Bumped version

* Disabled pro modal during onboarding

* Restructured the way onboarding changes settings

* Added better upload date template placeholder

.check.exs

@@ -9,18 +9,15 @@
   fix: true,
 
   ## don't retry automatically even if last run resulted in failures
-  # retry: false,
+  retry: false,
 
   ## list of tools (see `mix check` docs for a list of default curated tools)
   tools: [
     {:compiler, env: %{"MIX_ENV" => "test"}},
     {:formatter, env: %{"MIX_ENV" => "test"}},
     {:sobelow, "mix sobelow --config"},
-    # TODO: delete these and replace them with builtin ex_unit and formatter tools
+    {:prettier_formatting, "yarn run prettier . --check", fix: "yarn run prettier . --write"},
-    # once Elixir 1.16.2 is released (see: https://github.com/karolsluszniak/ex_check/issues/41#issuecomment-1921390413)
+    {:npm_test, false}
-    {:elixir_tests, "mix test"},
-    {:elixir_formatting, "mix format --check-formatted", fix: "mix format"},
-    {:prettier_formatting, "yarn run prettier . --check", fix: "yarn run prettier . --write"}
 
     ## curated tools may be disabled (e.g. the check for compilation warnings)
     # {:compiler, false},

.iex.exs

@@ -1,10 +1,11 @@
+import Ecto.Query, warn: false
 alias Pinchflat.Repo
 
 alias Pinchflat.Tasks.Task
+alias Pinchflat.Sources.Source
 alias Pinchflat.Media.MediaItem
 alias Pinchflat.Tasks.SourceTasks
 alias Pinchflat.Media.MediaMetadata
-alias Pinchflat.Sources.Source
 alias Pinchflat.Profiles.MediaProfile
 
 alias Pinchflat.Tasks
@@ -13,8 +14,12 @@ alias Pinchflat.Profiles
 alias Pinchflat.Sources
 alias Pinchflat.Settings
 
-alias Pinchflat.MediaClient.{SourceDetails, MediaDownloader}
+alias Pinchflat.MediaClient.MediaDownloader
-alias Pinchflat.Metadata.{Zipper, ThumbnailFetcher}
+alias Pinchflat.YtDlp.Backend.Media, as: YtDlpMedia
+alias Pinchflat.YtDlp.Backend.MediaCollection, as: YtDlpCollection
+
+alias Pinchflat.Api.YoutubeRss
+alias Pinchflat.Metadata.MetadataFileHelpers
 
 alias Pinchflat.Utils.FilesystemUtils.FileFollowerServer
 
@@ -31,6 +36,10 @@ defmodule IexHelpers do
     "https://www.youtube.com/watch?v=bR52O78ZIUw"
   end
 
+  def last_media_item do
+    Repo.one(from m in MediaItem, limit: 1)
+  end
+
   def details(type) do
     source =
       case type do
@@ -38,7 +47,7 @@ defmodule IexHelpers do
         :channel -> channel_url()
       end
 
-    SourceDetails.get_source_details(source)
+    YtDlpCollection.get_source_details(source)
   end
 
   def ids(type) do
@@ -48,7 +57,7 @@ defmodule IexHelpers do
         :channel -> channel_url()
       end
 
-    SourceDetails.get_media_attributes(source)
+    YtDlpCollection.get_media_attributes_for_collection(source)
   end
 end
 

config/config.exs

@@ -12,7 +12,7 @@ config :pinchflat,
   generators: [timestamp_type: :utc_datetime],
   # Specifying backend data here makes mocking and local testing SUPER easy
   yt_dlp_executable: System.find_executable("yt-dlp"),
-  yt_dlp_runner: Pinchflat.MediaClient.Backends.YtDlp.CommandRunner,
+  yt_dlp_runner: Pinchflat.YtDlp.Backend.CommandRunner,
   media_directory: "/downloads",
   # The user may or may not store metadata for their needs, but the app will always store its copy
   metadata_directory: "/config/metadata",
@@ -26,6 +26,8 @@ config :pinchflat,
 # Configures the endpoint
 config :pinchflat, PinchflatWeb.Endpoint,
   url: [host: "localhost", port: 8945],
+  # NOTE: this must be updated if ever deployed traditionally (ie: not self-hosted)
+  check_origin: false,
   adapter: Phoenix.Endpoint.Cowboy2Adapter,
   render_errors: [
     formats: [html: PinchflatWeb.ErrorHTML, json: PinchflatWeb.ErrorJSON],
@@ -40,7 +42,14 @@ config :pinchflat, Oban,
   # Keep old jobs for 30 days for display in the UI
   plugins: [{Oban.Plugins.Pruner, max_age: 30 * 24 * 60 * 60}],
   # TODO: consider making this an env var or something?
-  queues: [default: 10, media_indexing: 2, media_fetching: 2, media_local_metadata: 8]
+  queues: [
+    default: 10,
+    fast_indexing: 6,
+    media_indexing: 2,
+    media_collection_indexing: 2,
+    media_fetching: 2,
+    media_local_metadata: 8
+  ]
 
 # Configures the mailer
 #

config/dev.exs

@@ -21,7 +21,6 @@ config :pinchflat, PinchflatWeb.Endpoint,
   # Binding to loopback ipv4 address prevents access from other machines.
   # Change to `ip: {0, 0, 0, 0}` to allow access from other machines.
   http: [port: 4008],
-  check_origin: false,
   code_reloader: true,
   debug_errors: true,
   secret_key_base: "QLKKs3ypkUgJ/fMnWZaIYqpMbnA4IlPVEm3tvezsblhFDv4b67rdp+AmTpAFFURK",

dev.Dockerfile

@@ -1,9 +1,18 @@
-# Extend from the official Elixir image.
+ARG ELIXIR_VERSION=1.16.2
-FROM elixir:latest
+ARG OTP_VERSION=26.2.2
+ARG DEBIAN_VERSION=bookworm-20240130
+ARG DEV_IMAGE="hexpm/elixir:${ELIXIR_VERSION}-erlang-${OTP_VERSION}-debian-${DEBIAN_VERSION}"
+
+FROM ${DEV_IMAGE}
+
+# Set the locale deets
+ENV LANG en_US.UTF-8
+ENV LANGUAGE en_US:en
+ENV LC_ALL en_US.UTF-8
 
 # Install debian packages
 RUN apt-get update -qq
-RUN apt-get install -y inotify-tools ffmpeg \
+RUN apt-get install -y inotify-tools ffmpeg curl git openssh-client \
   python3 python3-pip python3-setuptools python3-wheel python3-dev
 
 # Install nodejs
@@ -28,6 +37,7 @@ COPY . ./
 RUN chmod +x ./docker-run.dev.sh
 
 # Install Elixir deps
+# RUN mix archive.install github hexpm/hex branch latest
 RUN mix deps.get
 # Gives us iex shell history
 ENV ERL_AFLAGS="-kernel shell_history enabled"

ideas.md

@@ -2,5 +2,3 @@
   - Use a UUID for the media database ID (or at least alongside it)
 - Look into this and its recommended plugins https://hexdocs.pm/ex_check/readme.html
 - Add output template option for the source's friendly name
-- TODO: Install Elixir 1.16.2 when available to fix bug with `ex_check` https://github.com/karolsluszniak/ex_check/issues/41#issuecomment-1921390413
-  - delete `{:elixir_tests, "mix test"}` and formatting check

lib/pinchflat/api/youtube_rss.ex

@@ -0,0 +1,51 @@
+defmodule Pinchflat.Api.YoutubeRss do
+  @moduledoc """
+  Methods for interacting with YouTube RSS feeds
+  """
+
+  require Logger
+
+  alias Pinchflat.Sources.Source
+
+  @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
+    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
+      {:ok, response} ->
+        response = to_string(response)
+        media_id_regex = ~r/<yt:videoId>(.*?)<\/yt:videoId>/
+
+        # Don't get on me about using regex to search XML.
+        # The content is known, well-formed, and simple.
+        media_ids =
+          media_id_regex
+          |> Regex.scan(response)
+          |> Enum.map(fn [_, id] -> String.trim(id) end)
+          |> Enum.filter(&(String.length(&1) > 0))
+          |> Enum.uniq()
+
+        Logger.debug("Media ids fetched from RSS: #{inspect(media_ids)}")
+
+        {:ok, media_ids}
+
+      {:error, _reason} ->
+        {:error, "Failed to fetch RSS feed"}
+    end
+  end
+
+  defp rss_url_for_source(source) do
+    case source.collection_type do
+      :channel -> "https://www.youtube.com/feeds/videos.xml?channel_id=#{source.collection_id}"
+      :playlist -> "https://www.youtube.com/feeds/videos.xml?playlist_id=#{source.collection_id}"
+    end
+  end
+
+  defp http_client do
+    Application.get_env(:pinchflat, :http_client, Pinchflat.HTTP.HTTPClient)
+  end
+end

lib/pinchflat/application.ex

@@ -10,9 +10,9 @@ defmodule Pinchflat.Application do
     children = [
       PinchflatWeb.Telemetry,
       Pinchflat.Repo,
-      # {Task, &run_startup_tasks/0},
+      # Must be before startup tasks
-      Pinchflat.StartupTasks,
       {Oban, Application.fetch_env!(:pinchflat, Oban)},
+      Pinchflat.StartupTasks,
       {DNSCluster, query: Application.get_env(:pinchflat, :dns_cluster_query) || :ignore},
       {Phoenix.PubSub, name: Pinchflat.PubSub},
       # Start the Finch HTTP client for sending emails

lib/pinchflat/http/http_behaviour.ex

@@ -4,5 +4,7 @@ defmodule Pinchflat.HTTP.HTTPBehaviour do
   so I can use Mox to create an HTTP mock
   """
 
+  @callback get(String.t()) :: {:ok, String.t()} | {:error, String.t()}
+  @callback get(String.t(), Keyword.t()) :: {:ok, String.t()} | {:error, String.t()}
   @callback get(String.t(), Keyword.t(), Keyword.t()) :: {:ok, String.t()} | {:error, String.t()}
 end

lib/pinchflat/media.ex

@@ -31,6 +31,22 @@ defmodule Pinchflat.Media do
     |> Repo.all()
   end
 
+  @doc """
+  Fetches all media items belonging to a given source that have a media_id in the given list.
+  Useful for determining the what media items we DON'T already have for fast indexing.
+
+  NOTE: These queries are getting a little tedious. When I have the time, I should see about
+  implementing a query pattern and having these compose queries from a common base. This would
+  also let me compose simple queries in the module using them for one-off methods
+
+  Returns [%MediaItem{}, ...].
+  """
+  def list_media_items_by_media_id_for(%Source{} = source, media_ids) do
+    MediaItem
+    |> where([mi], mi.source_id == ^source.id and mi.media_id in ^media_ids)
+    |> Repo.all()
+  end
+
   @doc """
   Returns a list of pending media_items for a given source, where
   pending means the `media_filepath` is `nil` AND the media_item
@@ -48,6 +64,7 @@ defmodule Pinchflat.Media do
     MediaItem
     |> where([mi], mi.source_id == ^source.id and is_nil(mi.media_filepath))
     |> where(^build_format_clauses(media_profile))
+    |> where(^maybe_apply_cutoff_date(source))
     |> Repo.maybe_limit(limit)
     |> Repo.all()
   end
@@ -76,11 +93,12 @@ defmodule Pinchflat.Media do
   Returns boolean()
   """
   def pending_download?(%MediaItem{} = media_item) do
-    media_profile = Repo.preload(media_item, source: :media_profile).source.media_profile
+    media_item = Repo.preload(media_item, source: :media_profile)
 
     MediaItem
     |> where([mi], mi.id == ^media_item.id and is_nil(mi.media_filepath))
-    |> where(^build_format_clauses(media_profile))
+    |> where(^build_format_clauses(media_item.source.media_profile))
+    |> where(^maybe_apply_cutoff_date(media_item.source))
     |> Repo.exists?()
   end
 
@@ -156,7 +174,9 @@ defmodule Pinchflat.Media do
   end
 
   @doc """
-  Creates a media_item. Returns {:ok, %MediaItem{}} | {:error, %Ecto.Changeset{}}.
+  Creates a media_item.
+
+  Returns {:ok, %MediaItem{}} | {:error, %Ecto.Changeset{}}
   """
   def create_media_item(attrs) do
     %MediaItem{}
@@ -165,7 +185,32 @@ defmodule Pinchflat.Media do
   end
 
   @doc """
-  Updates a media_item. Returns {:ok, %MediaItem{}} | {:error, %Ecto.Changeset{}}.
+  Creates a media item from the attributes returned by the video backend
+  (read: yt-dlp).
+
+  Unlike `create_media_item`, this will attempt an update if the media_item
+  already exists. This is so that future indexing can pick up attributes that
+  we may not have asked for in the past (eg: upload_date)
+
+  Returns {:ok, %MediaItem{}} | {:error, %Ecto.Changeset{}}
+  """
+  def create_media_item_from_backend_attrs(source, media_attrs_struct) do
+    attrs = Map.merge(%{source_id: source.id}, Map.from_struct(media_attrs_struct))
+
+    %MediaItem{}
+    |> MediaItem.changeset(attrs)
+    |> Repo.insert(
+      on_conflict: [
+        set: Map.to_list(attrs)
+      ],
+      conflict_target: [:source_id, :media_id]
+    )
+  end
+
+  @doc """
+  Updates a media_item.
+
+  Returns {:ok, %MediaItem{}} | {:error, %Ecto.Changeset{}}
   """
   def update_media_item(%MediaItem{} = media_item, attrs) do
     media_item
@@ -177,7 +222,7 @@ defmodule Pinchflat.Media do
   Deletes a media_item and its associated tasks.
   Can optionally delete the media_item's files.
 
-  Returns {:ok, %MediaItem{}} | {:error, %Ecto.Changeset{}}.
+  Returns {:ok, %MediaItem{}} | {:error, %Ecto.Changeset{}}
   """
   def delete_media_item(%MediaItem{} = media_item, opts \\ []) do
     delete_files = Keyword.get(opts, :delete_files, false)
@@ -217,6 +262,14 @@ defmodule Pinchflat.Media do
     {:ok, media_item}
   end
 
+  defp maybe_apply_cutoff_date(source) do
+    if source.download_cutoff_date do
+      dynamic([mi], mi.upload_date >= ^source.download_cutoff_date)
+    else
+      dynamic(true)
+    end
+  end
+
   defp build_format_clauses(media_profile) do
     mapped_struct = Map.from_struct(media_profile)
 
@@ -225,7 +278,7 @@ defmodule Pinchflat.Media do
         {{:shorts_behaviour, :only}, %{livestream_behaviour: :only}} ->
           dynamic(
             [mi],
-            ^dynamic and (mi.livestream == true or fragment("LOWER(?) LIKE LOWER(?)", mi.original_url, "%/shorts/%"))
+            ^dynamic and (mi.livestream == true or mi.short_form_content == true)
           )
 
         # Technically redundant, but makes the other clauses easier to parse
@@ -234,16 +287,13 @@ defmodule Pinchflat.Media do
           dynamic
 
         {{:shorts_behaviour, :only}, _} ->
-          # return records with /shorts/ in the original_url
+          dynamic([mi], ^dynamic and mi.short_form_content == true)
-          dynamic([mi], ^dynamic and fragment("LOWER(?) LIKE LOWER(?)", mi.original_url, "%/shorts/%"))
 
         {{:livestream_behaviour, :only}, _} ->
-          # return records with livestream: true
           dynamic([mi], ^dynamic and mi.livestream == true)
 
         {{:shorts_behaviour, :exclude}, %{livestream_behaviour: lb}} when lb != :only ->
-          # return records without /shorts/ in the original_url
+          dynamic([mi], ^dynamic and mi.short_form_content == false)
-          dynamic([mi], ^dynamic and fragment("LOWER(?) NOT LIKE LOWER(?)", mi.original_url, "%/shorts/%"))
 
         {{:livestream_behaviour, :exclude}, %{shorts_behaviour: sb}} when sb != :only ->
           # return records with livestream: false

lib/pinchflat/media/media_item.ex

@@ -12,14 +12,16 @@ defmodule Pinchflat.Media.MediaItem do
   alias Pinchflat.Media.MediaItemSearchIndex
 
   @allowed_fields [
-    # these fields are captured on indexing
+    # these fields are captured on indexing (and again on download)
     :title,
     :media_id,
     :description,
     :original_url,
     :livestream,
     :source_id,
-    # these fields are captured on download
+    :short_form_content,
+    :upload_date,
+    # these fields are captured only on download
     :media_downloaded_at,
     :media_filepath,
     :media_size_bytes,
@@ -27,7 +29,16 @@ defmodule Pinchflat.Media.MediaItem do
     :thumbnail_filepath,
     :metadata_filepath
   ]
-  @required_fields ~w(title original_url livestream media_id source_id)a
+  # Pretty much all the fields captured at index are required.
+  @required_fields ~w(
+    title
+    original_url
+    livestream
+    media_id
+    source_id
+    upload_date
+    short_form_content
+    )a
 
   schema "media_items" do
     field :title, :string
@@ -35,7 +46,9 @@ defmodule Pinchflat.Media.MediaItem do
     field :description, :string
     field :original_url, :string
     field :livestream, :boolean, default: false
+    field :short_form_content, :boolean, default: false
     field :media_downloaded_at, :utc_datetime
+    field :upload_date, :date
 
     field :media_filepath, :string
     field :media_size_bytes, :integer

lib/pinchflat/media_client/backends/yt_dlp/media.ex

@@ -1,26 +0,0 @@
-defmodule Pinchflat.MediaClient.Backends.YtDlp.Media do
-  @moduledoc """
-  Contains utilities for working with singular pieces of media
-  """
-
-  @doc """
-  Downloads a single piece of media (and possibly its metadata) directly to its
-  final destination. Returns the parsed JSON output from yt-dlp.
-
-  Returns {:ok, map()} | {:error, any, ...}.
-  """
-  def download(url, command_opts \\ []) do
-    opts = [:no_simulate] ++ command_opts
-
-    with {:ok, output} <- backend_runner().run(url, opts, "after_move:%()j"),
-         {:ok, parsed_json} <- Phoenix.json_library().decode(output) do
-      {:ok, parsed_json}
-    else
-      err -> err
-    end
-  end
-
-  defp backend_runner do
-    Application.get_env(:pinchflat, :yt_dlp_runner)
-  end
-end

lib/pinchflat/media_client/source_details.ex

@@ -1,47 +0,0 @@
-defmodule Pinchflat.MediaClient.SourceDetails do
-  @moduledoc """
-  This is the integration layer for actually working with sources.
-
-  Technically hardcodes the yt-dlp backend for now, but should leave
-  it open-ish for future expansion (just in case).
-  """
-
-  alias Pinchflat.Sources.Source
-  alias Pinchflat.MediaClient.Backends.YtDlp.MediaCollection, as: YtDlpSource
-
-  @doc """
-  Gets a source's ID and name from its URL using the given backend.
-
-  Returns {:ok, map()} | {:error, any, ...}.
-  """
-  def get_source_details(source_url, backend \\ :yt_dlp) do
-    source_module(backend).get_source_details(source_url)
-  end
-
-  @doc """
-  Returns a list of basic media data maps for the given source URL OR
-  source record using the given backend.
-
-  Options:
-    - :file_listener_handler - a function that will be called with the path to the
-      file that will be written to by yt-dlp. This is useful for
-      setting up a file watcher to read the file as it gets written to.
-
-  Returns {:ok, [map()]} | {:error, any, ...}.
-  """
-  def get_media_attributes(sourceable, opts \\ [], backend \\ :yt_dlp)
-
-  def get_media_attributes(%Source{} = source, opts, backend) do
-    get_media_attributes(source.collection_id, opts, backend)
-  end
-
-  def get_media_attributes(source_url, opts, backend) when is_binary(source_url) do
-    source_module(backend).get_media_attributes(source_url, opts)
-  end
-
-  defp source_module(backend) do
-    case backend do
-      :yt_dlp -> YtDlpSource
-    end
-  end
-end

lib/pinchflat/metadata/metadata_file_helpers.ex

@@ -1,4 +1,4 @@
-defmodule Pinchflat.MediaClient.Backends.YtDlp.MetadataFileHelpers do
+defmodule Pinchflat.Metadata.MetadataFileHelpers do
   @moduledoc """
   Provides methods for creating/downloading/storing related metadata
   out-of-band of the normal yt-dlp backend process.

lib/pinchflat/metadata/metadata_parser.ex

@@ -1,4 +1,4 @@
-defmodule Pinchflat.MediaClient.Backends.YtDlp.MetadataParser do
+defmodule Pinchflat.Metadata.MetadataParser do
   @moduledoc """
   yt-dlp offers a LOT of metadata in its JSON response, some of which
   needs to be extracted and included in various models.
@@ -25,9 +25,12 @@ defmodule Pinchflat.MediaClient.Backends.YtDlp.MetadataParser do
 
   defp parse_media_metadata(metadata) do
     %{
+      media_id: metadata["id"],
       title: metadata["title"],
+      original_url: metadata["original_url"],
       description: metadata["description"],
-      media_filepath: metadata["filepath"]
+      media_filepath: metadata["filepath"],
+      livestream: metadata["was_live"]
     }
   end
 

lib/pinchflat/profiles/media_profile.ex

@@ -30,7 +30,7 @@ defmodule Pinchflat.Profiles.MediaProfile do
     field :name, :string
 
     field :output_path_template, :string,
-      default: "/{{ source_custom_name }}/{{ title }}/{{ title }} [{{ id }}].{{ ext }}"
+      default: "/{{ source_custom_name }}/{{ upload_yyyy_mm_dd }} {{ title }}/{{ title }} [{{ id }}].{{ ext }}"
 
     field :download_subs, :boolean, default: true
     field :download_auto_subs, :boolean, default: true

lib/pinchflat/profiles/output_path_builder.ex

@@ -1,8 +1,6 @@
-defmodule Pinchflat.Profiles.Options.YtDlp.OutputPathBuilder do
+defmodule Pinchflat.Profiles.OutputPathBuilder do
   @moduledoc """
   Builds yt-dlp-friendly output paths for downloaded media
-
-  IDEA: consider making this a behaviour so I can add other backends later
   """
 
   alias Pinchflat.RenderedString.Parser, as: TemplateParser
@@ -41,7 +39,8 @@ defmodule Pinchflat.Profiles.Options.YtDlp.OutputPathBuilder do
       # Individual parts of the upload date
       "upload_year" => "%(upload_date>%Y)S",
       "upload_month" => "%(upload_date>%m)S",
-      "upload_day" => "%(upload_date>%d)S"
+      "upload_day" => "%(upload_date>%d)S",
+      "upload_yyyy_mm_dd" => "%(upload_date>%Y)S-%(upload_date>%m)S-%(upload_date>%d)S"
     }
   end
 end

lib/pinchflat/sources.ex

@@ -11,7 +11,7 @@ defmodule Pinchflat.Sources do
   alias Pinchflat.Sources.Source
   alias Pinchflat.Tasks.SourceTasks
   alias Pinchflat.Profiles.MediaProfile
-  alias Pinchflat.MediaClient.SourceDetails
+  alias Pinchflat.YtDlp.Backend.MediaCollection
 
   @doc """
   Returns the list of sources. Returns [%Source{}, ...]
@@ -41,12 +41,24 @@ defmodule Pinchflat.Sources do
   original_url (if provided). Will attempt to start indexing the source's
   media if successfully inserted.
 
