWorker` for more info on the
   order of operations and how this fits into the indexing process.
 
@@ -26,7 +27,7 @@ defmodule Pinchflat.FastIndexing.FastIndexingHelpers do
   downloaded_.
   """
   def kickoff_download_tasks_from_youtube_rss_feed(%Source{} = source) do
-    {:ok, media_ids} = YoutubeRss.get_recent_media_ids_from_rss(source)
+    {:ok, media_ids} = get_recent_media_ids(source)
     existing_media_items = list_media_items_by_media_id_for(source, media_ids)
     new_media_ids = media_ids -- Enum.map(existing_media_items, & &1.media_id)
 
@@ -47,6 +48,17 @@ defmodule Pinchflat.FastIndexing.FastIndexingHelpers do
     Enum.filter(maybe_new_media_items, & &1)
   end
 
+  # If possible, use the YouTube API to fetch media IDs. If that fails, fall back to the RSS feed.
+  # If the YouTube API isn't set up, just use the RSS feed.
+  defp get_recent_media_ids(source) do
+    with true <- YoutubeApi.enabled?(),
+         {:ok, media_ids} <- YoutubeApi.get_recent_media_ids(source) do
+      {:ok, media_ids}
+    else
+      _ -> YoutubeRss.get_recent_media_ids(source)
+    end
+  end
+
   defp list_media_items_by_media_id_for(source, media_ids) do
     MediaQuery.new()
     |> where(^dynamic([mi], ^MediaQuery.for_source(source) and mi.media_id in ^media_ids))

lib/pinchflat/fast_indexing/youtube_api.ex

@@ -0,0 +1,92 @@
+defmodule Pinchflat.FastIndexing.YoutubeApi do
+  @moduledoc """
+  Methods for interacting with the YouTube API for fast indexing
+  """
+
+  require Logger
+
+  alias Pinchflat.Settings
+  alias Pinchflat.Sources.Source
+  alias Pinchflat.Utils.FunctionUtils
+  alias Pinchflat.FastIndexing.YoutubeBehaviour
+
+  @behaviour YoutubeBehaviour
+
+  @doc """
+  Determines if the YouTube API is enabled for fast indexing by checking
+  if the user has an API key set
+
+  Returns boolean()
+  """
+  @impl YoutubeBehaviour
+  def enabled?(), do: is_binary(api_key())
+
+  @doc """
+  Fetches the recent media IDs from the YouTube API for a given source.
+
+  Returns {:ok, [binary()]} | {:error, binary()}
+  """
+  @impl YoutubeBehaviour
+  def get_recent_media_ids(%Source{} = source) do
+    api_response =
+      source
+      |> determine_playlist_id()
+      |> do_api_request()
+
+    case api_response do
+      {:ok, parsed_json} -> get_media_ids_from_response(parsed_json)
+      {:error, reason} -> {:error, reason}
+    end
+  end
+
+  # The UC prefix is for channels which won't work with this API endpoint. Swapping
+  # the prefix to UU will get us the playlist that represents the channel's uploads
+  defp determine_playlist_id(%{collection_id: c_id}) do
+    String.replace_prefix(c_id, "UC", "UU")
+  end
+
+  defp do_api_request(playlist_id) do
+    Logger.debug("Fetching recent media IDs from YouTube API for playlist: #{playlist_id}")
+
+    playlist_id
+    |> construct_api_endpoint()
+    |> http_client().get(accept: "application/json")
+    |> case do
+      {:ok, response} ->
+        Phoenix.json_library().decode(response)
+
+      {:error, reason} ->
+        Logger.error("Failed to fetch YouTube API: #{inspect(reason)}")
+        {:error, reason}
+    end
+  end
+
+  defp get_media_ids_from_response(parsed_json) do
+    parsed_json
+    |> Map.get("items", [])
+    |> Enum.map(fn item ->
+      item
+      |> Map.get("contentDetails", %{})
+      |> Map.get("videoId", nil)
+    end)
+    |> Enum.reject(&is_nil/1)
+    |> Enum.uniq()
+    |> FunctionUtils.wrap_ok()
+  end
+
+  defp api_key do
+    Settings.get!(:youtube_api_key)
+  end
+
+  defp construct_api_endpoint(playlist_id) do
+    api_base = "https://youtube.googleapis.com/youtube/v3/playlistItems"
+    property_type = "contentDetails"
+    max_results = 50
+
+    "#{api_base}?part=#{property_type}&maxResults=#{max_results}&playlistId=#{playlist_id}&key=#{api_key()}"
+  end
+
+  defp http_client do
+    Application.get_env(:pinchflat, :http_client, Pinchflat.HTTP.HTTPClient)
+  end
+end

