Introduce group basic federation, event new page and notifications

Signed-off-by: Thomas Citharel <tcit@tcit.fr>

lib/service/rich_media/parsers/fallback.ex

@@ -0,0 +1,41 @@
+# Portions of this file are derived from Pleroma:
+# Pleroma: A lightweight social networking server
+# Copyright © 2017-2020 Pleroma Authors <https://pleroma.social/>
+# SPDX-License-Identifier: AGPL-3.0-only
+
+defmodule Mobilizon.Service.RichMedia.Parsers.Fallback do
+  @moduledoc """
+  Module to parse fallback data in HTML pages (plain old title and meta description)
+  """
+  @spec parse(String.t(), map()) :: {:ok, map()} | {:error, String.t()}
+  def parse(html, data) do
+    data =
+      data
+      |> maybe_put(html, :title)
+      |> maybe_put(html, :description)
+
+    if Enum.empty?(data) do
+      {:error, "Not even a title"}
+    else
+      {:ok, data}
+    end
+  end
+
+  defp maybe_put(meta, html, attr) do
+    case get_page(html, attr) do
+      "" -> meta
+      content -> Map.put_new(meta, attr, content)
+    end
+  end
+
+  defp get_page(html, :title) do
+    html |> Floki.find("html head title") |> List.first() |> Floki.text() |> String.trim()
+  end
+
+  defp get_page(html, :description) do
+    case html |> Floki.find("html head meta[name='description']") |> List.first() do
+      nil -> ""
+      elem -> elem |> Floki.attribute("content") |> List.first() |> String.trim()
+    end
+  end
+end

lib/service/rich_media/parsers/meta_tags_parser.ex

@@ -0,0 +1,76 @@
+# Portions of this file are derived from Pleroma:
+# Pleroma: A lightweight social networking server
+# Copyright © 2017-2020 Pleroma Authors <https://pleroma.social/>
+# SPDX-License-Identifier: AGPL-3.0-only
+
+defmodule Mobilizon.Service.RichMedia.Parsers.MetaTagsParser do
+  @moduledoc """
+  Module to parse meta tags data in HTML pages
+  """
+  def parse(html, data, prefix, error_message, key_name, value_name \\ "content") do
+    meta_data =
+      html
+      |> get_elements(key_name, prefix)
+      |> Enum.reduce(data, fn el, acc ->
+        attributes = normalize_attributes(el, prefix, key_name, value_name)
+
+        Map.merge(acc, attributes)
+      end)
+      |> maybe_put_title(html)
+      |> maybe_put_description(html)
+
+    if Enum.empty?(meta_data) do
+      {:error, error_message}
+    else
+      {:ok, meta_data}
+    end
+  end
+
+  defp get_elements(html, key_name, prefix) do
+    html |> Floki.find("meta[#{key_name}^='#{prefix}:']")
+  end
+
+  defp normalize_attributes(html_node, prefix, key_name, value_name) do
+    {_tag, attributes, _children} = html_node
+
+    data =
+      Enum.into(attributes, %{}, fn {name, value} ->
+        {name, String.trim_leading(value, "#{prefix}:")}
+      end)
+
+    %{String.to_atom(data[key_name]) => data[value_name]}
+  end
+
+  defp maybe_put_title(%{title: _} = meta, _), do: meta
+
+  defp maybe_put_title(meta, html) when meta != %{} do
+    case get_page_title(html) do
+      "" -> meta
+      title -> Map.put_new(meta, :title, title)
+    end
+  end
+
+  defp maybe_put_title(meta, _), do: meta
+
+  defp maybe_put_description(%{description: _} = meta, _), do: meta
+
+  defp maybe_put_description(meta, html) when meta != %{} do
+    case get_page_description(html) do
+      "" -> meta
+      description -> Map.put_new(meta, :description, description)
+    end
+  end
+
+  defp maybe_put_description(meta, _), do: meta
+
+  defp get_page_title(html) do
+    html |> Floki.find("html head title") |> List.first() |> Floki.text()
+  end
+
+  defp get_page_description(html) do
+    case html |> Floki.find("html head meta[name='description']") |> List.first() do
+      nil -> ""
+      elem -> Floki.attribute(elem, "content")
+    end
+  end
+end