+  Runs an initial `change_source` check to ensure most of the source is valid
+  before making an expensive API call. Runs it through `Repo.insert` even
+  though we know it's going to fail so it picks up any addl. database errors
+  and fulfills our return contract.
+
   Returns {:ok, %Source{}} | {:error, %Ecto.Changeset{}}
   """
   def create_source(attrs) do
-    %Source{}
+    case change_source(%Source{}, attrs, :initial) do
-    |> change_source_from_url(attrs)
+      %Ecto.Changeset{valid?: true} ->
-    |> commit_and_handle_tasks()
+        %Source{}
+        |> change_source_from_url(attrs)
+        |> maybe_change_indexing_frequency()
+        |> commit_and_handle_tasks()
+
+      changeset ->
+        Repo.insert(changeset)
+    end
   end
 
   @doc """
@@ -57,12 +69,24 @@ defmodule Pinchflat.Sources do
   Existing indexing tasks will be cancelled if the indexing frequency has been
   changed (logic in `SourceTasks.kickoff_indexing_task`)
 
+  Runs an initial `change_source` check to ensure most of the source is valid
+  before making an expensive API call. Runs it through `Repo.update` even
+  though we know it's going to fail so it picks up any addl. database errors
+  and fulfills our return contract.
+
   Returns {:ok, %Source{}} | {:error, %Ecto.Changeset{}}
   """
   def update_source(%Source{} = source, attrs) do
-    source
+    case change_source(source, attrs, :initial) do
-    |> change_source_from_url(attrs)
+      %Ecto.Changeset{valid?: true} ->
-    |> commit_and_handle_tasks()
+        source
+        |> change_source_from_url(attrs)
+        |> maybe_change_indexing_frequency()
+        |> commit_and_handle_tasks()
+
+      changeset ->
+        Repo.update(changeset)
+    end
   end
 
   @doc """
@@ -87,8 +111,8 @@ defmodule Pinchflat.Sources do
   @doc """
   Returns an `%Ecto.Changeset{}` for tracking source changes.
   """
-  def change_source(%Source{} = source, attrs \\ %{}) do
+  def change_source(%Source{} = source, attrs \\ %{}, validation_stage \\ :pre_insert) do
-    Source.changeset(source, attrs)
+    Source.changeset(source, attrs, validation_stage)
   end
 
   @doc """
@@ -116,7 +140,7 @@ defmodule Pinchflat.Sources do
   defp add_source_details_to_changeset(source, changeset) do
     %Ecto.Changeset{changes: changes} = changeset
 
-    case SourceDetails.get_source_details(changes.original_url) do
+    case MediaCollection.get_source_details(changes.original_url) do
       {:ok, source_details} ->
         add_source_details_by_collection_type(source, changeset, source_details)
 
@@ -151,6 +175,20 @@ defmodule Pinchflat.Sources do
     change_source(source, Map.merge(changes, collection_changes))
   end
 
+  defp maybe_change_indexing_frequency(changeset) do
+    fast_index = Ecto.Changeset.get_field(changeset, :fast_index)
+
+    if fast_index do
+      Ecto.Changeset.put_change(
+        changeset,
+        :index_frequency_minutes,
+        Source.index_frequency_when_fast_indexing()
+      )
+    else
+      changeset
+    end
+  end
+
   defp commit_and_handle_tasks(changeset) do
     case Repo.insert_or_update(changeset) do
       {:ok, %Source{} = source} ->
@@ -188,13 +226,38 @@ defmodule Pinchflat.Sources do
       # If the record has been persisted, only run indexing if the
       # indexing frequency has been changed and is now greater than 0
       %{__meta__: %{state: :loaded}} ->
-        case changeset.changes do
+        maybe_update_slow_indexing_task(changeset, source)
-          %{index_frequency_minutes: mins} when mins > 0 -> SourceTasks.kickoff_indexing_task(source)
+        maybe_update_fast_indexing_task(changeset, source)
-          %{index_frequency_minutes: _} -> Tasks.delete_pending_tasks_for(source, "MediaIndexingWorker")
-          _ -> :ok
-        end
     end
 
     {:ok, source}
   end
+
+  defp maybe_update_slow_indexing_task(changeset, source) do
+    case changeset.changes do
+      %{index_frequency_minutes: mins} when mins > 0 ->
+        SourceTasks.kickoff_indexing_task(source)
+
+      %{index_frequency_minutes: _} ->
+        Tasks.delete_pending_tasks_for(source, "FastIndexingWorker")
+        Tasks.delete_pending_tasks_for(source, "MediaIndexingWorker")
+        Tasks.delete_pending_tasks_for(source, "MediaCollectionIndexingWorker")
+
+      _ ->
+        :ok
+    end
+  end
+
+  defp maybe_update_fast_indexing_task(changeset, source) do
+    case changeset.changes do
+      %{fast_index: true} ->
+        SourceTasks.kickoff_fast_indexing_task(source)
+
+      %{fast_index: false} ->
+        Tasks.delete_pending_tasks_for(source, "FastIndexingWorker")
+
+      _ ->
+        :ok
+    end
+  end
 end

lib/pinchflat/sources/source.ex

@@ -17,22 +17,33 @@ defmodule Pinchflat.Sources.Source do
     collection_type
     custom_name
     index_frequency_minutes
+    fast_index
     download_media
     last_indexed_at
     original_url
+    download_cutoff_date
+    media_profile_id
+  )a
+
+  # Expensive API calls are made when a source is inserted/updated so
+  # we want to ensure that the source is valid before making the call.
+  # This way, we check that the other attributes are valid before ensuring
+  # that all fields are valid.
+  @initially_required_fields ~w(
+    index_frequency_minutes
+    fast_index
+    download_media
+    original_url
     media_profile_id
   )a
 
-  @required_fields ~w(
+  @pre_insert_required_fields @initially_required_fields ++
-    collection_name
+                                ~w(
-    collection_id
+                                  custom_name
-    collection_type
+                                  collection_name
-    custom_name
+                                  collection_id
-    index_frequency_minutes
+                                  collection_type
-    download_media
+                                )a
-    original_url
-    media_profile_id
-  )a
 
   schema "sources" do
     field :custom_name, :string
@@ -40,10 +51,11 @@ defmodule Pinchflat.Sources.Source do
     field :collection_id, :string
     field :collection_type, Ecto.Enum, values: [:channel, :playlist]
     field :index_frequency_minutes, :integer, default: 60 * 24
+    field :fast_index, :boolean, default: false
     field :download_media, :boolean, default: true
     field :last_indexed_at, :utc_datetime
-    # This should only be used for user reference going forward
+    # Only download media items that were published after this date
-    # as the collection_id should be used for all API calls
+    field :download_cutoff_date, :date
     field :original_url, :string
 
     belongs_to :media_profile, MediaProfile
@@ -55,11 +67,31 @@ defmodule Pinchflat.Sources.Source do
   end
 
   @doc false
-  def changeset(source, attrs) do
+  def changeset(source, attrs, validation_stage) do
+    # See above for rationale
+    required_fields =
+      if validation_stage == :initial do
+        @initially_required_fields
+      else
+        @pre_insert_required_fields
+      end
+
     source
     |> cast(attrs, @allowed_fields)
     |> dynamic_default(:custom_name, fn cs -> get_field(cs, :collection_name) end)
-    |> validate_required(@required_fields)
+    |> validate_required(required_fields)
     |> unique_constraint([:collection_id, :media_profile_id])
   end
+
+  @doc false
+  def index_frequency_when_fast_indexing do
+    # 30 days in minutes
+    60 * 24 * 30
+  end
+
+  @doc false
+  def fast_index_frequency do
+    # minutes
+    15
+  end
 end

lib/pinchflat/startup_tasks.ex

@@ -8,8 +8,11 @@ defmodule Pinchflat.StartupTasks do
 
   # restart: :temporary means that this process will never be restarted (ie: will run once and then die)
   use GenServer, restart: :temporary
+  import Ecto.Query, warn: false
 
+  alias Pinchflat.Repo
   alias Pinchflat.Settings
+  alias Pinchflat.Workers.DataBackfillWorker
 
   def start_link(opts \\ []) do
     GenServer.start_link(__MODULE__, %{}, opts)
@@ -21,16 +24,27 @@ defmodule Pinchflat.StartupTasks do
   Any code defined here will run every time the application starts. You must
   make sure that the code is idempotent and safe to run multiple times.
 
-  This is a good place to set up default settings, create initial records, stuff like that
+  This is a good place to set up default settings, create initial records, stuff like that.
+  Should be fast - anything with the potential to be slow should be kicked off as a job instead.
   """
   @impl true
   def init(state) do
     apply_default_settings()
+    enqueue_backfill_worker()
 
     {:ok, state}
   end
 
   defp apply_default_settings do
     Settings.fetch!(:onboarding, true)
+    Settings.fetch!(:pro_enabled, false)
+  end
+
+  defp enqueue_backfill_worker do
+    DataBackfillWorker.cancel_pending_backfill_jobs()
+
+    %{}
+    |> DataBackfillWorker.new()
+    |> Repo.insert_unique_job()
   end
 end

lib/pinchflat/tasks/media_item_tasks.ex

@@ -6,6 +6,11 @@ defmodule Pinchflat.Tasks.MediaItemTasks do
   do is also defined here. Essentially, a one-stop-shop for media-related tasks/workers.
   """
   alias Pinchflat.Media
+  alias Pinchflat.Tasks
+  alias Pinchflat.Sources.Source
+  alias Pinchflat.Workers.MediaDownloadWorker
+
+  alias Pinchflat.YtDlp.Backend.Media, as: YtDlpMedia
 
   @doc """
   Fetches the file size of a media item and saves it to the database.
@@ -21,4 +26,36 @@ defmodule Pinchflat.Tasks.MediaItemTasks do
         err
     end
   end
+
+  @doc """
+  Indexes a single media item for a source and enqueues a download job if the
+  media should be downloaded. This method creates the media item record so it's
+  the one-stop-shop for adding a media item (and possibly downloading it) just
+  by a URL and source.
+
+  Returns {:ok, media_item} | {:error, any()}
+  """
+  def index_and_enqueue_download_for_media_item(%Source{} = source, url) do
+    maybe_media_item = create_media_item_from_url(source, url)
+
+    case maybe_media_item do
+      {:ok, media_item} ->
+        if source.download_media && Media.pending_download?(media_item) do
+          %{id: media_item.id}
+          |> MediaDownloadWorker.new()
+          |> Tasks.create_job_with_task(media_item)
+        end
+
+        {:ok, media_item}
+
+      err ->
+        err
+    end
+  end
+
+  defp create_media_item_from_url(source, url) do
+    {:ok, media_attrs} = YtDlpMedia.get_media_attributes(url)
+
+    Media.create_media_item_from_backend_attrs(source, media_attrs)
+  end
 end

lib/pinchflat/tasks/source_tasks.ex

@@ -12,31 +12,72 @@ defmodule Pinchflat.Tasks.SourceTasks do
   alias Pinchflat.Tasks
   alias Pinchflat.Sources
   alias Pinchflat.Sources.Source
+  alias Pinchflat.Api.YoutubeRss
   alias Pinchflat.Media.MediaItem
-  alias Pinchflat.MediaClient.SourceDetails
+  alias Pinchflat.Workers.FastIndexingWorker
-  alias Pinchflat.Workers.MediaIndexingWorker
   alias Pinchflat.Workers.MediaDownloadWorker
+  alias Pinchflat.Workers.MediaIndexingWorker
+  alias Pinchflat.YtDlp.Backend.MediaCollection
+  alias Pinchflat.Workers.MediaCollectionIndexingWorker
   alias Pinchflat.Utils.FilesystemUtils.FileFollowerServer
 
+  alias Pinchflat.YtDlp.Backend.Media, as: YtDlpMedia
+
   @doc """
   Starts tasks for indexing a source's media regardless of the source's indexing
-  frequency. It's assumed the caller will check for that.
+  frequency. It's assumed the caller will check for indexing frequency.
 
   Returns {:ok, %Task{}}.
   """
   def kickoff_indexing_task(%Source{} = source) do
+    Tasks.delete_pending_tasks_for(source, "FastIndexingWorker")
     Tasks.delete_pending_tasks_for(source, "MediaIndexingWorker")
+    Tasks.delete_pending_tasks_for(source, "MediaCollectionIndexingWorker")
 
-    source
+    %{id: source.id}
-    |> Map.take([:id])
     # Schedule this one immediately, but future ones will be on an interval
-    |> MediaIndexingWorker.new()
+    |> MediaCollectionIndexingWorker.new()
     |> Tasks.create_job_with_task(source)
-    |> case do
+  end
-      # This should never return {:error, :duplicate_job} since we just deleted
+
-      # any pending tasks. I'm being assertive about it so it's obvious if I'm wrong
+  @doc """
-      {:ok, task} -> {:ok, task}
+  Starts tasks for running a fast indexing task for a source's media
-    end
+  regardless of the source's fast_index state. It's assumed the
+  caller will check for fast_index.
+
+  This is used for running fast index tasks on update. On creation, the
+  fast index is enqueued after the slow index is complete.
+
+  Returns {:ok, %Task{}}.
+  """
+  def kickoff_fast_indexing_task(%Source{} = source) do
+    Tasks.delete_pending_tasks_for(source, "FastIndexingWorker")
+
+    %{id: source.id}
+    # Schedule this one immediately, but future ones will be on an interval
+    |> FastIndexingWorker.new()
+    |> Tasks.create_job_with_task(source)
+  end
+
+  @doc """
+  Fetches new media IDs from a source's YouTube RSS feed and kicks off indexing tasks
+  for any new media items. See comments in `MediaIndexingWorker` for more info on the
+  order of operations and how this fits into the indexing process.
+
+  Returns :ok
+  """
+  def kickoff_indexing_tasks_from_youtube_rss_feed(%Source{} = source) do
+    {:ok, media_ids} = YoutubeRss.get_recent_media_ids_from_rss(source)
+    existing_media_items = Media.list_media_items_by_media_id_for(source, media_ids)
+    new_media_ids = media_ids -- Enum.map(existing_media_items, & &1.media_id)
+
+    Enum.each(new_media_ids, fn media_id ->
+      url = "https://www.youtube.com/watch?v=#{media_id}"
+
+      %{id: source.id, media_url: url}
+      |> MediaIndexingWorker.new()
+      |> Tasks.create_job_with_task(source)
+    end)
   end
 
   @doc """
@@ -55,19 +96,17 @@ defmodule Pinchflat.Tasks.SourceTasks do
   job run. This should ensure that any stragglers are caught if, for some reason, they
   weren't enqueued or somehow got de-queued.
 
-  Since indexing returns all media data EVERY TIME, we rely on the unique index of the
+  Since indexing returns all media data EVERY TIME, we that that opportunity to update
-  media_id to prevent duplicates. Due to both the file follower and the fact that future
+  indexing metadata for media items that have already been created.
-  indexing will index a lot of existing data, this method will MOSTLY return error
-  changesets (from the unique index violation) and not media items. This is intended.
 
-  Returns [%MediaItem{}, ...] | [%Ecto.Changeset{}, ...]
+  Returns [%MediaItem{}, ...]
   """
   def index_and_enqueue_download_for_media_items(%Source{} = source) do
     # See the method definition below for more info on how file watchers work
     # (important reading if you're not familiar with it)
-    {:ok, media_attributes} = get_media_attributes_and_setup_file_watcher(source)
+    {:ok, media_attributes} = get_media_attributes_for_collection_and_setup_file_watcher(source)
-
     result = Enum.map(media_attributes, fn media_attrs -> create_media_item_from_attributes(source, media_attrs) end)
+
     Sources.update_source(source, %{last_indexed_at: DateTime.utc_now()})
     enqueue_pending_media_tasks(source)
 
@@ -89,8 +128,7 @@ defmodule Pinchflat.Tasks.SourceTasks do
     source
     |> Media.list_pending_media_items_for()
     |> Enum.each(fn media_item ->
-      media_item
+      %{id: media_item.id}
-      |> Map.take([:id])
       |> MediaDownloadWorker.new()
       |> Tasks.create_job_with_task(media_item)
     end)
@@ -116,7 +154,7 @@ defmodule Pinchflat.Tasks.SourceTasks do
   # lines (ie: you should gracefully fail if you can't parse a line).
   #
   # This works in-tandem with the normal (blocking) media indexing behaviour. When
-  # the `get_media_attributes` method completes it'll return the FULL result to
+  # the `get_media_attributes_for_collection` method completes it'll return the FULL result to
   # the caller for parsing. Ideally, every item in the list will have already
   # been processed by the file follower, but if not, the caller handles creation
   # of any media items that were missed/initially failed.
@@ -124,11 +162,11 @@ defmodule Pinchflat.Tasks.SourceTasks do
   # It attempts a graceful shutdown of the file follower after the indexing is done,
   # but the FileFollowerServer will also stop itself if it doesn't see any activity
   # for a sufficiently long time.
-  defp get_media_attributes_and_setup_file_watcher(source) do
+  defp get_media_attributes_for_collection_and_setup_file_watcher(source) do
     {:ok, pid} = FileFollowerServer.start_link()
 
     handler = fn filepath -> setup_file_follower_watcher(pid, filepath, source) end
-    result = SourceDetails.get_media_attributes(source.original_url, file_listener_handler: handler)
+    result = MediaCollection.get_media_attributes_for_collection(source.original_url, file_listener_handler: handler)
 
     FileFollowerServer.stop(pid)
 
@@ -141,7 +179,8 @@ defmodule Pinchflat.Tasks.SourceTasks do
         {:ok, media_attrs} ->
           Logger.debug("FileFollowerServer Handler: Got media attributes: #{inspect(media_attrs)}")
 
-          create_media_item_and_enqueue_download(source, media_attrs)
+          media_struct = YtDlpMedia.response_to_struct(media_attrs)
+          create_media_item_and_enqueue_download(source, media_struct)
 
         err ->
           Logger.debug("FileFollowerServer Handler: Error decoding JSON: #{inspect(err)}")
@@ -159,8 +198,7 @@ defmodule Pinchflat.Tasks.SourceTasks do
         if source.download_media && Media.pending_download?(media_item) do
           Logger.debug("FileFollowerServer Handler: Enqueuing download task for #{inspect(media_attrs)}")
 
-          media_item
+          %{id: media_item.id}
-          |> Map.take([:id])
           |> MediaDownloadWorker.new()
           |> Tasks.create_job_with_task(media_item)
         end
@@ -171,16 +209,7 @@ defmodule Pinchflat.Tasks.SourceTasks do
   end
 
   defp create_media_item_from_attributes(source, media_attrs) do
-    attrs = %{
+    case Media.create_media_item_from_backend_attrs(source, media_attrs) do
-      source_id: source.id,
-      title: media_attrs["title"],
-      media_id: media_attrs["id"],
-      original_url: media_attrs["original_url"],
-      livestream: media_attrs["was_live"],
-      description: media_attrs["description"]
-    }
-
-    case Media.create_media_item(attrs) do
       {:ok, media_item} -> media_item
       {:error, changeset} -> changeset
     end

lib/pinchflat/workers/data_backfill_worker.ex

@@ -0,0 +1,75 @@
+defmodule Pinchflat.Workers.DataBackfillWorker do
+  @moduledoc false
+
+  use Oban.Worker,
+    queue: :media_local_metadata,
+    unique: [period: :infinity, states: [:available, :scheduled, :retryable]],
+    tags: ["media_item", "media_metadata", "local_metadata", "data_backfill"]
+
+  # This one is going to be a little more self-contained
+  # instead of relying on outside modules for the methods.
+  # That's because, for now, these methods are not intended
+  # to be used elsewhere.
+  #
+  # I'm just trying out that pattern and seeing if I like it better
+  # so this may change.
+  import Ecto.Query, warn: false
+  require Logger
+
+  alias __MODULE__
+  alias Pinchflat.Repo
+  alias Pinchflat.Media.MediaItem
+
+  @doc """
+  Cancels all pending backfill jobs. Useful for ensuring worker runs immediately
+  on app boot.
+
+  Returns {:ok, integer()}
+  """
+  def cancel_pending_backfill_jobs do
+    Oban.Job
+    |> where(worker: "Pinchflat.Workers.DataBackfillWorker")
+    |> Oban.cancel_all_jobs()
+  end
+
+  @impl Oban.Worker
+  @doc """
+  Performs one-off tasks to get data in the right shape.
+  This can be needed when we add new features or change the way
+  we store data. Must be idempotent. All new data should already
+  conform to the expected schema so this should only be needed
+  for existing data. Still runs periodically to be safe.
+
+  Returns :ok
+  """
+  def perform(%Oban.Job{}) do
+    Logger.info("Running data backfill worker")
+    backfill_shorts_data()
+
+    reschedule_backfill()
+
+    :ok
+  end
+
+  defp backfill_shorts_data do
+    query =
+      from(
+        m in MediaItem,
+        where: fragment("? like ?", m.original_url, "%/shorts/%"),
+        where: m.short_form_content == false
+      )
+
+    {count, _} = Repo.update_all(query, set: [short_form_content: true])
+
+    Logger.info("Backfill worker set short_form_content to true for #{count} media items.")
+  end
+
+  defp reschedule_backfill do
+    # Run hourly
+    next_run_in = 60 * 60
+
+    %{}
+    |> DataBackfillWorker.new(schedule_in: next_run_in)
+    |> Repo.insert_unique_job()
+  end
+end

lib/pinchflat/workers/fast_indexing_worker.ex

@@ -0,0 +1,46 @@
+defmodule Pinchflat.Workers.FastIndexingWorker do
+  @moduledoc false
+
+  use Oban.Worker,
+    queue: :fast_indexing,
+    unique: [period: :infinity, states: [:available, :scheduled, :retryable]],
+    tags: ["media_source", "fast_indexing"]
+
+  alias __MODULE__
+  alias Pinchflat.Tasks
+  alias Pinchflat.Sources
+  alias Pinchflat.Sources.Source
+  alias Pinchflat.Tasks.SourceTasks
+
+  @impl Oban.Worker
+  @doc """
+  Kicks off the fast indexing process for a source, reschedules the job to run again
+  once complete. See `MediaCollectionIndexingWorker` and `MediaIndexingWorker` comments
+  for more
+
+  Returns :ok | {:ok, :job_exists} | {:ok, %Task{}}
+  """
+  def perform(%Oban.Job{args: %{"id" => source_id}}) do
+    source = Sources.get_source!(source_id)
+
+    if source.fast_index do
+      SourceTasks.kickoff_indexing_tasks_from_youtube_rss_feed(source)
+
+      reschedule_indexing(source)
+    else
+      :ok
+    end
+  end
+
+  defp reschedule_indexing(source) do
+    next_run_in = Source.fast_index_frequency() * 60
+
+    %{id: source.id}
+    |> FastIndexingWorker.new(schedule_in: next_run_in)
+    |> Tasks.create_job_with_task(source)
+    |> case do
+      {:ok, task} -> {:ok, task}
+      {:error, :duplicate_job} -> {:ok, :job_exists}
+    end
+  end
+end