lib/pinchflat/fast_indexing/youtube_behaviour.ex

@@ -0,0 +1,11 @@
+defmodule Pinchflat.FastIndexing.YoutubeBehaviour do
+  @moduledoc """
+  This module defines the behaviour for clients that interface with YouTube
+  for the purpose of fast indexing.
+  """
+
+  alias Pinchflat.Sources.Source
+
+  @callback enabled?() :: boolean()
+  @callback get_recent_media_ids(%Source{}) :: {:ok, [String.t()]} | {:error, String.t()}
+end

lib/pinchflat/fast_indexing/youtube_rss.ex

@@ -1,18 +1,31 @@
 defmodule Pinchflat.FastIndexing.YoutubeRss do
   @moduledoc """
-  Methods for interacting with YouTube RSS feeds
+  Methods for interacting with YouTube RSS feeds for fast indexing
   """
 
   require Logger
 
   alias Pinchflat.Sources.Source
+  alias Pinchflat.FastIndexing.YoutubeBehaviour
+
+  @behaviour YoutubeBehaviour
+
+  @doc """
+  Determines if the YouTube RSS feed is enabled for fast indexing. Used to satisfy
+  the `YoutubeBehaviour` behaviour.
+
+  Returns true
+  """
+  @impl YoutubeBehaviour
+  def enabled?(), do: true
 
   @doc """
   Fetches the recent media IDs from a YouTube RSS feed for a given source.
 
   Returns {:ok, [binary()]} | {:error, binary()}
   """
-  def get_recent_media_ids_from_rss(%Source{} = source) do
+  @impl YoutubeBehaviour
+  def get_recent_media_ids(%Source{} = source) do
     Logger.debug("Fetching recent media IDs from YouTube RSS feed for source: #{source.collection_id}")
 
     case http_client().get(rss_url_for_source(source)) do

lib/pinchflat/http/http_client.ex

@@ -21,9 +21,11 @@ defmodule Pinchflat.HTTP.HTTPClient do
   """
   @impl HTTPBehaviour
   def get(url, headers \\ [], opts \\ []) do
+    headers = parse_headers(headers)
+
     case :httpc.request(:get, {url, headers}, [], opts) do
       {:ok, {{_version, 200, _reason_phrase}, _headers, body}} ->
-        {:ok, body}
+        {:ok, to_string(body)}
 
       {:ok, {{_version, status_code, reason_phrase}, _headers, _body}} ->
         {:error, "HTTP request failed with status code #{status_code}: #{reason_phrase}"}
@@ -32,4 +34,8 @@ defmodule Pinchflat.HTTP.HTTPClient do
         {:error, "HTTP request failed: #{reason}"}
     end
   end
+
+  defp parse_headers(headers) do
+    Enum.map(headers, fn {k, v} -> {to_charlist(k), to_charlist(v)} end)
+  end
 end

lib/pinchflat/settings/setting.ex

@@ -13,7 +13,8 @@ defmodule Pinchflat.Settings.Setting do
     :apprise_version,
     :apprise_server,
     :video_codec_preference,
-    :audio_codec_preference
+    :audio_codec_preference,
+    :youtube_api_key
   ]
 
   @required_fields ~w(
@@ -29,6 +30,7 @@ defmodule Pinchflat.Settings.Setting do
     field :yt_dlp_version, :string
     field :apprise_version, :string
     field :apprise_server, :string
+    field :youtube_api_key, :string
 
     field :video_codec_preference, :string
     field :audio_codec_preference, :string

lib/pinchflat/sources/source.ex

@@ -129,7 +129,7 @@ defmodule Pinchflat.Sources.Source do
   @doc false
   def fast_index_frequency do
     # minutes
-    15
+    10
   end
 
   @doc false