lib/service/rich_media/parsers/oembed_parser.ex

@@ -0,0 +1,83 @@
+# Portions of this file are derived from Pleroma:
+# Pleroma: A lightweight social networking server
+# Copyright © 2017-2020 Pleroma Authors <https://pleroma.social/>
+# SPDX-License-Identifier: AGPL-3.0-only
+
+defmodule Mobilizon.Service.RichMedia.Parsers.OEmbed do
+  @moduledoc """
+  Module to parse OEmbed data in HTML pages
+  """
+  alias Mobilizon.Service.Formatter.HTML
+  require Logger
+
+  @http_options [
+    follow_redirect: true,
+    ssl: [{:versions, [:"tlsv1.2"]}]
+  ]
+
+  def parse(html, _data) do
+    Logger.debug("Using OEmbed parser")
+
+    with elements = [_ | _] <- get_discovery_data(html),
+         {:ok, oembed_url} <- get_oembed_url(elements),
+         {:ok, oembed_data} <- get_oembed_data(oembed_url),
+         oembed_data <- filter_oembed_data(oembed_data) do
+      Logger.debug("Data found with OEmbed parser")
+      Logger.debug(inspect(oembed_data))
+      {:ok, oembed_data}
+    else
+      _e ->
+        {:error, "No OEmbed data found"}
+    end
+  end
+
+  defp get_discovery_data(html) do
+    html |> Floki.find("link[type='application/json+oembed']")
+  end
+
+  defp get_oembed_url(nodes) do
+    {"link", attributes, _children} = nodes |> hd()
+
+    {:ok, Enum.into(attributes, %{})["href"]}
+  end
+
+  defp get_oembed_data(url) do
+    with {:ok, %HTTPoison.Response{body: json}} <- HTTPoison.get(url, [], @http_options),
+         {:ok, data} <- Jason.decode(json),
+         data <- data |> Map.new(fn {k, v} -> {String.to_atom(k), v} end) do
+      {:ok, data}
+    end
+  end
+
+  defp filter_oembed_data(data) do
+    case Map.get(data, :type) do
+      nil ->
+        {:error, "No type declared for OEmbed data"}
+
+      "link" ->
+        Map.put(data, :image_remote_url, Map.get(data, :thumbnail_url))
+
+      "photo" ->
+        if Map.get(data, :url, "") == "" do
+          {:error, "No URL for photo OEmbed data"}
+        else
+          data
+          |> Map.put(:image_remote_url, Map.get(data, :url))
+          |> Map.put(:width, Map.get(data, :width, 0))
+          |> Map.put(:height, Map.get(data, :height, 0))
+        end
+
+      "video" ->
+        {:ok, html} = data |> Map.get(:html, "") |> HTML.filter_tags_for_oembed()
+
+        data
+        |> Map.put(:html, html)
+        |> Map.put(:width, Map.get(data, :width, 0))
+        |> Map.put(:height, Map.get(data, :height, 0))
+        |> Map.put(:image_remote_url, Map.get(data, :thumbnail_url))
+
+      "rich" ->
+        {:error, "OEmbed data has rich type, which we don't support"}
+    end
+  end
+end

lib/service/rich_media/parsers/ogp.ex

@@ -0,0 +1,36 @@
+# Portions of this file are derived from Pleroma:
+# Pleroma: A lightweight social networking server
+# Copyright © 2017-2020 Pleroma Authors <https://pleroma.social/>
+# SPDX-License-Identifier: AGPL-3.0-only
+
+defmodule Mobilizon.Service.RichMedia.Parsers.OGP do
+  @moduledoc """
+  Module to parse OpenGraph data in HTML pages
+  """
+  require Logger
+  alias Mobilizon.Service.RichMedia.Parsers.MetaTagsParser
+
+  def parse(html, data) do
+    Logger.debug("Using OpenGraph card parser")
+
+    with {:ok, data} <-
+           MetaTagsParser.parse(
+             html,
+             data,
+             "og",
+             "No OGP metadata found",
+             "property"
+           ) do
+      data = transform_tags(data)
+      Logger.debug("Data found with OpenGraph card parser")
+      {:ok, data}
+    end
+  end
+
+  defp transform_tags(data) do
+    data
+    |> Map.put(:image_remote_url, Map.get(data, :image))
+    |> Map.put(:width, Map.get(data, :"image:width"))
+    |> Map.put(:height, Map.get(data, :"image:height"))
+  end
+end

lib/service/rich_media/parsers/twitter_card.ex

@@ -0,0 +1,34 @@
+# Portions of this file are derived from Pleroma:
+# Pleroma: A lightweight social networking server
+# Copyright © 2017-2020 Pleroma Authors <https://pleroma.social/>
+# SPDX-License-Identifier: AGPL-3.0-only
+
+defmodule Mobilizon.Service.RichMedia.Parsers.TwitterCard do
+  @moduledoc """
+  Module to parse Twitter tags data in HTML pages
+  """
+  alias Mobilizon.Service.RichMedia.Parsers.MetaTagsParser
+  require Logger
+
+  @spec parse(String.t(), map()) :: {:ok, map()} | {:error, String.t()}
+  def parse(html, data) do
+    Logger.debug("Using Twitter card parser")
+
+    res =
+      data
+      |> parse_name_attrs(html)
+      |> parse_property_attrs(html)
+
+    Logger.debug("Data found with Twitter card parser")
+    Logger.debug(inspect(res))
+    res
+  end
+
+  defp parse_name_attrs(data, html) do
+    MetaTagsParser.parse(html, data, "twitter", %{}, "name")
+  end
+
+  defp parse_property_attrs({_, data}, html) do
+    MetaTagsParser.parse(html, data, "twitter", "No twitter card metadata found", "property")
+  end
+end