lib/pinchflat/workers/media_collection_indexing_worker.ex

@@ -0,0 +1,109 @@
+defmodule Pinchflat.Workers.MediaCollectionIndexingWorker do
+  @moduledoc false
+
+  use Oban.Worker,
+    queue: :media_collection_indexing,
+    unique: [period: :infinity, states: [:available, :scheduled, :retryable]],
+    tags: ["media_source", "media_collection_indexing"]
+
+  alias __MODULE__
+  alias Pinchflat.Tasks
+  alias Pinchflat.Sources
+  alias Pinchflat.Sources.Source
+  alias Pinchflat.Tasks.SourceTasks
+  alias Pinchflat.Workers.FastIndexingWorker
+
+  @impl Oban.Worker
+  @doc """
+  The ID is that of a source _record_, not a YouTube channel/playlist ID. Indexes
+  the provided source, kicks off downloads for each new MediaItem, and
+  reschedules the job to run again in the future. It will ALWAYS index a source
+  if it's never been indexed before, but rescheduling is determined by the
+  `index_frequency_minutes` field.
+
+  README: Re-scheduling here works a little different than you may expect.
+  The reschedule time is relative to the time the job has actually _completed_.
+  This has some benefits but also side effects to be aware of:
+
+  - Benefit: No chance for jobs to overlap if a job takes longer than the
+    scheduled interval. Less likely to hit API rate limits.
+  - Side effect: Intervals are "soft" and _always_ walk forward. This may cause
+    user confusion since a 30-minute job scheduled for every hour will
+    actually run every 1 hour and 30 minutes. The tradeoff of not inundating
+    the API with requests and also not overlapping jobs is worth it, IMO.
+
+  Order of operations:
+    1. The user saves a source
+    2. This job is automatically scheduled immediately. This happens in all cases.
+    3. This job indexes all content for the given source. A download job is
+       enqueued for each media item that should be downloaded. This can be impacted
+       by the `download_media` field on the source as well as the profile's
+       shorts/livestream behaviour. At this step we also attach a file reader
+       to the `yt-dlp` output file so we can create media items as they come in
+       for a little speedup (see SourceTasks comments for more)
+    4. If this job is meant to reschedule (ie: has an index frequency > 0),
+       it reschedules itself. If not, it runs once and does not reschedule
+    5. If the source uses fast indexing, that job is kicked off as well. It
+       uses RSS to run a smaller, faster, and more frequent index. That job
+       handles rescheduling itself but largely has a similar behaviour to this
+       job in that it kicks off index and maybe download jobs. The biggest difference
+       is that an index job is kicked off _for each new media item_ as opposed
+       to one larger index job. Check out `MediaIndexingWorker` comments for more.
+    6. If the job reschedules, the cycle from step 3 repeats until the heat death
+       of the universe. The user changing things like the index frequency can
+       dequeue or reschedule jobs as well
+
+  NOTE: Since indexing can take a LONG time, I should check what happens if an
+  application restart occurs while a job is running. Will the job be lost?
+
+  IDEA: Should I use paging and do indexing in chunks? Is that even faster?
+
+  Returns :ok | {:ok, %Task{}}
+  """
+  def perform(%Oban.Job{args: %{"id" => source_id}}) do
+    source = Sources.get_source!(source_id)
+
+    case {source.index_frequency_minutes, source.last_indexed_at} do
+      {index_freq, _} when index_freq > 0 ->
+        # If the indexing is on a schedule simply run indexing and reschedule
+        SourceTasks.index_and_enqueue_download_for_media_items(source)
+        maybe_enqueue_fast_indexing_task(source)
+        reschedule_indexing(source)
+
+      {_, nil} ->
+        # If the source has never been indexed, index it once
+        # even if it's not meant to reschedule
+        SourceTasks.index_and_enqueue_download_for_media_items(source)
+        :ok
+
+      _ ->
+        # If the source HAS been indexed and is not meant to reschedule,
+        # perform a no-op
+        :ok
+    end
+  end
+
+  defp reschedule_indexing(source) do
+    next_run_in = source.index_frequency_minutes * 60
+
+    %{id: source.id}
+    |> MediaCollectionIndexingWorker.new(schedule_in: next_run_in)
+    |> Tasks.create_job_with_task(source)
+    |> case do
+      {:ok, task} -> {:ok, task}
+      {:error, :duplicate_job} -> {:ok, :job_exists}
+    end
+  end
+
+  defp maybe_enqueue_fast_indexing_task(source) do
+    if source.fast_index do
+      Tasks.delete_pending_tasks_for(source, "FastIndexingWorker")
+
+      next_run_in = Source.fast_index_frequency() * 60
+
+      %{id: source.id}
+      |> FastIndexingWorker.new(schedule_in: next_run_in)
+      |> Tasks.create_job_with_task(source)
+    end
+  end
+end

lib/pinchflat/workers/media_download_worker.ex

@@ -45,8 +45,7 @@ defmodule Pinchflat.Workers.MediaDownloadWorker do
   end
 
   defp schedule_filesystem_data_worker(media_item) do
-    media_item
+    %{id: media_item.id}
-    |> Map.take([:id])
     |> FilesystemDataWorker.new()
     |> Tasks.create_job_with_task(media_item)
     |> case do

lib/pinchflat/workers/media_indexing_worker.ex

@@ -6,67 +6,48 @@ defmodule Pinchflat.Workers.MediaIndexingWorker do
     unique: [period: :infinity, states: [:available, :scheduled, :retryable]],
     tags: ["media_source", "media_indexing"]
 
-  alias __MODULE__
+  require Logger
-  alias Pinchflat.Tasks
+
   alias Pinchflat.Sources
-  alias Pinchflat.Tasks.SourceTasks
+  alias Pinchflat.Tasks.MediaItemTasks
 
   @impl Oban.Worker
   @doc """
-  The ID is that of a source _record_, not a YouTube channel/playlist ID. Indexes
+  Similar to `MediaCollectionIndexingWorker`, but for individual media items.
-  the provided source, kicks off downloads for each new MediaItem, and
+  Does not reschedule or check anything to do with a source's indexing
-  reschedules the job to run again in the future. It will ALWAYS index a source
+  frequency - only collects initial metadata then kicks off a download.
-  if it's never been indexed before, but rescheduling is determined by the
+  `MediaCollectionIndexingWorker` should be preferred in general, but this is
-  `index_frequency_minutes` field.
+  useful for downloading one-off media items based on a URL (like for fast indexing).
 
-  README: Re-scheduling here works a little different than you may expect.
+  Only downloads media that _should_ be downloaded (ie: the source is set to download
-  The reschedule time is relative to the time the job has actually _completed_.
+  and the media matches the profile's format preferences)
-  This has some benefits but also side effects to be aware of:
 
-  - Benefit: No chance for jobs to overlap if a job takes longer than the
+  Order of operations:
-    scheduled interval. Less likely to hit API rate limits.
+    1. SourceTasks.kickoff_indexing_tasks_from_youtube_rss_feed/1 (which is running
-  - Side effect: Intervals are "soft" and _always_ walk forward. This may cause
+       in its own worker) periodically checks the YouTube RSS feed for new media
-    user confusion since a 30-minute job scheduled for every hour will
+    2. If new media is found, it enqueues a MediaIndexingWorker (this module) for each new media
-    actually run every 1 hour and 30 minutes. The tradeoff of not inundating
+       item
-    the API with requests and also not overlapping jobs is worth it, IMO.
+    3. This worker fetches the media metadata and uses that to determine if it should be
+       downloaded. If so, it enqueues a MediaDownloadWorker
 
-  NOTE: Since indexing can take a LONG time, I should check what happens if an
+  Each is a worker because they all either need to be scheduled periodically or call out to
-  application restart occurs while a job is running. Will the job be lost?
+  an external service and will be long-running. They're split into different jobs to separate
+  retry logic for each step and allow us to better optimize various queues (eg: the indexing
+  steps can keep running while the slow download steps are worked through).
 
-  IDEA: Should I use paging and do indexing in chunks? Is that even faster?
+  Returns :ok
-
-  Returns :ok | {:ok, %Task{}}
   """
-  def perform(%Oban.Job{args: %{"id" => source_id}}) do
+  def perform(%Oban.Job{args: %{"id" => source_id, "media_url" => media_url}}) do
     source = Sources.get_source!(source_id)
 
-    case {source.index_frequency_minutes, source.last_indexed_at} do
+    case MediaItemTasks.index_and_enqueue_download_for_media_item(source, media_url) do
-      {index_freq, _} when index_freq > 0 ->
+      {:ok, media_item} ->
-        # If the indexing is on a schedule simply run indexing and reschedule
+        Logger.debug("Indexed and enqueued download for url: #{media_url} (media item: #{media_item.id})")
-        SourceTasks.index_and_enqueue_download_for_media_items(source)
-        reschedule_indexing(source)
 
-      {_, nil} ->
+      {:error, reason} ->
-        # If the source has never been indexed, index it once
+        Logger.debug("Failed to index and enqueue download for url: #{media_url} (reason: #{inspect(reason)})")
-        # even if it's not meant to reschedule
-        SourceTasks.index_and_enqueue_download_for_media_items(source)
-        :ok
-
-      _ ->
-        # If the source HAS been indexed and is not meant to reschedule,
-        # perform a no-op
-        :ok
     end
-  end
 
-  defp reschedule_indexing(source) do
+    :ok
-    source
-    |> Map.take([:id])
-    |> MediaIndexingWorker.new(schedule_in: source.index_frequency_minutes * 60)
-    |> Tasks.create_job_with_task(source)
-    |> case do
-      {:ok, task} -> {:ok, task}
-      {:error, :duplicate_job} -> {:ok, :job_exists}
-    end
   end
 end

lib/pinchflat/yt_dlp/backend/backend_command_runner.ex

@@ -1,6 +1,9 @@
-defmodule Pinchflat.MediaClient.Backends.BackendCommandRunner do
+defmodule Pinchflat.YtDlp.Backend.BackendCommandRunner do
   @moduledoc """
-  A behaviour for running CLI commands against a downloader backend
+  A behaviour for running CLI commands against a downloader backend (yt-dlp).
+
+  Used so we can implement Mox for testing without actually running the
+  yt-dlp command.
   """
 
   @callback run(binary(), keyword(), binary()) :: {:ok, binary()} | {:error, binary(), integer()}

lib/pinchflat/yt_dlp/backend/command_runner.ex

@@ -1,4 +1,4 @@
-defmodule Pinchflat.MediaClient.Backends.YtDlp.CommandRunner do
+defmodule Pinchflat.YtDlp.Backend.CommandRunner do
   @moduledoc """
   Runs yt-dlp commands using the `System.cmd/3` function
   """
@@ -7,7 +7,7 @@ defmodule Pinchflat.MediaClient.Backends.YtDlp.CommandRunner do
 
   alias Pinchflat.Utils.StringUtils
   alias Pinchflat.Utils.FilesystemUtils, as: FSUtils
-  alias Pinchflat.MediaClient.Backends.BackendCommandRunner
+  alias Pinchflat.YtDlp.Backend.BackendCommandRunner
 
   @behaviour BackendCommandRunner
 
@@ -25,6 +25,7 @@ defmodule Pinchflat.MediaClient.Backends.YtDlp.CommandRunner do
   """
   @impl BackendCommandRunner
   def run(url, command_opts, output_template, addl_opts \\ []) do
+    # This approach lets us mock the command for testing
     command = backend_executable()
     # These must stay in exactly this order, hence why I'm giving it its own variable.
     # Also, can't use RAM file since yt-dlp needs a concrete filepath.

lib/pinchflat/yt_dlp/backend/media.ex

@@ -0,0 +1,116 @@
+defmodule Pinchflat.YtDlp.Backend.Media do
+  @moduledoc """
+  Contains utilities for working with singular pieces of media
+  """
+
+  @enforce_keys [
+    :media_id,
+    :title,
+    :description,
+    :original_url,
+    :livestream,
+    :short_form_content,
+    :upload_date
+  ]
+
+  defstruct [
+    :media_id,
+    :title,
+    :description,
+    :original_url,
+    :livestream,
+    :short_form_content,
+    :upload_date
+  ]
+
+  alias __MODULE__
+  alias Pinchflat.Utils.FunctionUtils
+
+  @doc """
+  Downloads a single piece of media (and possibly its metadata) directly to its
+  final destination. Returns the parsed JSON output from yt-dlp.
+
+  Returns {:ok, map()} | {:error, any, ...}.
+  """
+  def download(url, command_opts \\ []) do
+    opts = [:no_simulate] ++ command_opts
+
+    with {:ok, output} <- backend_runner().run(url, opts, "after_move:%()j"),
+         {:ok, parsed_json} <- Phoenix.json_library().decode(output) do
+      {:ok, parsed_json}
+    else
+      err -> err
+    end
+  end
+
+  @doc """
+  Returns a map representing the media at the given URL.
+
+  Returns {:ok, [map()]} | {:error, any, ...}.
+  """
+  def get_media_attributes(url) do
+    runner = Application.get_env(:pinchflat, :yt_dlp_runner)
+    command_opts = [:simulate, :skip_download]
+    output_template = indexing_output_template()
+
+    case runner.run(url, command_opts, output_template) do
+      {:ok, output} ->
+        output
+        |> Phoenix.json_library().decode!()
+        |> response_to_struct()
+        |> FunctionUtils.wrap_ok()
+
+      res ->
+        res
+    end
+  end
+
+  @doc """
+  Returns the output template for yt-dlp's indexing command.
+  """
+  def indexing_output_template do
+    "%(.{id,title,was_live,webpage_url,description,aspect_ratio,duration,upload_date})j"
+  end
+
+  @doc """
+  Transforms a response from yt-dlp into a struct. Interprets the response to
+  determine if the media is short-form content.
+
+  Returns %Media{}.
+  """
+  def response_to_struct(response) do
+    %Media{
+      media_id: response["id"],
+      title: response["title"],
+      description: response["description"],
+      original_url: response["webpage_url"],
+      livestream: response["was_live"],
+      short_form_content: short_form_content?(response),
+      upload_date: parse_upload_date(response["upload_date"])
+    }
+  end
+
+  defp short_form_content?(response) do
+    if String.contains?(response["webpage_url"], "/shorts/") do
+      true
+    else
+      # Sometimes shorts are returned without /shorts/ in the URL,
+      # so we need to do our best to determine if it's a short. This
+      # WILL returns false positives, but it's a best-effort approach
+      # that should work for most cases. The aspect_ratio check is
+      # based on a gut feeling and may need to be tweaked.
+      response["duration"] <= 60 && response["aspect_ratio"] < 0.8
+    end
+  end
+
+  defp parse_upload_date(upload_date) do
+    <<year::binary-size(4)>> <> <<month::binary-size(2)>> <> <<day::binary-size(2)>> = upload_date
+
+    Date.from_iso8601!("#{year}-#{month}-#{day}")
+  end
+
+  defp backend_runner do
+    # This approach lets us mock the command for testing
+    Application.get_env(:pinchflat, :yt_dlp_runner)
+  end
+end

lib/pinchflat/yt_dlp/backend/media_collection.ex

@@ -1,4 +1,4 @@
-defmodule Pinchflat.MediaClient.Backends.YtDlp.MediaCollection do
+defmodule Pinchflat.YtDlp.Backend.MediaCollection do
   @moduledoc """
   Contains utilities for working with collections of
   media (aka: a source [ie: channels, playlists]).
@@ -8,6 +8,7 @@ defmodule Pinchflat.MediaClient.Backends.YtDlp.MediaCollection do
 
   alias Pinchflat.Utils.FunctionUtils
   alias Pinchflat.Utils.FilesystemUtils
+  alias Pinchflat.YtDlp.Backend.Media, as: YtDlpMedia
 
   @doc """
   Returns a list of maps representing the media in the collection.
@@ -19,10 +20,10 @@ defmodule Pinchflat.MediaClient.Backends.YtDlp.MediaCollection do
 
   Returns {:ok, [map()]} | {:error, any, ...}.
   """
-  def get_media_attributes(url, addl_opts \\ []) do
+  def get_media_attributes_for_collection(url, addl_opts \\ []) do
     runner = Application.get_env(:pinchflat, :yt_dlp_runner)
     command_opts = [:simulate, :skip_download]
-    output_template = "%(.{id,title,was_live,original_url,description})j"
+    output_template = YtDlpMedia.indexing_output_template()
     output_filepath = FilesystemUtils.generate_metadata_tmpfile(:json)
     file_listener_handler = Keyword.get(addl_opts, :file_listener_handler, false)
 
@@ -35,6 +36,7 @@ defmodule Pinchflat.MediaClient.Backends.YtDlp.MediaCollection do
         output
         |> String.split("\n", trim: true)
         |> Enum.map(&Phoenix.json_library().decode!/1)
+        |> Enum.map(&YtDlpMedia.response_to_struct/1)
         |> FunctionUtils.wrap_ok()
 
       res ->
@@ -73,6 +75,7 @@ defmodule Pinchflat.MediaClient.Backends.YtDlp.MediaCollection do
   end
 
   defp backend_runner do
+    # This approach lets us mock the command for testing
     Application.get_env(:pinchflat, :yt_dlp_runner)
   end
 end

lib/pinchflat/yt_dlp/download_option_builder.ex

@@ -1,12 +1,10 @@
-defmodule Pinchflat.Profiles.Options.YtDlp.DownloadOptionBuilder do
+defmodule Pinchflat.YtDlp.DownloadOptionBuilder do
   @moduledoc """
   Builds the options for yt-dlp to download media based on the given media profile.
-
-  IDEA: consider making this a behaviour so I can add other backends later
   """
 
   alias Pinchflat.Media.MediaItem
-  alias Pinchflat.Profiles.Options.YtDlp.OutputPathBuilder
+  alias Pinchflat.Profiles.OutputPathBuilder
 
   @doc """
   Builds the options for yt-dlp to download media based on the given media's profile.
@@ -63,11 +61,11 @@ defmodule Pinchflat.Profiles.Options.YtDlp.DownloadOptionBuilder do
     mapped_struct = Map.from_struct(media_profile)
 
     Enum.reduce(mapped_struct, [], fn attr, acc ->
-      case {attr, media_profile} do
+      case attr do
-        {{:download_thumbnail, true}, _} ->
+        {:download_thumbnail, true} ->
-          acc ++ [:write_thumbnail]
+          acc ++ [:write_thumbnail, convert_thumbnail: "jpg"]
 
-        {{:embed_thumbnail, true}, %{preferred_resolution: pr}} when pr != :audio ->
+        {:embed_thumbnail, true} ->
           acc ++ [:embed_thumbnail]
 
         _ ->
@@ -80,31 +78,26 @@ defmodule Pinchflat.Profiles.Options.YtDlp.DownloadOptionBuilder do
     mapped_struct = Map.from_struct(media_profile)
 
     Enum.reduce(mapped_struct, [], fn attr, acc ->
-      case {attr, media_profile} do
+      case attr do
-        {{:download_metadata, true}, _} ->
+        {:download_metadata, true} -> acc ++ [:write_info_json, :clean_info_json]
-          acc ++ [:write_info_json, :clean_info_json]
+        {:embed_metadata, true} -> acc ++ [:embed_metadata]
-
+        _ -> acc
-        {{:embed_metadata, true}, %{preferred_resolution: pr}} when pr != :audio ->
-          acc ++ [:embed_metadata]
-
-        _ ->
-          acc
       end
     end)
   end
 
   defp quality_options(media_profile) do
-    codec_options = "+codec:avc:m4a"
+    video_codec_options = "+codec:avc:m4a"
 
     case media_profile.preferred_resolution do
-      # Also be aware that :audio disabled all embedding options for thumbnails, subtitles, and metadata
+      # Also be aware that :audio disabled all embedding options for subtitles
-      :audio -> [format_sort: "ext", format: "bestaudio"]
+      :audio -> [:extract_audio, format: "bestaudio[ext=m4a]"]
-      :"360p" -> [format_sort: "res:360,#{codec_options}"]
+      :"360p" -> [format_sort: "res:360,#{video_codec_options}"]
-      :"480p" -> [format_sort: "res:480,#{codec_options}"]
+      :"480p" -> [format_sort: "res:480,#{video_codec_options}"]
-      :"720p" -> [format_sort: "res:720,#{codec_options}"]
+      :"720p" -> [format_sort: "res:720,#{video_codec_options}"]
-      :"1080p" -> [format_sort: "res:1080,#{codec_options}"]
+      :"1080p" -> [format_sort: "res:1080,#{video_codec_options}"]
-      :"1440p" -> [format_sort: "res:1440,#{codec_options}"]
+      :"1440p" -> [format_sort: "res:1440,#{video_codec_options}"]
-      :"2160p" -> [format_sort: "res:2160,#{codec_options}"]
+      :"2160p" -> [format_sort: "res:2160,#{video_codec_options}"]
     end
   end
 

lib/pinchflat/yt_dlp/media_downloader.ex

@@ -1,25 +1,22 @@
 defmodule Pinchflat.MediaClient.MediaDownloader do
   @moduledoc """
-  This is the integration layer for actually downloading medias.
+  This is the integration layer for actually downloading media.
   It takes into account the media profile's settings in order
   to download the media with the desired options.
-
-  Technically hardcodes the yt-dlp backend for now, but should leave
-  it open-ish for future expansion (just in case).
   """
 
   alias Pinchflat.Repo
   alias Pinchflat.Media
   alias Pinchflat.Media.MediaItem
 
-  alias Pinchflat.MediaClient.Backends.YtDlp.Media, as: YtDlpMedia
+  alias Pinchflat.YtDlp.Backend.Media, as: YtDlpMedia
-  alias Pinchflat.Profiles.Options.YtDlp.DownloadOptionBuilder, as: YtDlpDownloadOptionBuilder
+  alias Pinchflat.YtDlp.DownloadOptionBuilder, as: YtDlpDownloadOptionBuilder
-  alias Pinchflat.MediaClient.Backends.YtDlp.MetadataParser, as: YtDlpMetadataParser
+  alias Pinchflat.Metadata.MetadataParser, as: YtDlpMetadataParser
-  alias Pinchflat.MediaClient.Backends.YtDlp.MetadataFileHelpers, as: YtDlpMetadataHelpers
+  alias Pinchflat.Metadata.MetadataFileHelpers, as: YtDlpMetadataHelpers
 
   @doc """
   Downloads media for a media item, updating the media item based on the metadata
-  returned by the backend. Also saves the entire metadata response to the associated
+  returned by yt-dlp. Also saves the entire metadata response to the associated
   media_metadata record.
 
   NOTE: related methods (like the download worker) won't download if the media item's source
@@ -28,12 +25,12 @@ defmodule Pinchflat.MediaClient.MediaDownloader do
 
   Returns {:ok, %MediaItem{}} | {:error, any, ...any}
   """
-  def download_for_media_item(%MediaItem{} = media_item, backend \\ :yt_dlp) do
+  def download_for_media_item(%MediaItem{} = media_item) do
     item_with_preloads = Repo.preload(media_item, [:metadata, source: :media_profile])
 
-    case download_with_options(media_item.original_url, item_with_preloads, backend) do
+    case download_with_options(media_item.original_url, item_with_preloads) do
       {:ok, parsed_json} ->
-        {parser, helpers} = metadata_parsers(backend)
+        {parser, helpers} = {YtDlpMetadataParser, YtDlpMetadataHelpers}
 
         parsed_attrs =
           parsed_json
@@ -55,29 +52,14 @@ defmodule Pinchflat.MediaClient.MediaDownloader do
     end
   end
 
-  defp download_with_options(url, item_with_preloads, backend) do
+  # def download_for_source(source, url) do
-    option_builder = option_builder(backend)
+  #   # Create MI from source and URL
-    media_backend = media_backend(backend)
+  #   media_item = nil
-    {:ok, options} = option_builder.build(item_with_preloads)
+  # end
 
-    media_backend.download(url, options)
+  defp download_with_options(url, item_with_preloads) do
-  end
+    {:ok, options} = YtDlpDownloadOptionBuilder.build(item_with_preloads)
 
-  defp option_builder(backend) do
+    YtDlpMedia.download(url, options)
-    case backend do
-      :yt_dlp -> YtDlpDownloadOptionBuilder
-    end
-  end
-
-  defp media_backend(backend) do
-    case backend do
-      :yt_dlp -> YtDlpMedia
-    end
-  end
-
-  defp metadata_parsers(backend) do
-    case backend do
-      :yt_dlp -> {YtDlpMetadataParser, YtDlpMetadataHelpers}
-    end
   end
 end

lib/pinchflat_web.ex

@@ -54,8 +54,9 @@ defmodule PinchflatWeb do
 
   def live_view do
     quote do
-      use Phoenix.LiveView,
+      use Phoenix.Component, global_prefixes: ~w(x-)
-        layout: {PinchflatWeb.Layouts, :app}
+
+      use Phoenix.LiveView
 
       alias Pinchflat.Settings
 
@@ -75,7 +76,7 @@ defmodule PinchflatWeb do
 
   def html do
     quote do
-      use Phoenix.Component
+      use Phoenix.Component, global_prefixes: ~w(x-)
 
       # Import convenience functions from controllers
       import Phoenix.Controller,

lib/pinchflat_web/components/core_components.ex

@@ -40,6 +40,7 @@ defmodule PinchflatWeb.CoreComponents do
   """
   attr :id, :string, required: true
   attr :show, :boolean, default: false
+  attr :allow_close, :boolean, default: true
   attr :on_cancel, JS, default: %JS{}
   slot :inner_block, required: true
 
@@ -50,9 +51,9 @@ defmodule PinchflatWeb.CoreComponents do
       phx-mounted={@show && show_modal(@id)}
       phx-remove={hide_modal(@id)}
       data-cancel={JS.exec(@on_cancel, "phx-remove")}
-      class="relative z-50 hidden"
+      class="relative z-99999 hidden"
     >
-      <div id={"#{@id}-bg"} class="bg-zinc-50/90 fixed inset-0 transition-opacity" aria-hidden="true" />
+      <div id={"#{@id}-bg"} class="bg-black-2/80 fixed inset-0 transition-opacity" aria-hidden="true" />
       <div
         class="fixed inset-0 overflow-y-auto"
         aria-labelledby={"#{@id}-title"}
@@ -62,28 +63,28 @@ defmodule PinchflatWeb.CoreComponents do
         tabindex="0"
       >
         <div class="flex min-h-full items-center justify-center">
-          <div class="w-full max-w-3xl p-4 sm:p-6 lg:py-8">
+          <div class="w-full max-w-3xl p-2 sm:p-6 lg:py-8">
-            <.focus_wrap
+            <div
               id={"#{@id}-container"}
-              phx-window-keydown={JS.exec("data-cancel", to: "##{@id}")}
+              phx-window-keydown={@allow_close && JS.exec("data-cancel", to: "##{@id}")}
               phx-key="escape"
-              phx-click-away={JS.exec("data-cancel", to: "##{@id}")}
+              phx-click-away={@allow_close && JS.exec("data-cancel", to: "##{@id}")}
-              class="shadow-zinc-700/10 ring-zinc-700/10 relative hidden rounded-2xl bg-white p-14 shadow-lg ring-1 transition"
+              class="shadow-zinc-700/10 ring-zinc-700/10 relative hidden rounded-2xl bg-graydark p-8 sm:p-14 shadow-lg ring-1 transition"
             >
-              <div class="absolute top-6 right-5">
+              <div :if={@allow_close} class="absolute top-6 right-5">
                 <button
                   phx-click={JS.exec("data-cancel", to: "##{@id}")}
                   type="button"
-                  class="-m-3 flex-none p-3 opacity-20 hover:opacity-40"
+                  class="-m-3 flex-none p-3 opacity-60 hover:opacity-80"
                   aria-label={gettext("close")}
                 >
-                  <.icon name="hero-x-mark-solid" class="h-5 w-5" />
+                  <.icon name="hero-x-mark-solid" class="h-5 w-5 text-white" />
                 </button>
               </div>
               <div id={"#{@id}-content"}>
                 <%= render_slot(@inner_block) %>
               </div>
-            </.focus_wrap>
+            </div>
           </div>
         </div>
       </div>
@@ -243,6 +244,7 @@ defmodule PinchflatWeb.CoreComponents do
   attr :id, :any, default: nil
   attr :name, :any
   attr :label, :string, default: nil
+  attr :label_suffix, :string, default: nil
   attr :value, :any
   attr :help, :string, default: nil
 
@@ -294,6 +296,7 @@ defmodule PinchflatWeb.CoreComponents do
           {@rest}
         />
         <%= @label %>
+        <span :if={@label_suffix} class="text-xs text-bodydark"><%= @label_suffix %></span>
       </label>
       <.help :if={@help}><%= @help %></.help>
       <.error :for={msg <- @errors}><%= msg %></.error>
@@ -309,7 +312,10 @@ defmodule PinchflatWeb.CoreComponents do
 
     ~H"""
     <div x-data={"{ enabled: #{@checked}}"}>
-      <.label for={@id}><%= @label %></.label>
+      <.label for={@id}>
+        <%= @label %>
+        <span :if={@label_suffix} class="text-xs text-bodydark"><%= @label_suffix %></span>
+      </.label>
       <div class="relative">
         <input type="hidden" name={@name} value="false" />
         <input
@@ -343,7 +349,9 @@ defmodule PinchflatWeb.CoreComponents do
   def input(%{type: "select"} = assigns) do
     ~H"""
     <div phx-feedback-for={@name}>
-      <.label for={@id}><%= @label %></.label>
+      <.label for={@id}>
+        <%= @label %><span :if={@label_suffix} class="text-xs text-bodydark"><%= @label_suffix %></span>
+      </.label>
       <select
         id={@id}
         name={@name}
@@ -367,7 +375,9 @@ defmodule PinchflatWeb.CoreComponents do
   def input(%{type: "textarea"} = assigns) do
     ~H"""
     <div phx-feedback-for={@name}>
-      <.label for={@id}><%= @label %></.label>
+      <.label for={@id}>
+        <%= @label %><span :if={@label_suffix} class="text-xs text-bodydark"><%= @label_suffix %></span>
+      </.label>
       <textarea
         id={@id}
         name={@name}
@@ -390,7 +400,9 @@ defmodule PinchflatWeb.CoreComponents do
   def input(assigns) do
     ~H"""
     <div phx-feedback-for={@name}>
-      <.label for={@id}><%= @label %></.label>
+      <.label for={@id}>
+        <%= @label %><span :if={@label_suffix} class="text-xs text-bodydark"><%= @label_suffix %></span>
+      </.label>
       <input
         type={@type}
         name={@name}
@@ -586,9 +598,14 @@ defmodule PinchflatWeb.CoreComponents do
   def list_items_from_map(assigns) do
     attrs =
       Enum.filter(assigns.map, fn
-        {_, %{__struct__: _}} -> false
+        {_, %{__struct__: s}} when s not in [Date, DateTime] ->
-        {_, [%{__meta__: _} | _]} -> false
+          false
-        _ -> true
+
+        {_, [%{__meta__: _} | _]} ->
+          false
+
+        _ ->
+          true
       end)
 
     assigns = assign(assigns, iterable_attributes: attrs)
@@ -608,15 +625,15 @@ defmodule PinchflatWeb.CoreComponents do
 
   ## Examples
 
-      <.back navigate={~p"/posts"}>Back to posts</.back>
+      <.back href={~p"/posts"}>Back to posts</.back>
   """
-  attr :navigate, :any, required: true
+  attr :href, :any, required: true
   slot :inner_block, required: true
 
   def back(assigns) do
     ~H"""
     <div class="mt-16">
-      <.link navigate={@navigate} class="text-sm font-semibold leading-6 text-zinc-900 hover:text-zinc-700">
+      <.link href={@href} class="text-sm font-semibold leading-6 text-zinc-900 hover:text-zinc-700">
         <.icon name="hero-arrow-left-solid" class="h-3 w-3" />
         <%= render_slot(@inner_block) %>
       </.link>
@@ -685,7 +702,6 @@ defmodule PinchflatWeb.CoreComponents do
     )
     |> show("##{id}-container")
     |> JS.add_class("overflow-hidden", to: "body")
-    |> JS.focus_first(to: "##{id}-content")
   end
 
   def hide_modal(js \\ %JS{}, id) do
@@ -697,7 +713,6 @@ defmodule PinchflatWeb.CoreComponents do
     |> hide("##{id}-container")
     |> JS.hide(to: "##{id}", transition: {"block", "block", "hidden"})
     |> JS.remove_class("overflow-hidden", to: "body")
-    |> JS.pop_focus()
   end
 
   @doc """

lib/pinchflat_web/components/custom_components/button_components.ex

@@ -14,7 +14,9 @@ defmodule PinchflatWeb.CustomComponents.ButtonComponents do
   attr :color, :string, default: "bg-primary"
   attr :rounding, :string, default: "rounded-sm"
   attr :class, :string, default: ""
+  attr :type, :string, default: "submit"
   attr :disabled, :boolean, default: false
+  attr :rest, :global
 
   slot :inner_block, required: true
 
@@ -26,9 +28,11 @@ defmodule PinchflatWeb.CustomComponents.ButtonComponents do
         "#{@rounding} inline-flex items-center justify-center px-8 py-4",
         "#{@color}",
         "hover:bg-opacity-90 lg:px-8 xl:px-10",
+        "disabled:bg-opacity-50 disabled:cursor-not-allowed disabled:text-gray-2",
         @class
       ]}
       disabled={@disabled}
+      {@rest}
     >
       <%= render_slot(@inner_block) %>
     </button>

lib/pinchflat_web/components/layouts.ex

@@ -6,14 +6,16 @@ defmodule PinchflatWeb.Layouts do
 
   attr :icon, :string, required: true
   attr :text, :string, required: true
-  attr :navigate, :any, required: true
+  attr :href, :any, required: true
+  attr :target, :any, default: "_self"
 
   def sidebar_item(assigns) do
     # I'm testing out grouping classes here. Tentative order: font, layout, color, animation, state-modifiers
     ~H"""
     <li>
       <.link
-        navigate={@navigate}
+        href={@href}
+        target={@target}
         class={[
           "font-medium text-bodydark1",
           "group relative flex items-center gap-2.5 rounded-sm px-4 py-2 duration-300 ease-in-out",

lib/pinchflat_web/components/layouts/partials/donate_modal.heex

@@ -0,0 +1,26 @@
+<.modal id="donate-modal" allow_close={true}>
+  <section x-data="{ inputValue: '' }">
+    <h3 class="text-2xl text-white">Donate</h3>
+    <p class="text-sm">Thank you for your support :&rpar;</p>
+    <p class="mt-4">
+      If you find the project valuable and want to support its development, a
+      <.inline_link href="https://www.paypal.me/kieraneglin">donation</.inline_link>
+      would be greatly appreciated.
+    </p>
+    <p class="mt-4">
+      Plus, $5 USD from any donation over $10 USD will be donated to
+      <.inline_link href="https://www.eff.org/">The Electronic Frontier Foundation</.inline_link>
+      who defend your online liberties and backed
+      <.inline_code>youtube-dl</.inline_code>
+      when Google took them down<.inline_link href="https://github.com/github/dmca/blob/9a85e0f021f7967af80e186b890776a50443f06c/2020/11/2020-11-16-RIAA-reversal-effletter.pdf">
+        <.icon name="hero-arrow-top-right-on-square" class="h-3 w-3" />
+      </.inline_link>.
+    </p>
+
+    <.link href="https://www.paypal.me/kieraneglin" target="_blank">
+      <.button color="bg-primary" class="w-full mt-8">
+        Donate
+      </.button>
+    </.link>
+  </section>
+</.modal>

lib/pinchflat_web/components/layouts/partials/sidebar.html.heex

@@ -1,31 +1,58 @@
 <aside
   x-bind:class="sidebarVisible ? 'translate-x-0' : '-translate-x-full'"
   class={[
-    "-translate-x-full absolute left-0 top-0 z-9999 flex h-screen w-60 flex-col overflow-y-hidden",
+    "-translate-x-full absolute left-0 top-0 z-9999 flex h-screen w-60 flex-col overflow-y-hidden justify-between",
     "bg-black duration-300 ease-linear shadow-lg sm:shadow-none dark:bg-boxdark lg:static lg:translate-x-0"
   ]}
   @click.outside="sidebarVisible = false"
 >
-  <div class="flex items-center justify-between gap-2 px-6 py-5.5 lg:py-6.5">
+  <section>
-    <a href="/" class="flex items-center">
+    <div class="flex items-center justify-between gap-2 px-6 py-5.5 lg:py-6.5">
-      <img src={~p"/images/logo.png?cachebust=2024-02-29"} alt="Pinchflat" class="w-9 h-9" />
+      <a href="/" class="flex items-center">
-      <h2 class="text-xl font-bold text-white pl-2">Pinchflat</h2>
+        <img src={~p"/images/logo.png?cachebust=2024-02-29"} alt="Pinchflat" class="w-9 h-9" />
-    </a>
+        <h2 class="text-xl font-bold text-white pl-2">Pinchflat</h2>
+      </a>
 
-    <button class="block lg:hidden" @click.stop="sidebarVisible = !sidebarVisible">
+      <button class="block lg:hidden" @click.stop="sidebarVisible = !sidebarVisible">
-      <.icon name="hero-arrow-left" class="fill-current" />
+        <.icon name="hero-arrow-left" class="fill-current" />
-    </button>
+      </button>
-  </div>
+    </div>
-  <div class="no-scrollbar flex flex-col overflow-y-auto duration-300 ease-linear">
+    <div class="no-scrollbar flex flex-col overflow-y-auto duration-300 ease-linear">
-    <nav class="mt-5 px-4 py-4 lg:mt-9 lg:px-6">
+      <nav class="mt-5 px-4 py-4 lg:mt-9 lg:px-6">
-      <div>
         <h3 class="mb-4 ml-4 text-sm font-medium text-bodydark2">MENU</h3>
-        <ul class="mb-6 flex flex-col gap-1.5">
+        <div class="flex flex-col justify-between">
-          <.sidebar_item icon="hero-home" text="Home" navigate={~p"/"} />
+          <ul class="mb-6 flex flex-col gap-1.5">
-          <.sidebar_item icon="hero-tv" text="Sources" navigate={~p"/sources"} />
+            <.sidebar_item icon="hero-home" text="Home" href={~p"/"} />
-          <.sidebar_item icon="hero-adjustments-vertical" text="Media Profiles" navigate={~p"/media_profiles"} />
+            <.sidebar_item icon="hero-tv" text="Sources" href={~p"/sources"} />
-        </ul>
+            <.sidebar_item icon="hero-adjustments-vertical" text="Media Profiles" href={~p"/media_profiles"} />
-      </div>
+          </ul>
+        </div>
+      </nav>
+    </div>
+  </section>
+  <section>
+    <nav class="mt-5 px-4 py-4 lg:mt-9 lg:px-6">
+      <ul class="mb-6 flex flex-col gap-1.5">
+        <.sidebar_item
+          icon="hero-code-bracket"
+          text="Github"
+          target="_blank"
+          href="https://github.com/kieraneglin/pinchflat"
+        />
+        <li>
+          <span
+            class={[
+              "font-medium text-bodydark1",
+              "group relative flex items-center gap-2.5 rounded-sm px-4 py-2 duration-300 ease-in-out",
+              "duration-300 ease-in-out cursor-pointer",
+              "hover:bg-graydark dark:hover:bg-meta-4"
+            ]}
+            phx-click={show_modal("donate-modal")}
+          >
+            <.icon name="hero-currency-dollar" /> Donate
+          </span>
+        </li>
+      </ul>
     </nav>
-  </div>
+  </section>
 </aside>

lib/pinchflat_web/components/layouts/partials/upgrade_button_live.ex

@@ -0,0 +1,40 @@
+defmodule Pinchflat.UpgradeButtonLive do
+  use PinchflatWeb, :live_view
+
+  def render(assigns) do
+    ~H"""
+    <form phx-change="check_matching_text">
+      <.input type="text" name="unlock-pro-textbox" value="" />
+    </form>
+
+    <.button
+      class="w-full mt-4"
+      type="button"
+      disabled={@button_disabled}
+      phx-click={hide_modal("upgrade-modal")}
+      x-on:click="setTimeout(() => { proEnabled = true }, 200)"
+    >
+      Unlock Pro
+    </.button>
+    """
+  end
+
+  def mount(_params, _session, socket) do
+    {:ok, assign(socket, :button_disabled, true)}
+  end
+
+  def handle_event("check_matching_text", %{"unlock-pro-textbox" => text}, socket) do
+    normalized_text =
+      text
+      |> String.trim()
+      |> String.downcase()
+
+    if normalized_text == "got it!" do
+      Settings.set!(:pro_enabled, true)
+
+      {:noreply, update(socket, :button_disabled, fn _ -> false end)}
+    else
+      {:noreply, update(socket, :button_disabled, fn _ -> true end)}
+    end
+  end
+end

lib/pinchflat_web/components/layouts/partials/upgrade_modal.heex

@@ -0,0 +1,28 @@
+<.modal id="upgrade-modal" allow_close={false}>
+  <section>
+    <h3 class="text-2xl text-white">Pro Mode</h3>
+    <p class="text-sm">Don't worry - Pinchflat is completely free :&rpar;</p>
+    <p class="mt-4">
+      If you find the project valuable and want to support its development, a
+      <.inline_link href="https://www.paypal.me/kieraneglin">donation</.inline_link>
+      would be greatly appreciated.
+    </p>
+    <p class="mt-4">
+      Plus, $5 USD from any donation over $10 USD will be donated to
+      <.inline_link href="https://www.eff.org/">The Electronic Frontier Foundation</.inline_link>
+      who defend your online liberties and backed
+      <.inline_code>youtube-dl</.inline_code>
+      when Google took them down<.inline_link href="https://github.com/github/dmca/blob/9a85e0f021f7967af80e186b890776a50443f06c/2020/11/2020-11-16-RIAA-reversal-effletter.pdf">
+        <.icon name="hero-arrow-top-right-on-square" class="h-3 w-3" />
+      </.inline_link>. <strong>You do not need to donate to unlock Pro</strong>. It's just a way to say thanks!
+    </p>
+
+    <p class="mt-4">
+      To unlock Pro, simply type
+      <.inline_code>got it!</.inline_code>
+      into the text box and press the button.
+    </p>
+
+    <%= live_render(@conn, Pinchflat.UpgradeButtonLive) %>
+  </section>
+</.modal>

lib/pinchflat_web/components/layouts/root.html.heex

@@ -12,7 +12,19 @@
     <script defer phx-track-static type="text/javascript" src={~p"/assets/app.js"}>
     </script>
   </head>
-  <body x-data="{ sidebarVisible: false }" class="dark text-bodydark bg-boxdark-2">
+  <body
+    x-data={"{ 
+      sidebarVisible: false,
+      proEnabled: #{Settings.get!(:pro_enabled)},
+      onboarding: #{Settings.get!(:onboarding)}
+    }"}
+    class="dark text-bodydark bg-boxdark-2"
+  >
     <%= @inner_content %>
+
+    <.donate_modal conn={@conn} />
+    <template x-if="!proEnabled && !onboarding">
+      <.upgrade_modal conn={@conn} />
+    </template>
   </body>
 </html>

lib/pinchflat_web/controllers/media_items/media_item_html/show.html.heex

@@ -1,6 +1,6 @@
 <div class="mb-6 flex gap-3 flex-row items-center justify-between">
   <div class="flex gap-3 items-center">
-    <.link navigate={~p"/sources/#{@media_item.source_id}"}>
+    <.link href={~p"/sources/#{@media_item.source_id}"}>
       <.icon name="hero-arrow-left" class="w-10 h-10 hover:dark:text-white" />
     </.link>
     <h2 class="text-title-md2 font-bold text-black dark:text-white ml-4">

lib/pinchflat_web/controllers/media_profiles/media_profile_html.ex

@@ -35,6 +35,7 @@ defmodule PinchflatWeb.MediaProfiles.MediaProfileHTML do
       upload_day: nil,
       upload_month: nil,
       upload_year: nil,
+      upload_yyyy_mm_dd: "the upload date in the format YYYY-MM-DD",
       source_custom_name: "the name of the sources that use this profile",
       source_collection_type: "the collection type of the sources that use this profile. Either 'channel' or 'playlist'"
     }

lib/pinchflat_web/controllers/media_profiles/media_profile_html/edit.html.heex

@@ -1,5 +1,5 @@
 <div class="mb-6 flex gap-3 flex-row items-center">
-  <.link navigate={~p"/media_profiles"}>
+  <.link href={~p"/media_profiles"}>
     <.icon name="hero-arrow-left" class="w-10 h-10 hover:dark:text-white" />
   </.link>
   <h2 class="text-title-md2 font-bold text-black dark:text-white ml-4">Edit Media Profile</h2>

lib/pinchflat_web/controllers/media_profiles/media_profile_html/index.html.heex

@@ -3,7 +3,7 @@
     Media Profiles
   </h2>
   <nav>
-    <.link navigate={~p"/media_profiles/new"}>
+    <.link href={~p"/media_profiles/new"}>
       <.button color="bg-primary" rounding="rounded-full">
         <span class="font-bold text-xl mx-2">+</span> New <span class="hidden sm:inline pl-1">Media Profile</span>
       </.button>

lib/pinchflat_web/controllers/media_profiles/media_profile_html/media_profile_form.html.heex

@@ -19,7 +19,7 @@
     type="text"
     inputclass="font-mono"
     label="Output path template"
-    help="Must end with .{{ ext }}. See below for more details. I promise the default is good for most cases (required)"
+    help="Must end with .{{ ext }}. See below for more details. The default is good for most cases (required)"
   />
 
   <h3 class="mt-8 text-2xl text-black dark:text-white">
@@ -109,7 +109,7 @@
     options={friendly_resolution_options()}
     type="select"
     label="Preferred Resolution"
-    help="Will grab the closest available resolution if your preferred is not available. Setting to 'Audio Only' negates embedding options."
+    help="Will grab the closest available resolution if your preferred is not available"
   />
 
   <.button class="my-10 sm:mb-7.5 w-full sm:w-auto">Save Media profile</.button>

lib/pinchflat_web/controllers/media_profiles/media_profile_html/new.html.heex

@@ -1,5 +1,5 @@
 <div class="mb-6 flex gap-3 flex-row items-center">
-  <.link :if={!Settings.get!(:onboarding)} navigate={~p"/media_profiles"}>
+  <.link :if={!Settings.get!(:onboarding)} href={~p"/media_profiles"}>
     <.icon name="hero-arrow-left" class="w-10 h-10 hover:dark:text-white" />
   </.link>
   <h2 class="text-title-md2 font-bold text-black dark:text-white ml-4">New Media Profile</h2>

lib/pinchflat_web/controllers/media_profiles/media_profile_html/show.html.heex

@@ -1,6 +1,6 @@
 <div class="mb-6 flex gap-3 flex-row items-center justify-between">
   <div class="flex items-center">
-    <.link navigate={~p"/media_profiles"}>
+    <.link href={~p"/media_profiles"}>
       <.icon name="hero-arrow-left" class="w-10 h-10 hover:dark:text-white" />
     </.link>
     <h2 class="text-title-md2 font-bold text-black dark:text-white ml-2">
@@ -9,7 +9,7 @@
   </div>
 
   <nav>
-    <.link navigate={~p"/media_profiles/#{@media_profile}/edit"}>
+    <.link href={~p"/media_profiles/#{@media_profile}/edit"}>
       <.button color="bg-primary" rounding="rounded-full">
         <.icon name="hero-pencil-square" class="mr-2" />Edit <span class="hidden sm:inline pl-1">Media Profile</span>
       </.button>

lib/pinchflat_web/controllers/pages/page_controller.ex

@@ -8,38 +8,34 @@ defmodule PinchflatWeb.Pages.PageController do
   alias Pinchflat.Profiles.MediaProfile
 
   def home(conn, params) do
-    force_onboarding = params["onboarding"]
+    done_onboarding = params["onboarding"] == "0"
-    media_profiles_exist = Repo.exists?(MediaProfile)
+    force_onboarding = params["onboarding"] == "1"
-    sources_exist = Repo.exists?(Source)
 
-    if !force_onboarding && media_profiles_exist && sources_exist do
+    if done_onboarding, do: Settings.set!(:onboarding, false)
-      render_home_page(conn)
+
+    if force_onboarding || Settings.get!(:onboarding) do
+      render_onboarding_page(conn)
     else
-      render_onboarding_page(conn, media_profiles_exist, sources_exist)
+      render_home_page(conn)
     end
   end
 
   defp render_home_page(conn) do
-    Settings.set!(:onboarding, false)
-    media_profile_count = Repo.aggregate(MediaProfile, :count, :id)
-    source_count = Repo.aggregate(Source, :count, :id)
-    media_item_count = Repo.aggregate(MediaItem, :count, :id)
-
     conn
     |> render(:home,
-      media_profile_count: media_profile_count,
+      media_profile_count: Repo.aggregate(MediaProfile, :count, :id),
-      source_count: source_count,
+      source_count: Repo.aggregate(Source, :count, :id),
-      media_item_count: media_item_count
+      media_item_count: Repo.aggregate(MediaItem, :count, :id)
     )
   end
 
-  defp render_onboarding_page(conn, media_profiles_exist, sources_exist) do
+  defp render_onboarding_page(conn) do
     Settings.set!(:onboarding, true)
 
     conn
     |> render(:onboarding_checklist,
-      media_profiles_exist: media_profiles_exist,
+      media_profiles_exist: Repo.exists?(MediaProfile),
-      sources_exist: sources_exist,
+      sources_exist: Repo.exists?(Source),
       layout: {Layouts, :onboarding}
     )
   end

lib/pinchflat_web/controllers/pages/page_html/onboarding_checklist.html.heex

@@ -10,7 +10,7 @@
         <p class="text-md text-bodydark">Media Profiles set your preferences for fetching and downloading media.</p>
         <p class="text-md text-bodydark">Don't worry, you can create more Media Profiles later!</p>
         <div class="mt-8">
-          <.link navigate={~p"/media_profiles/new"}>
+          <.link href={~p"/media_profiles/new"}>
             <.button color="bg-primary" rounding="rounded-full" disabled={@media_profiles_exist}>
               <span class="font-bold mx-2">+</span> New Media Profile
             </.button>
@@ -24,7 +24,7 @@
           Each Media Profile can control many Sources so it's easy to add more content!
         </p>
         <div class="mt-8">
-          <.link navigate={~p"/sources/new"}>
+          <.link href={~p"/sources/new"}>
             <.button color="bg-primary" rounding="rounded-full" disabled={not @media_profiles_exist}>
               <span class="font-bold mx-2">+</span> New Source
             </.button>
@@ -39,7 +39,7 @@
         </p>
         <p class="text-md text-bodydark">Feel free to add more Media Profiles or Sources in the meantime!</p>
         <div class="mt-8">
-          <.link navigate={~p"/"}>
+          <.link href={~p"/?onboarding=0"}>
             <.button color="bg-primary" rounding="rounded-full" disabled={not @sources_exist}>
               Let's Go <span class="font-bold mx-2">🚀</span>
             </.button>

lib/pinchflat_web/controllers/searches/search_html/show.html.heex

@@ -17,7 +17,7 @@
           </:col>
           <:col :let={result} label="" class="flex place-content-evenly">
             <.link
-              navigate={~p"/sources/#{result.source_id}/media/#{result.id}"}
+              href={~p"/sources/#{result.source_id}/media/#{result.id}"}
               class="hover:text-secondary duration-200 ease-in-out mx-0.5"
             >
               <.icon name="hero-eye" />

lib/pinchflat_web/controllers/sources/source_controller.ex

@@ -3,8 +3,9 @@ defmodule PinchflatWeb.Sources.SourceController do
 
   alias Pinchflat.Repo
   alias Pinchflat.Media
-  alias Pinchflat.Profiles
+  alias Pinchflat.Tasks
   alias Pinchflat.Sources
+  alias Pinchflat.Profiles
   alias Pinchflat.Sources.Source
 
   def index(conn, _params) do
@@ -43,15 +44,18 @@ defmodule PinchflatWeb.Sources.SourceController do
   end
 
   def show(conn, %{"id" => id}) do
-    source =
+    source = Repo.preload(Sources.get_source!(id), :media_profile)
-      id
-      |> Sources.get_source!()
-      |> Repo.preload([:media_profile, tasks: [:job]])
 
+    pending_tasks = Repo.preload(Tasks.list_pending_tasks_for(:source_id, source.id), :job)
     pending_media = Media.list_pending_media_items_for(source, limit: 100)
     downloaded_media = Media.list_downloaded_media_items_for(source, limit: 100)
 
-    render(conn, :show, source: source, pending_media: pending_media, downloaded_media: downloaded_media)
+    render(conn, :show,
+      source: source,
+      pending_tasks: pending_tasks,
+      pending_media: pending_media,
+      downloaded_media: downloaded_media
+    )
   end
 
   def edit(conn, %{"id" => id}) do

lib/pinchflat_web/controllers/sources/source_html/edit.html.heex

@@ -1,5 +1,5 @@
 <div class="mb-6 flex gap-3 flex-row items-center">
-  <.link navigate={~p"/sources"}>
+  <.link href={~p"/sources"}>
     <.icon name="hero-arrow-left" class="w-10 h-10 hover:dark:text-white" />
   </.link>
   <h2 class="text-title-md2 font-bold text-black dark:text-white ml-4">Edit Source</h2>

lib/pinchflat_web/controllers/sources/source_html/fast_indexing_help.html.heex

@@ -0,0 +1,21 @@
+<aside>
+  <h2 class="text-xl font-bold mb-2">What is fast indexing (experimental)?</h2>
+  <section class="ml-2 md:ml-4 mb-4 max-w-prose">
+    <p>
+      Indexing is the act of scanning a channel or playlist (aka: source) for new media.
+    </p>
+    <p class="mt-2">
+      Normal indexing uses <code class="text-sm">yt-dlp</code>
+      to scan the entire source on your specified frequency, but it's very slow for large sources. This is the most accurate way to find uploaded media with the tradeoff being that pairing a large source with a low index frequency will result in you spending most of your time indexing. Only so many indexing operations can be running at the same time, so this can impact your other source's ability to index.
+    </p>
+    <p class="mt-2">
+      Fast indexing takes a different approach. It still does an initial scan the slow way but after that it uses an RSS feed to frequently check for new videos. This has the potential to be hundreds of times faster, but it can miss videos if the uploader un-privates an old video or uploads dozens of videos in the space of a few minutes. It works well for most channels or playlists but it's not perfect.
+    </p>
+    <p class="mt-2">
+      To make up for this limitation, a normal index is still run monthly to catch any videos that were missed by fast indexing. Fast indexing overrides the normal index frequency.
+    </p>
+    <p class="mt-2">
+      Fast indexing is experimental so please report any issues on GitHub. It's only recommended for sources with over 200-ish videos and that upload frequently. Not recommended for small or inactive sources.
+    </p>
+  </section>
+</aside>

lib/pinchflat_web/controllers/sources/source_html/index.html.heex

@@ -1,7 +1,7 @@
 <div class="mb-6 flex gap-3 flex-row items-center justify-between">
   <h2 class="text-title-md2 font-bold text-black dark:text-white">Sources</h2>
   <nav>
-    <.link navigate={~p"/sources/new"}>
+    <.link href={~p"/sources/new"}>
       <.button color="bg-primary" rounding="rounded-full">
         <span class="font-bold mx-2">+</span> New <span class="hidden sm:inline pl-1">Source</span>
       </.button>
@@ -22,7 +22,7 @@
         </:col>
         <:col :let={source} label="Media Profile">
           <.link
-            navigate={~p"/media_profiles/#{source.media_profile_id}"}
+            href={~p"/media_profiles/#{source.media_profile_id}"}
             class="hover:text-secondary duration-200 ease-in-out"
           >
             <%= source.media_profile.name %>

lib/pinchflat_web/controllers/sources/source_html/new.html.heex

@@ -1,5 +1,5 @@
 <div class="mb-6 flex gap-3 flex-row items-center">
-  <.link :if={!Settings.get!(:onboarding)} navigate={~p"/sources"}>
+  <.link :if={!Settings.get!(:onboarding)} href={~p"/sources"}>
     <.icon name="hero-arrow-left" class="w-10 h-10 hover:dark:text-white" />
   </.link>
   <h2 class="text-title-md2 font-bold text-black dark:text-white ml-4">New Source</h2>

lib/pinchflat_web/controllers/sources/source_html/show.html.heex

@@ -1,6 +1,6 @@
 <div class="mb-6 flex gap-3 flex-row items-center justify-between">
   <div class="flex gap-3 items-center">
-    <.link navigate={~p"/sources"}>
+    <.link href={~p"/sources"}>
       <.icon name="hero-arrow-left" class="w-10 h-10 hover:dark:text-white" />
     </.link>
     <h2 class="text-title-md2 font-bold text-black dark:text-white ml-4">
@@ -9,7 +9,7 @@
   </div>
 
   <nav>
-    <.link navigate={~p"/sources/#{@source}/edit"}>
+    <.link href={~p"/sources/#{@source}/edit"}>
       <.button color="bg-primary" rounding="rounded-full">
         <.icon name="hero-pencil-square" class="mr-2" /> Edit <span class="hidden sm:inline pl-1">Source</span>
       </.button>
@@ -84,9 +84,9 @@
           <p class="text-black dark:text-white">Nothing Here!</p>
         <% end %>
       </:tab>
-      <:tab title="Tasks">
+      <:tab title="Pending Tasks">
-        <%= if match?([_|_], @source.tasks) do %>
+        <%= if match?([_|_], @pending_tasks) do %>
-          <.table rows={@source.tasks} table_class="text-black dark:text-white">
+          <.table rows={@pending_tasks} table_class="text-black dark:text-white">
             <:col :let={task} label="Worker">
               <%= task.job.worker %>
             </:col>

lib/pinchflat_web/controllers/sources/source_html/source_form.html.heex

@@ -3,6 +3,10 @@
     Oops, something went wrong! Please check the errors below.
   </.error>
 
+  <h3 class="mt-8 text-2xl text-black dark:text-white">
+    General Options
+  </h3>
+
   <.input
     field={f[:custom_name]}
     type="text"
@@ -19,6 +23,10 @@
     label="Media Profile"
   />
 
+  <h3 class="mt-8 text-2xl text-black dark:text-white">
+    Indexing Options
+  </h3>
+
   <.input
     field={f[:index_frequency_minutes]}
     options={friendly_index_frequencies()}
@@ -27,6 +35,21 @@
     help="Time between one index of this source finishing and the next one starting. Setting to 'On Create' will still run an initial index but no subsequent ones"
   />
 
+  <%!-- TODO: use Alpine to disable the index frequency when fast indexing is enabled --%>
+  <div phx-click={show_modal("upgrade-modal")}>
+    <.input
+      field={f[:fast_index]}
+      type="toggle"
+      label="Use Fast Indexing?"
+      label_suffix="(pro)"
+      help="Experimental. Ignores 'Index Frequency'. Recommended for large channels that upload frequently. See below for more info"
+    />
+  </div>
+
+  <h3 class="mt-8 text-2xl text-black dark:text-white">
+    Downloading Options
+  </h3>
+
   <.input
     field={f[:download_media]}
     type="toggle"
@@ -34,7 +57,17 @@
     help="Unchecking still indexes media but it won't be downloaded until you enable this option"
   />
 
-  <:actions>
+  <.input
-    <.button class="my-10 sm:mb-7.5 w-full sm:w-auto">Save Source</.button>
+    field={f[:download_cutoff_date]}
-  </:actions>
+    type="text"
+    label="Download Cutoff Date"
+    placeholder="YYYY-MM-DD"
+    help="Only download media uploaded after this date. Leave blank to download all media. Must be in YYYY-MM-DD format"
+  />
+
+  <.button class="my-10 sm:mb-7.5 w-full sm:w-auto">Save Source</.button>
+
+  <div class="rounded-sm dark:bg-meta-4 p-4 md:p-6 mb-5">
+    <.fast_indexing_help />
+  </div>
 </.simple_form>

mix.exs

@@ -4,7 +4,7 @@ defmodule Pinchflat.MixProject do
   def project do
     [
       app: :pinchflat,
-      version: "0.1.0-alpha",
+      version: "0.1.0-alpha.2",
       elixir: "~> 1.16",
       elixirc_paths: elixirc_paths(Mix.env()),
       start_permanent: Mix.env() == :prod,

priv/repo/migrations/20240309052602_add_short_form_to_media_items.exs

@@ -0,0 +1,9 @@
+defmodule Pinchflat.Repo.Migrations.AddShortFormToMediaItems do
+  use Ecto.Migration
+
+  def change do
+    alter table(:media_items) do
+      add :short_form_content, :boolean, null: false, default: false
+    end
+  end
+end

priv/repo/migrations/20240309182418_add_fast_index_to_sources.exs

@@ -0,0 +1,9 @@
+defmodule Pinchflat.Repo.Migrations.AddFastIndexToSources do
+  use Ecto.Migration
+
+  def change do
+    alter table(:sources) do
+      add :fast_index, :boolean, null: false, default: false
+    end
+  end
+end

priv/repo/migrations/20240310230713_add_uploaded_at_to_media_items.exs

@@ -0,0 +1,12 @@
+defmodule Pinchflat.Repo.Migrations.AddUploadedAtToMediaItems do
+  use Ecto.Migration
+
+  def change do
+    alter table(:media_items) do
+      # Setting default to unix epoch so I can enforce not null BUT also easily
+      # identify records that were created before this column was added.
+      # Not a DateTime because yt-dlp only returns the date
+      add :upload_date, :date, default: "1970-01-01", null: false
+    end
+  end
+end

priv/repo/migrations/20240311033214_add_download_cutoff_date_to_sources.exs

@@ -0,0 +1,9 @@
+defmodule Pinchflat.Repo.Migrations.AddDownloadCutoffDateToSources do
+  use Ecto.Migration
+
+  def change do
+    alter table(:sources) do
+      add :download_cutoff_date, :date
+    end
+  end
+end

selfhosted.Dockerfile

@@ -11,9 +11,9 @@
 #   - https://pkgs.org/ - resource for finding needed packages
 #   - Ex: hexpm/elixir:1.16.0-erlang-26.2.1-debian-bullseye-20231009-slim
 #
-ARG ELIXIR_VERSION=1.16.0
+ARG ELIXIR_VERSION=1.16.2
-ARG OTP_VERSION=26.2.1
+ARG OTP_VERSION=26.2.2
-ARG DEBIAN_VERSION=bullseye-20231009-slim
+ARG DEBIAN_VERSION=bookworm-20240130-slim
 
 ARG BUILDER_IMAGE="hexpm/elixir:${ELIXIR_VERSION}-erlang-${OTP_VERSION}-debian-${DEBIAN_VERSION}"
 ARG RUNNER_IMAGE="debian:${DEBIAN_VERSION}"
@@ -76,12 +76,13 @@ FROM ${RUNNER_IMAGE}
 
 RUN apt-get update -y
 RUN apt-get install -y libstdc++6 openssl libncurses5 locales ca-certificates \
-  python3 python3-pip ffmpeg
+  ffmpeg curl git openssh-client
 RUN apt-get clean && rm -f /var/lib/apt/lists/*_*
 
-# Download YT-DLP
+# Download and update YT-DLP
-# NOTE: If you're seeing weird issues, consider using the FFMPEG released by yt-dlp
+RUN curl -L https://github.com/yt-dlp/yt-dlp/releases/latest/download/yt-dlp -o /usr/local/bin/yt-dlp
-RUN python3 -m pip install -U --pre yt-dlp
+RUN chmod a+rx /usr/local/bin/yt-dlp
+RUN yt-dlp -U
 
 # Set the locale
 RUN sed -i '/en_US.UTF-8/s/^# //g' /etc/locale.gen && locale-gen
@@ -106,6 +107,17 @@ ENV RUN_CONTEXT="selfhosted"
 # Only copy the final release from the build stage
 COPY --from=builder --chown=nobody:root /app/_build/${MIX_ENV}/rel/pinchflat ./
 
+# NEVER do this if you're running in an environment where you don't trust the user
+# (ie: most environments). This is only acceptable in a self-hosted environment.
+# The user could just run the whole container as root and bypass this anyway so
+# it's not a huge deal.
+# This removes the root password to allow users to assume root if needed. This is
+# preferrable to running the whole container as root so that the files/directories
+# created by the app aren't owned by root and are therefore easier for other users
+# and processes to interact with. If you want to just run the whole container as
+# root, use --user 0:0 or something.
+RUN passwd -d root
+
 USER nobody
 
 # If using an environment that doesn't automatically reap zombie processes, it is

test/pinchflat/api/youtube_rss_test.exs

@@ -0,0 +1,79 @@
+defmodule Pinchflat.Api.YoutubeRssTest do
+  use Pinchflat.DataCase
+  import Mox
+  import Pinchflat.SourcesFixtures
+
+  alias Pinchflat.Api.YoutubeRss
+
+  setup :verify_on_exit!
+
+  setup do
+    source = source_fixture()
+
+    {:ok, source: source}
+  end
+
+  describe "get_recent_media_ids_from_rss/1" do
+    test "calls the expected URL for channel sources" do
+      source = source_fixture(collection_type: :channel, collection_id: "channel_id")
+
+      expect(HTTPClientMock, :get, fn url ->
+        assert url =~ "https://www.youtube.com/feeds/videos.xml?channel_id=#{source.collection_id}"
+
+        {:ok, ""}
+      end)
+
+      assert {:ok, _} = YoutubeRss.get_recent_media_ids_from_rss(source)
+    end
+
+    test "calls the expected URL for playlist sources" do
+      source = source_fixture(collection_type: :playlist, collection_id: "playlist_id")
+
+      expect(HTTPClientMock, :get, fn url ->
+        assert url =~ "https://www.youtube.com/feeds/videos.xml?playlist_id=#{source.collection_id}"
+
+        {:ok, ""}
+      end)
+
+      assert {:ok, _} = YoutubeRss.get_recent_media_ids_from_rss(source)
+    end
+
+    test "returns an error if the HTTP request fails", %{source: source} do
+      expect(HTTPClientMock, :get, fn _url -> {:error, ""} end)
+
+      assert {:error, "Failed to fetch RSS feed"} = YoutubeRss.get_recent_media_ids_from_rss(source)
+    end
+
+    test "returns the media IDs from the RSS feed", %{source: source} do
+      expect(HTTPClientMock, :get, fn _url ->
+        {:ok, "<yt:videoId>test_1</yt:videoId><yt:videoId>test_2</yt:videoId>"}
+      end)
+
+      assert {:ok, ["test_1", "test_2"]} = YoutubeRss.get_recent_media_ids_from_rss(source)
+    end
+
+    test "strips whitespace from media IDs", %{source: source} do
+      expect(HTTPClientMock, :get, fn _url ->
+        {:ok, "<yt:videoId> test_1 </yt:videoId><yt:videoId> test_2 </yt:videoId>"}
+      end)
+
+      assert {:ok, ["test_1", "test_2"]} = YoutubeRss.get_recent_media_ids_from_rss(source)
+    end
+
+    test "removes empty media IDs", %{source: source} do
+      expect(HTTPClientMock, :get, fn _url ->
+        {:ok, "<yt:videoId>test_1</yt:videoId><yt:videoId></yt:videoId>"}
+      end)
+
+      assert {:ok, ["test_1"]} = YoutubeRss.get_recent_media_ids_from_rss(source)
+    end
+
+    test "removes duplicate media IDs", %{source: source} do
+      expect(HTTPClientMock, :get, fn _url ->
+        {:ok, "<yt:videoId>test_1</yt:videoId><yt:videoId>test_1</yt:videoId>"}
+      end)
+
+      assert {:ok, ["test_1"]} = YoutubeRss.get_recent_media_ids_from_rss(source)
+    end
+  end
+end

test/pinchflat/media_client/backends/yt_dlp/media_test.exs

@@ -1,56 +0,0 @@
-defmodule Pinchflat.MediaClient.Backends.YtDlp.MediaTest do
-  use Pinchflat.DataCase
-  import Mox
-
-  alias Pinchflat.MediaClient.Backends.YtDlp.Media
-
-  @media_url "https://www.youtube.com/watch?v=TiZPUDkDYbk"
-
-  setup :verify_on_exit!
-
-  # expect(YtDlpRunnerMock, :run, fn _url, [_, _, json_output_path | _] ->
-  # copy_metadata(json_output_path)
-
-  #   {:ok, ""}
-  # end)
-
-  describe "download/2" do
-    test "it calls the backend runner with the expected arguments" do
-      expect(YtDlpRunnerMock, :run, fn @media_url, opts, ot ->
-        assert [:no_simulate] = opts
-        assert "after_move:%()j" = ot
-
-        {:ok, render_metadata(:media_metadata)}
-      end)
-
-      assert {:ok, _} = Media.download(@media_url)
-    end
-
-    test "it passes along additional options" do
-      expect(YtDlpRunnerMock, :run, fn _url, opts, _ot ->
-        assert [:no_simulate, :custom_arg] = opts
-
-        {:ok, "{}"}
-      end)
-
-      assert {:ok, _} = Media.download(@media_url, [:custom_arg])
-    end
-
-    test "it parses and returns the generated file as JSON" do
-      expect(YtDlpRunnerMock, :run, fn _url, _opts, _ot ->
-        {:ok, render_metadata(:media_metadata)}
-      end)
-
-      assert {:ok, %{"title" => "Trying to Wheelie Without the Rear Brake"}} =
-               Media.download(@media_url)
-    end
-
-    test "it returns errors" do
-      expect(YtDlpRunnerMock, :run, fn _url, _opt, _ot ->
-        {:error, "something"}
-      end)
-
-      assert {:error, "something"} = Media.download(@media_url)
-    end
-  end
-end

test/pinchflat/media_client/source_details_test.exs

@@ -1,112 +0,0 @@
-defmodule Pinchflat.MediaClient.SourceDetailsTest do
-  use Pinchflat.DataCase
-  import Mox
-  import Pinchflat.ProfilesFixtures
-  import Pinchflat.SourcesFixtures
-
-  alias Pinchflat.MediaClient.SourceDetails
-
-  @channel_url "https://www.youtube.com/c/TheUselessTrials"
-
-  setup :verify_on_exit!
-
-  describe "get_source_details/2" do
-    test "it passes the expected arguments to the backend" do
-      expect(YtDlpRunnerMock, :run, fn @channel_url, opts, ot ->
-        assert opts == [:simulate, :skip_download, playlist_end: 1]
-        assert ot == "%(.{channel,channel_id,playlist_id,playlist_title})j"
-
-        {:ok, "{}"}
-      end)
-
-      assert {:ok, _} = SourceDetails.get_source_details(@channel_url)
-    end
-
-    test "it returns a map composed of the returned data" do
-      expect(YtDlpRunnerMock, :run, fn _url, _opts, _ot ->
-        Phoenix.json_library().encode(%{
-          channel: "TheUselessTrials",
-          channel_id: "UCQH2",
-          playlist_id: "PLQH2",
-          playlist_title: "TheUselessTrials - Videos"
-        })
-      end)
-
-      assert {:ok, res} = SourceDetails.get_source_details(@channel_url)
-
-      assert %{
-               channel_id: "UCQH2",
-               channel_name: "TheUselessTrials",
-               playlist_id: "PLQH2",
-               playlist_name: "TheUselessTrials - Videos"
-             } = res
-    end
-  end
-
-  describe "get_media_attributes/2 when passed a string" do
-    test "it passes the expected arguments to the backend" do
-      expect(YtDlpRunnerMock, :run, fn @channel_url, opts, ot, _addl_opts ->
-        assert opts == [:simulate, :skip_download]
-        assert ot == "%(.{id,title,was_live,original_url,description})j"
-
-        {:ok, ""}
-      end)
-
-      assert {:ok, _} = SourceDetails.get_media_attributes(@channel_url)
-    end
-
-    test "it returns a list of maps" do
-      expect(YtDlpRunnerMock, :run, fn _url, _opts, _ot, _addl_opts ->
-        {:ok, source_attributes_return_fixture()}
-      end)
-
-      assert {:ok, [%{}, %{}, %{}]} = SourceDetails.get_media_attributes(@channel_url)
-    end
-  end
-
-  describe "get_media_attributes/2 when passed a Source record" do
-    test "it calls the backend with the source's collection ID" do
-      source = source_fixture()
-
-      expect(YtDlpRunnerMock, :run, fn url, _opts, _ot, _addl_opts ->
-        assert source.collection_id == url
-        {:ok, source_attributes_return_fixture()}
-      end)
-
-      assert {:ok, _} = SourceDetails.get_media_attributes(source)
-    end
-
-    test "it builds options based on the source's media profile" do
-      expect(YtDlpRunnerMock, :run, fn _url, opts, _ot, _addl_opts ->
-        assert opts == [:simulate, :skip_download]
-        {:ok, ""}
-      end)
-
-      media_profile =
-        media_profile_fixture(
-          shorts_behaviour: :include,
-          livestream_behaviour: :exclude
-        )
-
-      source = source_fixture(media_profile_id: media_profile.id)
-      assert {:ok, _} = SourceDetails.get_media_attributes(source)
-    end
-
-    test "lets you pass through an optional file_listener_handler" do
-      expect(YtDlpRunnerMock, :run, fn _url, _opts, _ot, _addl_opts ->
-        {:ok, source_attributes_return_fixture()}
-      end)
-
-      source = source_fixture()
-      current_self = self()
-
-      handler = fn filename ->
-        send(current_self, {:handler, filename})
-      end
-
-      assert {:ok, _} = SourceDetails.get_media_attributes(source, file_listener_handler: handler)
-
-      assert_receive {:handler, _}
-    end
-  end
-end

test/pinchflat/media_test.exs

@@ -9,7 +9,9 @@ defmodule Pinchflat.MediaTest do
 
   alias Pinchflat.Media
   alias Pinchflat.Media.MediaItem
-  alias Pinchflat.MediaClient.Backends.YtDlp.MetadataFileHelpers
+  alias Pinchflat.Metadata.MetadataFileHelpers
+
+  alias Pinchflat.YtDlp.Backend.Media, as: YtDlpMedia
 
   setup :verify_on_exit!
 
@@ -45,6 +47,23 @@ defmodule Pinchflat.MediaTest do
     end
   end
 
+  describe "list_media_items_by_media_id_for/2" do
+    test "returns media_items for a given source and media_ids" do
+      source = source_fixture()
+      media_item = media_item_fixture(%{source_id: source.id, media_id: "123"})
+
+      assert Media.list_media_items_by_media_id_for(source, ["123"]) == [media_item]
+    end
+
+    test "does not return matching media_ids for a different source" do
+      source = source_fixture()
+      other_source = source_fixture()
+      _media_item = media_item_fixture(%{source_id: other_source.id, media_id: "123"})
+
+      assert Media.list_media_items_by_media_id_for(source, ["123"]) == []
+    end
+  end
+
   describe "list_pending_media_items_for/1" do
     test "it returns pending without a filepath for a given source" do
       source = source_fixture()
@@ -78,7 +97,7 @@ defmodule Pinchflat.MediaTest do
     test "returns shorts and normal media when shorts_behaviour is :include" do
       source = source_fixture(%{media_profile_id: media_profile_fixture(%{shorts_behaviour: :include}).id})
       normal = media_item_fixture(%{source_id: source.id, media_filepath: nil})
-      short = media_item_fixture(%{source_id: source.id, media_filepath: nil, original_url: "/shorts/"})
+      short = media_item_fixture(%{source_id: source.id, media_filepath: nil, short_form_content: true})
 
       assert Media.list_pending_media_items_for(source) == [normal, short]
     end
@@ -86,7 +105,7 @@ defmodule Pinchflat.MediaTest do
     test "returns only shorts when shorts_behaviour is :only" do
       source = source_fixture(%{media_profile_id: media_profile_fixture(%{shorts_behaviour: :only}).id})
       _normal = media_item_fixture(%{source_id: source.id, media_filepath: nil})
-      short = media_item_fixture(%{source_id: source.id, media_filepath: nil, original_url: "/shorts/"})
+      short = media_item_fixture(%{source_id: source.id, media_filepath: nil, short_form_content: true})
 
       assert Media.list_pending_media_items_for(source) == [short]
     end
@@ -94,7 +113,7 @@ defmodule Pinchflat.MediaTest do
     test "returns only normal media when shorts_behaviour is :exclude" do
       source = source_fixture(%{media_profile_id: media_profile_fixture(%{shorts_behaviour: :exclude}).id})
       normal = media_item_fixture(%{source_id: source.id, media_filepath: nil})
-      _short = media_item_fixture(%{source_id: source.id, media_filepath: nil, original_url: "/shorts/"})
+      _short = media_item_fixture(%{source_id: source.id, media_filepath: nil, short_form_content: true})
 
       assert Media.list_pending_media_items_for(source) == [normal]
     end
@@ -139,7 +158,7 @@ defmodule Pinchflat.MediaTest do
 
       normal = media_item_fixture(%{source_id: source.id, media_filepath: nil})
       livestream = media_item_fixture(%{source_id: source.id, media_filepath: nil, livestream: true})
-      short = media_item_fixture(%{source_id: source.id, media_filepath: nil, original_url: "/shorts/"})
+      short = media_item_fixture(%{source_id: source.id, media_filepath: nil, short_form_content: true})
 
       assert Media.list_pending_media_items_for(source) == [normal, livestream, short]
     end
@@ -156,7 +175,7 @@ defmodule Pinchflat.MediaTest do
 
       _normal = media_item_fixture(%{source_id: source.id, media_filepath: nil})
       livestream = media_item_fixture(%{source_id: source.id, media_filepath: nil, livestream: true})
-      short = media_item_fixture(%{source_id: source.id, media_filepath: nil, original_url: "/shorts/"})
+      short = media_item_fixture(%{source_id: source.id, media_filepath: nil, short_form_content: true})
 
       assert Media.list_pending_media_items_for(source) == [livestream, short]
     end
@@ -173,7 +192,7 @@ defmodule Pinchflat.MediaTest do
 
       normal = media_item_fixture(%{source_id: source.id, media_filepath: nil})
       _livestream = media_item_fixture(%{source_id: source.id, media_filepath: nil, livestream: true})
-      _short = media_item_fixture(%{source_id: source.id, media_filepath: nil, original_url: "/shorts/"})
+      _short = media_item_fixture(%{source_id: source.id, media_filepath: nil, short_form_content: true})
 
       assert Media.list_pending_media_items_for(source) == [normal]
     end
@@ -190,12 +209,36 @@ defmodule Pinchflat.MediaTest do
 
       _normal = media_item_fixture(%{source_id: source.id, media_filepath: nil})
       _livestream = media_item_fixture(%{source_id: source.id, media_filepath: nil, livestream: true})
-      short = media_item_fixture(%{source_id: source.id, media_filepath: nil, original_url: "/shorts/"})
+      short = media_item_fixture(%{source_id: source.id, media_filepath: nil, short_form_content: true})
 
       assert Media.list_pending_media_items_for(source) == [short]
     end
   end
 
+  describe "list_pending_media_items_for/1 when testing cutoff dates" do
+    test "does not return media items with an upload date before the cutoff date" do
+      source = source_fixture(%{download_cutoff_date: now_minus(1, :day)})
+
+      _old_media_item =
+        media_item_fixture(%{source_id: source.id, media_filepath: nil, upload_date: now_minus(2, :days)})
+
+      new_media_item = media_item_fixture(%{source_id: source.id, media_filepath: nil, upload_date: now()})
+
+      assert Media.list_pending_media_items_for(source) == [new_media_item]
+    end
+
+    test "does not apply a cutoff if there is no cutoff date" do
+      source = source_fixture(%{download_cutoff_date: nil})
+
+      old_media_item =
+        media_item_fixture(%{source_id: source.id, media_filepath: nil, upload_date: now_minus(2, :days)})
+
+      new_media_item = media_item_fixture(%{source_id: source.id, media_filepath: nil, upload_date: now()})
+
+      assert Media.list_pending_media_items_for(source) == [old_media_item, new_media_item]
+    end
+  end
+
   describe "list_downloaded_media_items_for/1" do
     test "returns only media items with a media_filepath" do
       source = source_fixture()
@@ -230,7 +273,7 @@ defmodule Pinchflat.MediaTest do
 
     test "returns false if the media hasn't been downloaded but the profile doesn't DL shorts" do
       source = source_fixture(%{media_profile_id: media_profile_fixture(%{shorts_behaviour: :exclude}).id})
-      media_item = media_item_fixture(%{source_id: source.id, media_filepath: nil, original_url: "/shorts/"})
+      media_item = media_item_fixture(%{source_id: source.id, media_filepath: nil, short_form_content: true})
 
       refute Media.pending_download?(media_item)
     end
@@ -241,6 +284,27 @@ defmodule Pinchflat.MediaTest do
 
       refute Media.pending_download?(media_item)
     end
+
+    test "returns true if there is a cutoff date before the media's upload date" do
+      source = source_fixture(%{download_cutoff_date: now_minus(2, :days)})
+      media_item = media_item_fixture(%{source_id: source.id, media_filepath: nil, upload_date: now_minus(1, :day)})
+
+      assert Media.pending_download?(media_item)
+    end
+
+    test "returns false if there is a cutoff date after the media's upload date" do
+      source = source_fixture(%{download_cutoff_date: now_minus(1, :day)})
+      media_item = media_item_fixture(%{source_id: source.id, media_filepath: nil, upload_date: now_minus(2, :days)})
+
+      refute Media.pending_download?(media_item)
+    end
+
+    test "returns true if there is no cutoff date" do
+      source = source_fixture(%{download_cutoff_date: nil})
+      media_item = media_item_fixture(%{source_id: source.id, media_filepath: nil, upload_date: now_minus(1, :day)})
+
+      assert Media.pending_download?(media_item)
+    end
   end
 
   describe "search/1" do
@@ -354,10 +418,12 @@ defmodule Pinchflat.MediaTest do
         title: Faker.Commerce.product_name(),
         media_filepath: "/video/#{Faker.File.file_name(:video)}",
         source_id: source_fixture().id,
-        original_url: "https://www.youtube.com/channel/#{Faker.String.base64(12)}"
+        original_url: "https://www.youtube.com/channel/#{Faker.String.base64(12)}",
+        upload_date: Date.utc_today()
       }
 
       assert {:ok, %MediaItem{} = media_item} = Media.create_media_item(valid_attrs)
+
       assert media_item.title == valid_attrs.title
       assert media_item.media_id == valid_attrs.media_id
       assert media_item.media_filepath == valid_attrs.media_filepath
@@ -368,6 +434,42 @@ defmodule Pinchflat.MediaTest do
     end
   end
 
+  describe "create_media_item_from_backend_attrs/2" do
+    test "creates a media item for a given source and attributes" do
+      source = source_fixture()
+
+      media_attrs =
+        media_attributes_return_fixture()
+        |> Phoenix.json_library().decode!()
+        |> YtDlpMedia.response_to_struct()
+
+      assert {:ok, %MediaItem{} = media_item} = Media.create_media_item_from_backend_attrs(source, media_attrs)
+
+      assert media_item.source_id == source.id
+      assert media_item.title == media_attrs.title
+      assert media_item.media_id == media_attrs.media_id
+      assert media_item.original_url == media_attrs.original_url
+      assert media_item.description == media_attrs.description
+    end
+
+    test "updates the media item if it already exists" do
+      source = source_fixture()
+
+      media_attrs =
+        media_attributes_return_fixture()
+        |> Phoenix.json_library().decode!()
+        |> YtDlpMedia.response_to_struct()
+
+      different_attrs = %YtDlpMedia{media_attrs | title: "Different title"}
+
+      assert {:ok, %MediaItem{} = media_item_1} = Media.create_media_item_from_backend_attrs(source, media_attrs)
+      assert {:ok, %MediaItem{} = media_item_2} = Media.create_media_item_from_backend_attrs(source, different_attrs)
+
+      assert media_item_1.id == media_item_2.id
+      assert media_item_2.title == different_attrs.title
+    end
+  end
+
   describe "update_media_item/2" do
     test "updating with valid data updates the media_item" do
       media_item = media_item_fixture()

test/pinchflat/metadata/metadata_file_helpers_test.exs

@@ -1,9 +1,9 @@
-defmodule Pinchflat.MediaClient.Backends.YtDlp.MetadataFileHelpersTest do
+defmodule Pinchflat.Metadata.MetadataFileHelpersTest do
   use Pinchflat.DataCase
   import Mox
   import Pinchflat.MediaFixtures
 
-  alias Pinchflat.MediaClient.Backends.YtDlp.MetadataFileHelpers, as: Helpers
+  alias Pinchflat.Metadata.MetadataFileHelpers, as: Helpers
 
   setup do
     media_item = media_item_fixture()

test/pinchflat/metadata/metadata_parser_test.exs

@@ -1,7 +1,7 @@
-defmodule Pinchflat.MediaClient.Backends.YtDlp.MediaParserTest do
+defmodule Pinchflat.YtDlp.Backend.MediaParserTest do
   use Pinchflat.DataCase
 
-  alias Pinchflat.MediaClient.Backends.YtDlp.MetadataParser, as: Parser
+  alias Pinchflat.Metadata.MetadataParser, as: Parser
 
   setup do
     json_filepath =
@@ -33,13 +33,31 @@ defmodule Pinchflat.MediaClient.Backends.YtDlp.MediaParserTest do
     test "it extracts the title", %{metadata: metadata} do
       result = Parser.parse_for_media_item(metadata)
 
-      assert result.title == "Trying to Wheelie Without the Rear Brake"
+      assert result.title == metadata["title"]
     end
 
     test "it extracts the description", %{metadata: metadata} do
       result = Parser.parse_for_media_item(metadata)
 
-      assert is_binary(result.description)
+      assert result.description == metadata["description"]
+    end
+
+    test "it extracts the original_url", %{metadata: metadata} do
+      result = Parser.parse_for_media_item(metadata)
+
+      assert result.original_url == metadata["original_url"]
+    end
+
+    test "it extracts the media_id", %{metadata: metadata} do
+      result = Parser.parse_for_media_item(metadata)
+
+      assert result.media_id == metadata["id"]
+    end
+
+    test "it extracts the livestream flag", %{metadata: metadata} do
+      result = Parser.parse_for_media_item(metadata)
+
+      assert result.livestream == metadata["was_live"]
     end
   end
 

test/pinchflat/sources_test.exs

@@ -9,8 +9,10 @@ defmodule Pinchflat.SourcesTest do
   alias Pinchflat.Sources
   alias Pinchflat.Tasks.SourceTasks
   alias Pinchflat.Sources.Source
-  alias Pinchflat.Workers.MediaIndexingWorker
+  alias Pinchflat.Workers.FastIndexingWorker
   alias Pinchflat.Workers.MediaDownloadWorker
+  alias Pinchflat.Workers.MediaIndexingWorker
+  alias Pinchflat.Workers.MediaCollectionIndexingWorker
 
   @invalid_source_attrs %{name: nil, collection_id: nil}
 
@@ -113,6 +115,12 @@ defmodule Pinchflat.SourcesTest do
       assert {:error, %Ecto.Changeset{}} = Sources.create_source(@invalid_source_attrs)
     end
 
+    test "creation with invalid data fails fast and does not call the runner" do
+      expect(YtDlpRunnerMock, :run, 0, &channel_mock/3)
+
+      assert {:error, %Ecto.Changeset{}} = Sources.create_source(@invalid_source_attrs)
+    end
+
     test "creation enforces uniqueness of collection_id scoped to the media_profile" do
       expect(YtDlpRunnerMock, :run, 2, fn _url, _opts, _ot ->
         {:ok,
@@ -166,7 +174,7 @@ defmodule Pinchflat.SourcesTest do
 
       assert {:ok, %Source{} = source} = Sources.create_source(valid_attrs)
 
-      assert_enqueued(worker: MediaIndexingWorker, args: %{"id" => source.id})
+      assert_enqueued(worker: MediaCollectionIndexingWorker, args: %{"id" => source.id})
     end
 
     test "creation schedules an index test even if the index frequency is 0" do
@@ -180,7 +188,37 @@ defmodule Pinchflat.SourcesTest do
 
       assert {:ok, %Source{} = source} = Sources.create_source(valid_attrs)
 
-      assert_enqueued(worker: MediaIndexingWorker, args: %{"id" => source.id})
+      assert_enqueued(worker: MediaCollectionIndexingWorker, args: %{"id" => source.id})
+    end
+
+    test "fast_index forces the index frequency to be a default value" do
+      expect(YtDlpRunnerMock, :run, &channel_mock/3)
+
+      valid_attrs = %{
+        media_profile_id: media_profile_fixture().id,
+        original_url: "https://www.youtube.com/channel/abc123",
+        fast_index: true,
+        index_frequency_minutes: 0
+      }
+
+      assert {:ok, %Source{} = source} = Sources.create_source(valid_attrs)
+
+      assert source.index_frequency_minutes == Source.index_frequency_when_fast_indexing()
+    end
+
+    test "disabling fast index will not change the index frequency" do
+      expect(YtDlpRunnerMock, :run, &channel_mock/3)
+
+      valid_attrs = %{
+        media_profile_id: media_profile_fixture().id,
+        original_url: "https://www.youtube.com/channel/abc123",
+        fast_index: false,
+        index_frequency_minutes: 0
+      }
+
+      assert {:ok, %Source{} = source} = Sources.create_source(valid_attrs)
+
+      assert source.index_frequency_minutes == 0
     end
   end
 
@@ -193,6 +231,14 @@ defmodule Pinchflat.SourcesTest do
       assert source.collection_name == "some updated name"
     end
 
+    test "updates with invalid data fails fast and does not call the runner" do
+      expect(YtDlpRunnerMock, :run, 0, &channel_mock/3)
+
+      source = source_fixture()
+
+      assert {:error, %Ecto.Changeset{}} = Sources.update_source(source, @invalid_source_attrs)
+    end
+
     test "updating the original_url will re-fetch the source details for channels" do
       expect(YtDlpRunnerMock, :run, &channel_mock/3)
 
@@ -230,7 +276,7 @@ defmodule Pinchflat.SourcesTest do
 
       assert {:ok, %Source{} = source} = Sources.update_source(source, update_attrs)
       assert source.index_frequency_minutes == 123
-      assert_enqueued(worker: MediaIndexingWorker, args: %{"id" => source.id})
+      assert_enqueued(worker: MediaCollectionIndexingWorker, args: %{"id" => source.id})
     end
 
     test "updating the index frequency to 0 will not re-schedule the indexing task" do
@@ -239,18 +285,25 @@ defmodule Pinchflat.SourcesTest do
 
       assert {:ok, %Source{}} = Sources.update_source(source, update_attrs)
 
-      refute_enqueued(worker: MediaIndexingWorker, args: %{"id" => source.id})
+      refute_enqueued(worker: MediaCollectionIndexingWorker, args: %{"id" => source.id})
     end
 
     test "updating the index frequency to 0 will delete any pending tasks" do
       source = source_fixture()
-      {:ok, job} = Oban.insert(MediaIndexingWorker.new(%{"id" => source.id}))
-      task = task_fixture(source_id: source.id, job_id: job.id)
       update_attrs = %{index_frequency_minutes: 0}
 
+      {:ok, job_1} = Oban.insert(FastIndexingWorker.new(%{"id" => source.id}))
+      task_1 = task_fixture(source_id: source.id, job_id: job_1.id)
+      {:ok, job_2} = Oban.insert(MediaIndexingWorker.new(%{"id" => source.id}))
+      task_2 = task_fixture(source_id: source.id, job_id: job_2.id)
+      {:ok, job_3} = Oban.insert(MediaCollectionIndexingWorker.new(%{"id" => source.id}))
+      task_3 = task_fixture(source_id: source.id, job_id: job_3.id)
+
       assert {:ok, %Source{}} = Sources.update_source(source, update_attrs)
 
-      assert_raise Ecto.NoResultsError, fn -> Repo.reload!(task) end
+      assert_raise Ecto.NoResultsError, fn -> Repo.reload!(task_1) end
+      assert_raise Ecto.NoResultsError, fn -> Repo.reload!(task_2) end
+      assert_raise Ecto.NoResultsError, fn -> Repo.reload!(task_3) end
     end
 
     test "not updating the index frequency will not re-schedule the indexing task or delete tasks" do
@@ -261,7 +314,7 @@ defmodule Pinchflat.SourcesTest do
       assert {:ok, %Source{}} = Sources.update_source(source, update_attrs)
 
       assert Repo.reload!(task)
-      refute_enqueued(worker: MediaIndexingWorker, args: %{"id" => source.id})
+      refute_enqueued(worker: MediaCollectionIndexingWorker, args: %{"id" => source.id})
     end
 
     test "enabling the download_media attribute will schedule a download task" do
@@ -285,6 +338,26 @@ defmodule Pinchflat.SourcesTest do
       refute_enqueued(worker: MediaDownloadWorker)
     end
 
+    test "enabling fast_index will schedule a fast indexing task" do
+      source = source_fixture(fast_index: false)
+      update_attrs = %{fast_index: true}
+
+      refute_enqueued(worker: FastIndexingWorker)
+      assert {:ok, %Source{}} = Sources.update_source(source, update_attrs)
+      assert_enqueued(worker: FastIndexingWorker, args: %{"id" => source.id})
+    end
+
+    test "disabling fast_index will cancel the fast indexing task" do
+      source = source_fixture(fast_index: true)
+      update_attrs = %{fast_index: false}
+      {:ok, job} = Oban.insert(FastIndexingWorker.new(%{"id" => source.id}))
+      task_fixture(source_id: source.id, job_id: job.id)
+
+      assert_enqueued(worker: FastIndexingWorker, args: %{"id" => source.id})
+      assert {:ok, %Source{}} = Sources.update_source(source, update_attrs)
+      refute_enqueued(worker: FastIndexingWorker)
+    end
+
     test "updates with invalid data returns error changeset" do
       source = source_fixture()
 
@@ -293,6 +366,24 @@ defmodule Pinchflat.SourcesTest do
 
       assert source == Sources.get_source!(source.id)
     end
+
+    test "fast_index forces the index frequency to be a default value" do
+      source = source_fixture(%{fast_index: true})
+      update_attrs = %{index_frequency_minutes: 0}
+
+      assert {:ok, source} = Sources.update_source(source, update_attrs)
+
+      assert source.index_frequency_minutes == Source.index_frequency_when_fast_indexing()
+    end
+
+    test "disabling fast index will not change the index frequency" do
+      source = source_fixture(%{fast_index: false})
+      update_attrs = %{index_frequency_minutes: 0}
+
+      assert {:ok, source} = Sources.update_source(source, update_attrs)
+
+      assert source.index_frequency_minutes == 0
+    end
   end
 
   describe "delete_source/2" do
@@ -353,7 +444,7 @@ defmodule Pinchflat.SourcesTest do
     end
   end
 
-  describe "change_source/2" do
+  describe "change_source/3" do
     test "it returns a changeset" do
       source = source_fixture()
 

test/pinchflat/tasks/media_items_tasks_test.exs

@@ -1,9 +1,19 @@
 defmodule Pinchflat.Tasks.MediaItemTasksTest do
   use Pinchflat.DataCase
 
+  import Mox
   import Pinchflat.MediaFixtures
+  import Pinchflat.SourcesFixtures
+  import Pinchflat.ProfilesFixtures
 
+  alias Pinchflat.Tasks
+  alias Pinchflat.Media.MediaItem
   alias Pinchflat.Tasks.MediaItemTasks
+  alias Pinchflat.Workers.MediaDownloadWorker
+
+  setup :verify_on_exit!
+
+  @media_url "https://www.youtube.com/watch?v=1234"
 
   describe "compute_and_save_media_filesize/1" do
     test "updates the media item with the file size" do
@@ -22,4 +32,72 @@ defmodule Pinchflat.Tasks.MediaItemTasksTest do
       assert {:error, _} = MediaItemTasks.compute_and_save_media_filesize(media_item)
     end
   end
+
+  describe "index_and_enqueue_download_for_media_item/2" do
+    setup do
+      stub(YtDlpRunnerMock, :run, fn _url, _opts, _ot ->
+        {:ok, media_attributes_return_fixture()}
+      end)
+
+      {:ok, [source: source_fixture()]}
+    end
+
+    test "creates a new media item based on the URL", %{source: source} do
+      assert Repo.aggregate(MediaItem, :count) == 0
+      assert {:ok, _} = MediaItemTasks.index_and_enqueue_download_for_media_item(source, @media_url)
+      assert Repo.aggregate(MediaItem, :count) == 1
+    end
+
+    test "won't duplicate media_items based on media_id and source", %{source: source} do
+      assert {:ok, mi_1} = MediaItemTasks.index_and_enqueue_download_for_media_item(source, @media_url)
+      assert {:ok, mi_2} = MediaItemTasks.index_and_enqueue_download_for_media_item(source, @media_url)
+
+      assert Repo.aggregate(MediaItem, :count) == 1
+      assert mi_1.id == mi_2.id
+    end
+
+    test "enqueues a download job", %{source: source} do
+      assert {:ok, media_item} = MediaItemTasks.index_and_enqueue_download_for_media_item(source, @media_url)
+
+      assert_enqueued(worker: MediaDownloadWorker, args: %{"id" => media_item.id})
+    end
+
+    test "creates a download task record", %{source: source} do
+      assert {:ok, media_item} = MediaItemTasks.index_and_enqueue_download_for_media_item(source, @media_url)
+
+      assert [_] = Tasks.list_tasks_for(:media_item_id, media_item.id, "MediaDownloadWorker")
+    end
+
+    test "does not enqueue a download job if the source does not allow it" do
+      source = source_fixture(%{download_media: false})
+
+      assert {:ok, _} = MediaItemTasks.index_and_enqueue_download_for_media_item(source, @media_url)
+
+      refute_enqueued(worker: MediaDownloadWorker)
+    end
+
+    test "does not enqueue a download job if the media item does not match the format rules" do
+      profile = media_profile_fixture(%{shorts_behaviour: :exclude})
+      source = source_fixture(%{media_profile_id: profile.id})
+
+      stub(YtDlpRunnerMock, :run, fn _url, _opts, _ot ->
+        output =
+          Phoenix.json_library().encode!(%{
+            id: "video2",
+            title: "Video 2",
+            webpage_url: "https://example.com/shorts/video2",
+            was_live: true,
+            description: "desc2",
+            aspect_ratio: 1.67,
+            duration: 345.67,
+            upload_date: "20210101"
+          })
+
+        {:ok, output}
+      end)
+
+      assert {:ok, _media_item} = MediaItemTasks.index_and_enqueue_download_for_media_item(source, @media_url)
+      refute_enqueued(worker: MediaDownloadWorker)
+    end
+  end
 end

test/pinchflat/tasks/source_tasks_test.exs

@@ -11,8 +11,10 @@ defmodule Pinchflat.Tasks.SourceTasksTest do
   alias Pinchflat.Tasks.Task
   alias Pinchflat.Tasks.SourceTasks
   alias Pinchflat.Media.MediaItem
-  alias Pinchflat.Workers.MediaIndexingWorker
+  alias Pinchflat.Workers.FastIndexingWorker
   alias Pinchflat.Workers.MediaDownloadWorker
+  alias Pinchflat.Workers.MediaIndexingWorker
+  alias Pinchflat.Workers.MediaCollectionIndexingWorker
 
   setup :verify_on_exit!
 
@@ -22,7 +24,7 @@ defmodule Pinchflat.Tasks.SourceTasksTest do
 
       assert {:ok, _} = SourceTasks.kickoff_indexing_task(source)
 
-      assert_enqueued(worker: MediaIndexingWorker, args: %{"id" => source.id})
+      assert_enqueued(worker: MediaCollectionIndexingWorker, args: %{"id" => source.id})
     end
 
     test "it creates and attaches a task" do
@@ -33,7 +35,17 @@ defmodule Pinchflat.Tasks.SourceTasksTest do
       assert task.source_id == source.id
     end
 
-    test "it deletes any pending tasks for the source" do
+    test "it deletes any pending media collection tasks for the source" do
+      source = source_fixture()
+      {:ok, job} = Oban.insert(MediaCollectionIndexingWorker.new(%{"id" => source.id}))
+      task = task_fixture(source_id: source.id, job_id: job.id)
+
+      assert {:ok, _} = SourceTasks.kickoff_indexing_task(source)
+
+      assert_raise Ecto.NoResultsError, fn -> Repo.reload!(task) end
+    end
+
+    test "it deletes any pending media tasks for the source" do
       source = source_fixture()
       {:ok, job} = Oban.insert(MediaIndexingWorker.new(%{"id" => source.id}))
       task = task_fixture(source_id: source.id, job_id: job.id)
@@ -42,6 +54,69 @@ defmodule Pinchflat.Tasks.SourceTasksTest do
 
       assert_raise Ecto.NoResultsError, fn -> Repo.reload!(task) end
     end
+
+    test "it deletes any fast indexing tasks for the source" do
+      source = source_fixture()
+      {:ok, job} = Oban.insert(FastIndexingWorker.new(%{"id" => source.id}))
+      task = task_fixture(source_id: source.id, job_id: job.id)
+
+      assert {:ok, _} = SourceTasks.kickoff_indexing_task(source)
+
+      assert_raise Ecto.NoResultsError, fn -> Repo.reload!(task) end
+    end
+  end
+
+  describe "kickoff_fast_indexing_task/1" do
+    test "it schedules a job" do
+      source = source_fixture()
+
+      assert {:ok, _} = SourceTasks.kickoff_fast_indexing_task(source)
+
+      assert_enqueued(worker: FastIndexingWorker, args: %{"id" => source.id})
+    end
+
+    test "it creates and attaches a task" do
+      source = source_fixture()
+
+      assert {:ok, %Task{} = task} = SourceTasks.kickoff_fast_indexing_task(source)
+
+      assert task.source_id == source.id
+    end
+
+    test "it deletes any fast indexing tasks for the source" do
+      source = source_fixture()
+      {:ok, job} = Oban.insert(FastIndexingWorker.new(%{"id" => source.id}))
+      task = task_fixture(source_id: source.id, job_id: job.id)
+
+      assert {:ok, _} = SourceTasks.kickoff_fast_indexing_task(source)
+
+      assert_raise Ecto.NoResultsError, fn -> Repo.reload!(task) end
+    end
+  end
+
+  describe "kickoff_indexing_tasks_from_youtube_rss_feed/1" do
+    setup do
+      {:ok, [source: source_fixture()]}
+    end
+
+    test "enqueues a new worker for each new media_id in the source's RSS feed", %{source: source} do
+      expect(HTTPClientMock, :get, fn _url -> {:ok, "<yt:videoId>test_1</yt:videoId>"} end)
+
+      assert :ok = SourceTasks.kickoff_indexing_tasks_from_youtube_rss_feed(source)
+
+      assert [worker] = all_enqueued(worker: MediaIndexingWorker)
+      assert worker.args["id"] == source.id
+      assert worker.args["media_url"] == "https://www.youtube.com/watch?v=test_1"
+    end
+
+    test "does not enqueue a new worker for the source's media IDs we already know about", %{source: source} do
+      expect(HTTPClientMock, :get, fn _url -> {:ok, "<yt:videoId>test_1</yt:videoId>"} end)
+      media_item_fixture(source_id: source.id, media_id: "test_1")
+
+      assert :ok = SourceTasks.kickoff_indexing_tasks_from_youtube_rss_feed(source)
+
+      refute_enqueued(worker: MediaIndexingWorker)
+    end
   end
 
   describe "index_and_enqueue_download_for_media_items/1" do
@@ -93,12 +168,14 @@ defmodule Pinchflat.Tasks.SourceTasksTest do
                Enum.map(media_items_other_source, & &1.media_id)
     end
 
-    test "it returns a list of media_items or changesets", %{source: source} do
+    test "it returns a list of media_items", %{source: source} do
       first_run = SourceTasks.index_and_enqueue_download_for_media_items(source)
       duplicate_run = SourceTasks.index_and_enqueue_download_for_media_items(source)
 
-      assert Enum.all?(first_run, fn %MediaItem{} -> true end)
+      first_ids = Enum.map(first_run, & &1.id)
-      assert Enum.all?(duplicate_run, fn %Ecto.Changeset{} -> true end)
+      duplicate_ids = Enum.map(duplicate_run, & &1.id)
+
+      assert first_ids == duplicate_ids
     end
 
     test "it updates the source's last_indexed_at field", %{source: source} do
@@ -203,9 +280,12 @@ defmodule Pinchflat.Tasks.SourceTasksTest do
           Phoenix.json_library().encode!(%{
             id: "video2",
             title: "Video 2",
-            original_url: "https://example.com/shorts/video2",
+            webpage_url: "https://example.com/shorts/video2",
             was_live: true,
-            description: "desc2"
+            description: "desc2",
+            aspect_ratio: 1.67,
+            duration: 345.67,
+            upload_date: "20210101"
           })
 
         File.write(filepath, contents)

test/pinchflat/workers/data_backfill_worker_test.exs

@@ -0,0 +1,70 @@
+defmodule Pinchflat.Workers.DataBackfillWorkerTest do
+  use Pinchflat.DataCase
+
+  import Pinchflat.MediaFixtures
+
+  alias Pinchflat.Workers.DataBackfillWorker
+  alias Pinchflat.Workers.FilesystemDataWorker
+
+  describe "cancel_pending_backfill_jobs/0" do
+    test "cancels all pending backfill jobs" do
+      %{}
+      |> DataBackfillWorker.new()
+      |> Repo.insert_unique_job()
+
+      assert_enqueued(worker: DataBackfillWorker)
+
+      DataBackfillWorker.cancel_pending_backfill_jobs()
+
+      refute_enqueued(worker: DataBackfillWorker)
+    end
+
+    test "does not cancel jobs for other workers" do
+      %{id: 0}
+      |> FilesystemDataWorker.new()
+      |> Repo.insert_unique_job()
+
+      assert_enqueued(worker: FilesystemDataWorker)
+
+      DataBackfillWorker.cancel_pending_backfill_jobs()
+
+      assert_enqueued(worker: FilesystemDataWorker)
+    end
+  end
+
+  describe "perform/1" do
+    setup do
+      DataBackfillWorker.cancel_pending_backfill_jobs()
+
+      :ok
+    end
+
+    test "reschedules itself once complete" do
+      perform_job(DataBackfillWorker, %{})
+
+      assert_enqueued(worker: DataBackfillWorker, scheduled_at: now_plus(60, :minutes))
+    end
+  end
+
+  describe "perform/1 when testing backfill_shorts_data" do
+    test "sets short_form_content to true for media items with shorts in the URL" do
+      media_item = media_item_with_attachments(%{original_url: "https://example.com/shorts/123"})
+
+      refute media_item.short_form_content
+
+      perform_job(DataBackfillWorker, %{})
+
+      assert Repo.reload!(media_item).short_form_content
+    end
+
+    test "does not set short_form_content to true for media items without shorts in the URL" do
+      media_item = media_item_with_attachments(%{original_url: "https://example.com/longs/123"})
+
+      refute media_item.short_form_content
+
+      perform_job(DataBackfillWorker, %{})
+
+      refute Repo.reload!(media_item).short_form_content
+    end
+  end
+end

test/pinchflat/workers/fast_indexing_worker_test.exs

@@ -0,0 +1,46 @@
+defmodule Pinchflat.Workers.FastIndexingWorkerTest do
+  use Pinchflat.DataCase
+
+  import Mox
+  import Pinchflat.SourcesFixtures
+
+  alias Pinchflat.Sources.Source
+  alias Pinchflat.Workers.FastIndexingWorker
+
+  setup :verify_on_exit!
+
+  describe "perform/1" do
+    test "calls out to Youtube RSS if enabled" do
+      expect(HTTPClientMock, :get, fn _url -> {:ok, ""} end)
+      source = source_fixture(fast_index: true)
+
+      perform_job(FastIndexingWorker, %{"id" => source.id})
+    end
+
+    test "reschedules itself if fast indexing is enabled" do
+      expect(HTTPClientMock, :get, fn _url -> {:ok, ""} end)
+      source = source_fixture(fast_index: true)
+      perform_job(FastIndexingWorker, %{"id" => source.id})
+
+      assert_enqueued(
+        worker: FastIndexingWorker,
+        args: %{"id" => source.id},
+        scheduled_at: now_plus(Source.fast_index_frequency(), :minutes)
+      )
+    end
+
+    test "does not call out to Youtube RSS if disabled" do
+      expect(HTTPClientMock, :get, 0, fn _url -> {:ok, ""} end)
+      source = source_fixture(fast_index: false)
+
+      perform_job(FastIndexingWorker, %{"id" => source.id})
+    end
+
+    test "does not reschedule itself if fast indexing is disabled" do
+      source = source_fixture(fast_index: false)
+      perform_job(FastIndexingWorker, %{"id" => source.id})
+
+      refute_enqueued(worker: FastIndexingWorker, args: %{"id" => source.id})
+    end
+  end
+end

test/pinchflat/workers/media_collection_indexing_worker_test.exs

@@ -0,0 +1,162 @@
+defmodule Pinchflat.Workers.MediaCollectionIndexingWorkerTest do
+  use Pinchflat.DataCase
+
+  import Mox
+  import Pinchflat.TasksFixtures
+  import Pinchflat.MediaFixtures
+  import Pinchflat.SourcesFixtures
+
+  alias Pinchflat.Tasks
+  alias Pinchflat.Sources.Source
+  alias Pinchflat.Workers.FastIndexingWorker
+  alias Pinchflat.Workers.MediaDownloadWorker
+  alias Pinchflat.Workers.MediaCollectionIndexingWorker
+
+  setup :verify_on_exit!
+
+  describe "perform/1" do
+    test "it indexes the source if it should be indexed" do
+      expect(YtDlpRunnerMock, :run, fn _url, _opts, _ot, _addl_opts -> {:ok, ""} end)
+
+      source = source_fixture(index_frequency_minutes: 10)
+
+      perform_job(MediaCollectionIndexingWorker, %{id: source.id})
+    end
+
+    test "it indexes the source no matter what if the source has never been indexed before" do
+      expect(YtDlpRunnerMock, :run, fn _url, _opts, _ot, _addl_opts -> {:ok, ""} end)
+
+      source = source_fixture(index_frequency_minutes: 0, last_indexed_at: nil)
+
+      perform_job(MediaCollectionIndexingWorker, %{id: source.id})
+    end
+
+    test "it does not do any indexing if the source has been indexed and shouldn't be rescheduled" do
+      expect(YtDlpRunnerMock, :run, 0, fn _url, _opts, _ot, _addl_opts -> {:ok, ""} end)
+
+      source = source_fixture(index_frequency_minutes: -1, last_indexed_at: DateTime.utc_now())
+
+      perform_job(MediaCollectionIndexingWorker, %{id: source.id})
+    end
+
+    test "it does not reschedule if the source shouldn't be indexed" do
+      stub(YtDlpRunnerMock, :run, fn _url, _opts, _ot, _addl_opts -> {:ok, ""} end)
+
+      source = source_fixture(index_frequency_minutes: -1)
+      perform_job(MediaCollectionIndexingWorker, %{id: source.id})
+
+      refute_enqueued(worker: MediaCollectionIndexingWorker, args: %{"id" => source.id})
+    end
+
+    test "it kicks off a download job for each pending media item" do
+      expect(YtDlpRunnerMock, :run, fn _url, _opts, _ot, _addl_opts ->
+        {:ok, source_attributes_return_fixture()}
+      end)
+
+      source = source_fixture(index_frequency_minutes: 10)
+      perform_job(MediaCollectionIndexingWorker, %{id: source.id})
+
+      assert length(all_enqueued(worker: MediaDownloadWorker)) == 3
+    end
+
+    test "it starts a job for any pending media item even if it's from another run" do
+      expect(YtDlpRunnerMock, :run, fn _url, _opts, _ot, _addl_opts ->
+        {:ok, source_attributes_return_fixture()}
+      end)
+
+      source = source_fixture(index_frequency_minutes: 10)
+      media_item_fixture(%{source_id: source.id, media_filepath: nil})
+      perform_job(MediaCollectionIndexingWorker, %{id: source.id})
+
+      assert length(all_enqueued(worker: MediaDownloadWorker)) == 4
+    end
+
+    test "it does not kick off a job for media items that could not be saved" do
+      expect(YtDlpRunnerMock, :run, fn _url, _opts, _ot, _addl_opts ->
+        {:ok, source_attributes_return_fixture()}
+      end)
+
+      source = source_fixture(index_frequency_minutes: 10)
+      media_item_fixture(%{source_id: source.id, media_filepath: nil, media_id: "video1"})
+      perform_job(MediaCollectionIndexingWorker, %{id: source.id})
+
+      # Only 3 jobs should be enqueued, since the first video is a duplicate
+      assert length(all_enqueued(worker: MediaDownloadWorker))
+    end
+
+    test "it reschedules the job based on the index frequency" do
+      expect(YtDlpRunnerMock, :run, fn _url, _opts, _ot, _addl_opts -> {:ok, ""} end)
+
+      source = source_fixture(index_frequency_minutes: 10)
+      perform_job(MediaCollectionIndexingWorker, %{id: source.id})
+
+      assert_enqueued(
+        worker: MediaCollectionIndexingWorker,
+        args: %{"id" => source.id},
+        scheduled_at: now_plus(source.index_frequency_minutes, :minutes)
+      )
+    end
+
+    test "it creates a task for the rescheduled job" do
+      expect(YtDlpRunnerMock, :run, fn _url, _opts, _ot, _addl_opts -> {:ok, ""} end)
+
+      source = source_fixture(index_frequency_minutes: 10)
+      task_count_fetcher = fn -> Enum.count(Tasks.list_tasks()) end
+
+      assert_changed([from: 0, to: 1], task_count_fetcher, fn ->
+        perform_job(MediaCollectionIndexingWorker, %{id: source.id})
+      end)
+    end
+
+    test "it creates a future task for fast indexing if appropriate" do
+      expect(YtDlpRunnerMock, :run, fn _url, _opts, _ot, _addl_opts -> {:ok, ""} end)
+
+      source = source_fixture(index_frequency_minutes: 10, fast_index: true)
+      perform_job(MediaCollectionIndexingWorker, %{id: source.id})
+
+      assert_enqueued(
+        worker: FastIndexingWorker,
+        args: %{"id" => source.id},
+        scheduled_at: now_plus(Source.fast_index_frequency(), :minutes)
+      )
+    end
+
+    test "it deletes existing fast indexing tasks if a new one is created" do
+      expect(YtDlpRunnerMock, :run, fn _url, _opts, _ot, _addl_opts -> {:ok, ""} end)
+
+      source = source_fixture(index_frequency_minutes: 10, fast_index: true)
+      {:ok, job} = Oban.insert(FastIndexingWorker.new(%{"id" => source.id}))
+      task = task_fixture(source_id: source.id, job_id: job.id)
+
+      perform_job(MediaCollectionIndexingWorker, %{id: source.id})
+
+      assert_raise Ecto.NoResultsError, fn -> Repo.reload!(task) end
+    end
+
+    test "it does not create a task for fast indexing otherwise" do
+      expect(YtDlpRunnerMock, :run, fn _url, _opts, _ot, _addl_opts -> {:ok, ""} end)
+
+      source = source_fixture(index_frequency_minutes: 10, fast_index: false)
+      perform_job(MediaCollectionIndexingWorker, %{id: source.id})
+
+      refute_enqueued(worker: FastIndexingWorker)
+    end
+
+    test "it creates the basic media_item records" do
+      expect(YtDlpRunnerMock, :run, fn _url, _opts, _ot, _addl_opts -> {:ok, source_attributes_return_fixture()} end)
+
+      source = source_fixture(index_frequency_minutes: 10)
+
+      media_item_fetcher = fn ->
+        source
+        |> Repo.preload(:media_items)
+        |> Map.get(:media_items)
+        |> Enum.map(fn media_item -> media_item.media_id end)
+      end
+
+      assert_changed([from: [], to: ["video1", "video2", "video3"]], media_item_fetcher, fn ->
+        perform_job(MediaCollectionIndexingWorker, %{id: source.id})
+      end)
+    end
+  end
+end

test/pinchflat/workers/media_indexing_worker_test.exs

@@ -5,121 +5,40 @@ defmodule Pinchflat.Workers.MediaIndexingWorkerTest do
   import Pinchflat.MediaFixtures
   import Pinchflat.SourcesFixtures
 
-  alias Pinchflat.Tasks
+  alias Pinchflat.Media.MediaItem
   alias Pinchflat.Workers.MediaIndexingWorker
   alias Pinchflat.Workers.MediaDownloadWorker
 
+  @media_url "https://www.youtube.com/watch?v=1234567890"
+
   setup :verify_on_exit!
 
+  setup do
+    source = source_fixture()
+
+    {:ok, source: source}
+  end
+
   describe "perform/1" do
-    test "it indexes the source if it should be indexed" do
+    test "indexes the media item and saves it to the database", %{source: source} do
-      expect(YtDlpRunnerMock, :run, fn _url, _opts, _ot, _addl_opts -> {:ok, ""} end)
+      expect(YtDlpRunnerMock, :run, fn _url, _opts, _ot ->
-
+        {:ok, media_attributes_return_fixture()}
-      source = source_fixture(index_frequency_minutes: 10)
-
-      perform_job(MediaIndexingWorker, %{id: source.id})
-    end
-
-    test "it indexes the source no matter what if the source has never been indexed before" do
-      expect(YtDlpRunnerMock, :run, fn _url, _opts, _ot, _addl_opts -> {:ok, ""} end)
-
-      source = source_fixture(index_frequency_minutes: 0, last_indexed_at: nil)
-
-      perform_job(MediaIndexingWorker, %{id: source.id})
-    end
-
-    test "it does not do any indexing if the source has been indexed and shouldn't be rescheduled" do
-      expect(YtDlpRunnerMock, :run, 0, fn _url, _opts, _ot, _addl_opts -> {:ok, ""} end)
-
-      source = source_fixture(index_frequency_minutes: -1, last_indexed_at: DateTime.utc_now())
-
-      perform_job(MediaIndexingWorker, %{id: source.id})
-    end
-
-    test "it does not reschedule if the source shouldn't be indexed" do
-      stub(YtDlpRunnerMock, :run, fn _url, _opts, _ot, _addl_opts -> {:ok, ""} end)
-
-      source = source_fixture(index_frequency_minutes: -1)
-      perform_job(MediaIndexingWorker, %{id: source.id})
-
-      refute_enqueued(worker: MediaIndexingWorker, args: %{"id" => source.id})
-    end
-
-    test "it kicks off a download job for each pending media item" do
-      expect(YtDlpRunnerMock, :run, fn _url, _opts, _ot, _addl_opts ->
-        {:ok, source_attributes_return_fixture()}
       end)
 
-      source = source_fixture(index_frequency_minutes: 10)
+      before = Repo.aggregate(MediaItem, :count, :id)
-      perform_job(MediaIndexingWorker, %{id: source.id})
+      perform_job(MediaIndexingWorker, %{id: source.id, media_url: @media_url})
 
-      assert length(all_enqueued(worker: MediaDownloadWorker)) == 3
+      assert Repo.aggregate(MediaItem, :count, :id) == before + 1
     end
 
-    test "it starts a job for any pending media item even if it's from another run" do
+    test "enqueues a download job for the media item", %{source: source} do
-      expect(YtDlpRunnerMock, :run, fn _url, _opts, _ot, _addl_opts ->
+      expect(YtDlpRunnerMock, :run, fn _url, _opts, _ot ->
-        {:ok, source_attributes_return_fixture()}
+        {:ok, media_attributes_return_fixture()}
       end)
 
-      source = source_fixture(index_frequency_minutes: 10)
+      perform_job(MediaIndexingWorker, %{id: source.id, media_url: @media_url})
-      media_item_fixture(%{source_id: source.id, media_filepath: nil})
-      perform_job(MediaIndexingWorker, %{id: source.id})
 
-      assert length(all_enqueued(worker: MediaDownloadWorker)) == 4
+      assert [_] = all_enqueued(worker: MediaDownloadWorker)
-    end
-
-    test "it does not kick off a job for media items that could not be saved" do
-      expect(YtDlpRunnerMock, :run, fn _url, _opts, _ot, _addl_opts ->
-        {:ok, source_attributes_return_fixture()}
-      end)
-
-      source = source_fixture(index_frequency_minutes: 10)
-      media_item_fixture(%{source_id: source.id, media_filepath: nil, media_id: "video1"})
-      perform_job(MediaIndexingWorker, %{id: source.id})
-
-      # Only 3 jobs should be enqueued, since the first video is a duplicate
-      assert length(all_enqueued(worker: MediaDownloadWorker))
-    end
-
-    test "it reschedules the job based on the index frequency" do
-      expect(YtDlpRunnerMock, :run, fn _url, _opts, _ot, _addl_opts -> {:ok, ""} end)
-
-      source = source_fixture(index_frequency_minutes: 10)
-      perform_job(MediaIndexingWorker, %{id: source.id})
-
-      assert_enqueued(
-        worker: MediaIndexingWorker,
-        args: %{"id" => source.id},
-        scheduled_at: now_plus(source.index_frequency_minutes, :minutes)
-      )
-    end
-
-    test "it creates a task for the rescheduled job" do
-      expect(YtDlpRunnerMock, :run, fn _url, _opts, _ot, _addl_opts -> {:ok, ""} end)
-
-      source = source_fixture(index_frequency_minutes: 10)
-      task_count_fetcher = fn -> Enum.count(Tasks.list_tasks()) end
-
-      assert_changed([from: 0, to: 1], task_count_fetcher, fn ->
-        perform_job(MediaIndexingWorker, %{id: source.id})
-      end)
-    end
-
-    test "it creates the basic media_item records" do
-      expect(YtDlpRunnerMock, :run, fn _url, _opts, _ot, _addl_opts -> {:ok, source_attributes_return_fixture()} end)
-
-      source = source_fixture(index_frequency_minutes: 10)
-
-      media_item_fetcher = fn ->
-        source
-        |> Repo.preload(:media_items)
-        |> Map.get(:media_items)
-        |> Enum.map(fn media_item -> media_item.media_id end)
-      end
-
-      assert_changed([from: [], to: ["video1", "video2", "video3"]], media_item_fetcher, fn ->
-        perform_job(MediaIndexingWorker, %{id: source.id})
-      end)
     end
   end
 end

test/pinchflat/yt_dlp/backend/command_runner_test.exs

@@ -1,7 +1,7 @@
-defmodule Pinchflat.MediaClient.Backends.YtDlp.CommandRunnerTest do
+defmodule Pinchflat.YtDlp.Backend.CommandRunnerTest do
   use ExUnit.Case, async: true
 
-  alias Pinchflat.MediaClient.Backends.YtDlp.CommandRunner, as: Runner
+  alias Pinchflat.YtDlp.Backend.CommandRunner, as: Runner
 
   @original_executable Application.compile_env(:pinchflat, :yt_dlp_executable)
   @media_url "https://www.youtube.com/watch?v=-LHXuyzpex0"

test/pinchflat/yt_dlp/backend/media_collection_test.exs

@@ -1,39 +1,40 @@
-defmodule Pinchflat.MediaClient.Backends.YtDlp.MediaCollectionTest do
+defmodule Pinchflat.YtDlp.Backend.MediaCollectionTest do
   use Pinchflat.DataCase
   import Mox
   import Pinchflat.SourcesFixtures
 
-  alias Pinchflat.MediaClient.Backends.YtDlp.MediaCollection
+  alias Pinchflat.YtDlp.Backend.Media
+  alias Pinchflat.YtDlp.Backend.MediaCollection
 
   @channel_url "https://www.youtube.com/c/TheUselessTrials"
 
   setup :verify_on_exit!
 
-  describe "get_media_attributes/2" do
+  describe "get_media_attributes_for_collection/2" do
     test "returns a list of video attributes with no blank elements" do
       expect(YtDlpRunnerMock, :run, fn _url, _opts, _ot, _addl_opts ->
         {:ok, source_attributes_return_fixture() <> "\n\n"}
       end)
 
-      assert {:ok, [%{"id" => "video1"}, %{"id" => "video2"}, %{"id" => "video3"}]} =
+      assert {:ok, [%Media{media_id: "video1"}, %Media{media_id: "video2"}, %Media{media_id: "video3"}]} =
-               MediaCollection.get_media_attributes(@channel_url)
+               MediaCollection.get_media_attributes_for_collection(@channel_url)
     end
 
     test "it passes the expected default args" do
       expect(YtDlpRunnerMock, :run, fn _url, opts, ot, _addl_opts ->
         assert opts == [:simulate, :skip_download]
-        assert ot == "%(.{id,title,was_live,original_url,description})j"
+        assert ot == Media.indexing_output_template()
 
         {:ok, ""}
       end)
 
-      assert {:ok, _} = MediaCollection.get_media_attributes(@channel_url)
+      assert {:ok, _} = MediaCollection.get_media_attributes_for_collection(@channel_url)
     end
 
     test "returns the error straight through when the command fails" do
       expect(YtDlpRunnerMock, :run, fn _url, _opts, _ot, _addl_opts -> {:error, "Big issue", 1} end)
 
-      assert {:error, "Big issue", 1} = MediaCollection.get_media_attributes(@channel_url)
+      assert {:error, "Big issue", 1} = MediaCollection.get_media_attributes_for_collection(@channel_url)
     end
 
     test "passes the explict tmpfile path to runner" do
@@ -44,7 +45,7 @@ defmodule Pinchflat.MediaClient.Backends.YtDlp.MediaCollectionTest do
         {:ok, ""}
       end)
 
-      assert {:ok, _} = MediaCollection.get_media_attributes(@channel_url)
+      assert {:ok, _} = MediaCollection.get_media_attributes_for_collection(@channel_url)
     end
 
     test "supports an optional file_listener_handler that gets passed a filename" do
@@ -55,7 +56,8 @@ defmodule Pinchflat.MediaClient.Backends.YtDlp.MediaCollectionTest do
         send(current_self, {:handler, filename})
       end
 
-      assert {:ok, _} = MediaCollection.get_media_attributes(@channel_url, file_listener_handler: handler)
+      assert {:ok, _} =
+               MediaCollection.get_media_attributes_for_collection(@channel_url, file_listener_handler: handler)
 
       assert_receive {:handler, filename}
       assert String.ends_with?(filename, ".json")

test/pinchflat/yt_dlp/backend/media_test.exs

@@ -0,0 +1,157 @@
+defmodule Pinchflat.YtDlp.Backend.MediaTest do
+  use Pinchflat.DataCase
+  import Mox
+  import Pinchflat.MediaFixtures
+
+  alias Pinchflat.YtDlp.Backend.Media
+
+  @media_url "https://www.youtube.com/watch?v=TiZPUDkDYbk"
+
+  setup :verify_on_exit!
+
+  describe "download/2" do
+    test "it calls the backend runner with the expected arguments" do
+      expect(YtDlpRunnerMock, :run, fn @media_url, opts, ot ->
+        assert [:no_simulate] = opts
+        assert "after_move:%()j" = ot
+
+        {:ok, render_metadata(:media_metadata)}
+      end)
+
+      assert {:ok, _} = Media.download(@media_url)
+    end
+
+    test "it passes along additional options" do
+      expect(YtDlpRunnerMock, :run, fn _url, opts, _ot ->
+        assert [:no_simulate, :custom_arg] = opts
+
+        {:ok, "{}"}
+      end)
+
+      assert {:ok, _} = Media.download(@media_url, [:custom_arg])
+    end
+
+    test "it parses and returns the generated file as JSON" do
+      expect(YtDlpRunnerMock, :run, fn _url, _opts, _ot ->
+        {:ok, render_metadata(:media_metadata)}
+      end)
+
+      assert {:ok, %{"title" => "Trying to Wheelie Without the Rear Brake"}} =
+               Media.download(@media_url)
+    end
+
+    test "it returns errors" do
+      expect(YtDlpRunnerMock, :run, fn _url, _opt, _ot ->
+        {:error, "something"}
+      end)
+
+      assert {:error, "something"} = Media.download(@media_url)
+    end
+  end
+
+  describe "get_media_attributes/1" do
+    test "returns a list of video attributes" do
+      expect(YtDlpRunnerMock, :run, fn _url, _opts, _ot ->
+        {:ok, media_attributes_return_fixture()}
+      end)
+
+      assert {:ok, %{description: _, media_id: _, original_url: _, title: _, livestream: _}} =
+               Media.get_media_attributes(@media_url)
+    end
+
+    test "it passes the expected default args" do
+      expect(YtDlpRunnerMock, :run, fn _url, opts, ot ->
+        assert opts == [:simulate, :skip_download]
+        assert ot == Media.indexing_output_template()
+
+        {:ok, media_attributes_return_fixture()}
+      end)
+
+      assert {:ok, _} = Media.get_media_attributes(@media_url)
+    end
+
+    test "returns the error straight through when the command fails" do
+      expect(YtDlpRunnerMock, :run, fn _url, _opts, _ot -> {:error, "Big issue", 1} end)
+
+      assert {:error, "Big issue", 1} = Media.get_media_attributes(@media_url)
+    end
+  end
+
+  describe "indexing_output_template/0" do
+    test "contains all the greatest hits" do
+      assert "%(.{id,title,was_live,webpage_url,description,aspect_ratio,duration,upload_date})j" ==
+               Media.indexing_output_template()
+    end
+  end
+
+  describe "response_to_struct/1" do
+    test "transforms a response into a struct" do
+      response = %{
+        "id" => "TiZPUDkDYbk",
+        "title" => "Trying to Wheelie Without the Rear Brake",
+        "description" => "I'm not sure what I expected.",
+        "webpage_url" => "https://www.youtube.com/watch?v=TiZPUDkDYbk",
+        "was_live" => false,
+        "aspect_ratio" => 1.0,
+        "duration" => 60,
+        "upload_date" => "20210101"
+      }
+
+      assert %Media{
+               media_id: "TiZPUDkDYbk",
+               title: "Trying to Wheelie Without the Rear Brake",
+               description: "I'm not sure what I expected.",
+               original_url: "https://www.youtube.com/watch?v=TiZPUDkDYbk",
+               livestream: false,
+               short_form_content: false,
+               upload_date: Date.from_iso8601!("2021-01-01")
+             } == Media.response_to_struct(response)
+    end
+
+    test "sets short_form_content to true if the URL contains /shorts/" do
+      response = %{
+        "webpage_url" => "https://www.youtube.com/shorts/TiZPUDkDYbk",
+        "aspect_ratio" => 1.0,
+        "duration" => 61,
+        "upload_date" => "20210101"
+      }
+
+      assert %Media{short_form_content: true} = Media.response_to_struct(response)
+    end
+
+    test "sets short_form_content to true if the aspect ratio are duration are right" do
+      response = %{
+        "webpage_url" => "https://www.youtube.com/watch?v=TiZPUDkDYbk",
+        "aspect_ratio" => 0.5,
+        "duration" => 59,
+        "upload_date" => "20210101"
+      }
+
+      assert %Media{short_form_content: true} = Media.response_to_struct(response)
+    end
+
+    test "sets short_form_content to false otherwise" do
+      response = %{
+        "webpage_url" => "https://www.youtube.com/watch?v=TiZPUDkDYbk",
+        "aspect_ratio" => 1.0,
+        "duration" => 61,
+        "upload_date" => "20210101"
+      }
+
+      assert %Media{short_form_content: false} = Media.response_to_struct(response)
+    end
+
+    test "parses the upload date" do
+      response = %{
+        "webpage_url" => "https://www.youtube.com/watch?v=TiZPUDkDYbk",
+        "aspect_ratio" => 1.0,
+        "duration" => 61,
+        "upload_date" => "20210101"
+      }
+
+      expected_date = Date.from_iso8601!("2021-01-01")
+
+      assert %Media{upload_date: ^expected_date} = Media.response_to_struct(response)
+    end
+  end
+end

test/pinchflat/yt_dlp/download_option_builder_test.exs

@@ -1,11 +1,11 @@
-defmodule Pinchflat.Profiles.Options.YtDlp.DownloadOptionBuilderTest do
+defmodule Pinchflat.YtDlp.DownloadOptionBuilderTest do
   use Pinchflat.DataCase
   import Pinchflat.MediaFixtures
   import Pinchflat.ProfilesFixtures
   import Pinchflat.SourcesFixtures
 
   alias Pinchflat.Profiles
-  alias Pinchflat.Profiles.Options.YtDlp.DownloadOptionBuilder
+  alias Pinchflat.YtDlp.DownloadOptionBuilder
 
   setup do
     media_profile = media_profile_fixture(%{output_path_template: "{{ title }}.%(ext)s"})
@@ -125,6 +125,14 @@ defmodule Pinchflat.Profiles.Options.YtDlp.DownloadOptionBuilderTest do
       assert :write_thumbnail in res
     end
 
+    test "convertes thumbnail to jpg when download_thumbnail is true", %{media_item: media_item} do
+      media_item = update_media_profile_attribute(media_item, %{download_thumbnail: true})
+
+      assert {:ok, res} = DownloadOptionBuilder.build(media_item)
+
+      assert {:convert_thumbnail, "jpg"} in res
+    end
+
     test "includes :embed_thumbnail option when specified", %{media_item: media_item} do
       media_item = update_media_profile_attribute(media_item, %{embed_thumbnail: true})
 
@@ -133,14 +141,6 @@ defmodule Pinchflat.Profiles.Options.YtDlp.DownloadOptionBuilderTest do
       assert :embed_thumbnail in res
     end
 
-    test "doesn't include :embed_thumbnail option when preferred_resolution is :audio", %{media_item: media_item} do
-      media_item = update_media_profile_attribute(media_item, %{embed_thumbnail: true, preferred_resolution: :audio})
-
-      assert {:ok, res} = DownloadOptionBuilder.build(media_item)
-
-      refute :embed_thumbnail in res
-    end
-
     test "doesn't include these options when not specified", %{media_item: media_item} do
       media_item = update_media_profile_attribute(media_item, %{embed_thumbnail: false, download_thumbnail: false})
 
@@ -169,14 +169,6 @@ defmodule Pinchflat.Profiles.Options.YtDlp.DownloadOptionBuilderTest do
       assert :embed_metadata in res
     end
 
-    test "doesn't include :embed_metadata option when preferred_resolution is :audio", %{media_item: media_item} do
-      media_item = update_media_profile_attribute(media_item, %{embed_metadata: true, preferred_resolution: :audio})
-
-      assert {:ok, res} = DownloadOptionBuilder.build(media_item)
-
-      refute :embed_metadata in res
-    end
-
     test "doesn't include these options when not specified", %{media_item: media_item} do
       media_item = update_media_profile_attribute(media_item, %{embed_metadata: false, download_metadata: false})
 
@@ -202,8 +194,8 @@ defmodule Pinchflat.Profiles.Options.YtDlp.DownloadOptionBuilderTest do
 
       assert {:ok, res} = DownloadOptionBuilder.build(media_item)
 
-      assert {:format, "bestaudio"} in res
+      assert :extract_audio in res
-      assert {:format_sort, "ext"} in res
+      assert {:format, "bestaudio[ext=m4a]"} in res
     end
   end
 

test/pinchflat_web/controllers/page_controller_test.exs

@@ -1,51 +1,31 @@
 defmodule PinchflatWeb.PageControllerTest do
   use PinchflatWeb.ConnCase
 
-  import Pinchflat.ProfilesFixtures
-  import Pinchflat.SourcesFixtures
-
   alias Pinchflat.Settings
 
   describe "GET / when testing onboarding" do
-    test "sets the onboarding session to true when onboarding", %{conn: conn} do
+    test "sets the onboarding setting to true when onboarding", %{conn: conn} do
       _conn = get(conn, ~p"/")
       assert Settings.get!(:onboarding)
     end
 
-    test "displays the onboarding page when no media profiles exist", %{conn: conn} do
-      conn = get(conn, ~p"/")
-      assert html_response(conn, 200) =~ "Welcome to Pinchflat"
-    end
-
-    test "displays the onboarding page when no sources exist", %{conn: conn} do
-      _ = media_profile_fixture()
-
-      conn = get(conn, ~p"/")
-      assert html_response(conn, 200) =~ "Welcome to Pinchflat"
-    end
-
     test "displays the onboarding page when onboarding is forced", %{conn: conn} do
-      _ = media_profile_fixture()
+      Settings.set!(:onboarding, false)
-      _ = source_fixture()
 
       conn = get(conn, ~p"/?onboarding=1")
       assert html_response(conn, 200) =~ "Welcome to Pinchflat"
     end
 
-    test "sets the onboarding session to false when not onboarding", %{conn: conn} do
+    test "sets the onboarding setting to false if you pass the corrent query param", %{conn: conn} do
       conn = get(conn, ~p"/")
       assert Settings.get!(:onboarding)
 
-      _ = media_profile_fixture()
+      _conn = get(conn, ~p"/?onboarding=0")
-      _ = source_fixture()
-
-      _conn = get(conn, ~p"/")
       refute Settings.get!(:onboarding)
     end
 
     test "displays the home page when not onboarding", %{conn: conn} do
-      _ = media_profile_fixture()
+      Settings.set!(:onboarding, false)
-      _ = source_fixture()
 
       conn = get(conn, ~p"/")
       assert html_response(conn, 200) =~ "MENU"

test/profiles/output_path_builder_test.exs

@@ -1,7 +1,7 @@
-defmodule Pinchflat.MediaClient.Backends.YtDlp.OutputPathBuilderTest do
+defmodule Pinchflat.Profiles.OutputPathBuilderTest do
   use Pinchflat.DataCase
 
-  alias Pinchflat.Profiles.Options.YtDlp.OutputPathBuilder
+  alias Pinchflat.Profiles.OutputPathBuilder
 
   describe "build/2" do
     test "it expands 'standard' curly brace variables in the template" do

test/support/fixtures/media_fixtures.ex

@@ -19,8 +19,10 @@ defmodule Pinchflat.MediaFixtures do
         title: Faker.Commerce.product_name(),
         original_url: "https://www.youtube.com/watch?v=#{media_id}",
         livestream: false,
+        short_form_content: false,
         media_filepath: "/video/#{Faker.File.file_name(:video)}",
-        source_id: SourcesFixtures.source_fixture().id
+        source_id: SourcesFixtures.source_fixture().id,
+        upload_date: DateTime.utc_now()
       })
       |> Pinchflat.Media.create_media_item()
 
@@ -65,4 +67,19 @@ defmodule Pinchflat.MediaFixtures do
     merged_attrs = Map.merge(attrs, %{media_filepath: stored_media_filepath})
     media_item_fixture(merged_attrs)
   end
+
+  def media_attributes_return_fixture do
+    media_attributes = %{
+      id: "video1",
+      title: "Video 1",
+      webpage_url: "https://example.com/video1",
+      was_live: false,
+      description: "desc1",
+      aspect_ratio: 1.67,
+      duration: 123.45,
+      upload_date: "20210101"
+    }
+
+    Phoenix.json_library().encode!(media_attributes)
+  end
 end

test/support/fixtures/sources_fixtures.ex

@@ -15,15 +15,19 @@ defmodule Pinchflat.SourcesFixtures do
     {:ok, source} =
       %Source{}
       |> Source.changeset(
-        Enum.into(attrs, %{
+        Enum.into(
-          collection_name: "Source ##{:rand.uniform(1_000_000)}",
+          attrs,
-          collection_id: Base.encode16(:crypto.hash(:md5, "#{:rand.uniform(1_000_000)}")),
+          %{
-          collection_type: "channel",
+            collection_name: "Source ##{:rand.uniform(1_000_000)}",
-          custom_name: "Cool and good internal name!",
+            collection_id: Base.encode16(:crypto.hash(:md5, "#{:rand.uniform(1_000_000)}")),
-          original_url: "https://www.youtube.com/channel/#{Faker.String.base64(12)}",
+            collection_type: "channel",
-          media_profile_id: ProfilesFixtures.media_profile_fixture().id,
+            custom_name: "Cool and good internal name!",
-          index_frequency_minutes: 60
+            original_url: "https://www.youtube.com/channel/#{Faker.String.base64(12)}",
-        })
+            media_profile_id: ProfilesFixtures.media_profile_fixture().id,
+            index_frequency_minutes: 60
+          }
+        ),
+        :pre_insert
       )
       |> Repo.insert()
 
@@ -35,23 +39,32 @@ defmodule Pinchflat.SourcesFixtures do
       %{
         id: "video1",
         title: "Video 1",
-        original_url: "https://example.com/video1",
+        webpage_url: "https://example.com/video1",
         was_live: false,
-        description: "desc1"
+        description: "desc1",
+        aspect_ratio: 1.67,
+        duration: 12.34,
+        upload_date: "20210101"
       },
       %{
         id: "video2",
         title: "Video 2",
-        original_url: "https://example.com/video2",
+        webpage_url: "https://example.com/video2",
         was_live: true,
-        description: "desc2"
+        description: "desc2",
+        aspect_ratio: 1.67,
+        duration: 345.67,
+        upload_date: "20220202"
       },
       %{
         id: "video3",
         title: "Video 3",
-        original_url: "https://example.com/video3",
+        webpage_url: "https://example.com/video3",
         was_live: false,
-        description: "desc3"
+        description: "desc3",
+        aspect_ratio: 1.0,
+        duration: 678.90,
+        upload_date: "20230303"
       }
     ]
 

test/support/testing_helper_methods.ex

@@ -11,6 +11,14 @@ defmodule Pinchflat.TestingHelperMethods do
     DateTime.add(now(), offset, :minute)
   end
 
+  def now_minus(offset, unit) when unit in [:minute, :minutes] do
+    DateTime.add(now(), -offset, :minute)
+  end
+
+  def now_minus(offset, unit) when unit in [:day, :days] do
+    DateTime.add(now(), -offset, :day)
+  end
+
   def assert_changed(checker_fun, action_fn) do
     before_res = checker_fun.()
     action_fn.()

test/test_helper.exs

@@ -1,4 +1,4 @@
-Mox.defmock(YtDlpRunnerMock, for: Pinchflat.MediaClient.Backends.BackendCommandRunner)
+Mox.defmock(YtDlpRunnerMock, for: Pinchflat.YtDlp.Backend.BackendCommandRunner)
 Application.put_env(:pinchflat, :yt_dlp_runner, YtDlpRunnerMock)
 
 Mox.defmock(HTTPClientMock, for: Pinchflat.HTTP.HTTPBehaviour)