refactor(cite-task): support cite artilce in side comment (#399)

* refactor(cite-task): support cite artilce in side comment

* refactor(cite-task): support cross comment citing

* refactor(cite-task): add some inline doc

* refactor(cite-task): fmt naming

* refactor(cite-task): fix citing self edge-case

* refactor(cite-task): fmt

Author: mydearxym

lib/groupher_server/cms/delegates/article_curd.ex

@@ -164,10 +164,10 @@ defmodule GroupherServer.CMS.Delegate.ArticleCURD do
       |> Multi.run(:update_user_published_meta, fn _, _ ->
         Accounts.update_published_states(uid, thread)
       end)
-      |> Multi.run(:block_tasks, fn _, %{create_article: article} ->
+      |> Multi.run(:after_tasks, fn _, %{create_article: article} ->
         Later.run({CiteTasks, :handle, [article]})
+        Later.run({__MODULE__, :notify_admin_new_article, [article]})
       end)
-      # TODO: run mini tasks
       |> Multi.run(:mention_users, fn _, %{create_article: article} ->
         # article.body |> Jason.decode!() |> 各种小 task
         Delivery.mention_from_content(community.raw, thread, article, attrs, %User{id: uid})
@@ -223,6 +223,9 @@ defmodule GroupherServer.CMS.Delegate.ArticleCURD do
     |> Multi.run(:update_edit_status, fn _, %{update_article: update_article} ->
       ArticleCommunity.update_edit_status(update_article)
     end)
+    |> Multi.run(:after_tasks, fn _, %{update_article: update_article} ->
+      Later.run({CiteTasks, :handle, [update_article]})
+    end)
     |> Repo.transaction()
     |> result()
   end
@@ -449,7 +452,6 @@ defmodule GroupherServer.CMS.Delegate.ArticleCURD do
 
   # create done
   defp result({:ok, %{set_active_at_timestamp: result}}) do
-    Later.run({__MODULE__, :notify_admin_new_article, [result]})
     {:ok, result}
   end
 

lib/groupher_server/cms/delegates/cite_tasks.ex

@@ -29,7 +29,7 @@ defmodule GroupherServer.CMS.Delegate.CiteTasks do
   import Helper.ErrorCode
 
   alias GroupherServer.{CMS, Repo}
-  alias CMS.Model.CitedContent
+  alias CMS.Model.{CitedContent, Comment}
   alias Helper.ORM
 
   alias Ecto.Multi
@@ -38,16 +38,16 @@ defmodule GroupherServer.CMS.Delegate.CiteTasks do
   @article_threads get_config(:article, :threads)
   @valid_article_prefix Enum.map(@article_threads, &"#{@site_host}/#{&1}/")
 
-  def handle(%{body: body} = article) do
+  def handle(%{body: body} = content) do
     with {:ok, %{"blocks" => blocks}} <- Jason.decode(body),
-         article <- Repo.preload(article, author: :user) do
+         content <- preload_content_author(content) do
       Multi.new()
       |> Multi.run(:delete_all_cited_contents, fn _, _ ->
-        delete_all_cited_contents(article)
+        delete_all_cited_contents(content)
       end)
       |> Multi.run(:update_cited_info, fn _, _ ->
         blocks
-        |> Enum.reduce([], &(&2 ++ parse_cited_info_per_block(article, &1)))
+        |> Enum.reduce([], &(&2 ++ parse_cited_info_per_block(content, &1)))
         |> merge_same_cited_article_block
         |> update_cited_info
       end)
@@ -56,9 +56,17 @@ defmodule GroupherServer.CMS.Delegate.CiteTasks do
     end
   end
 
+  def preload_content_author(%Comment{} = comment), do: comment
+  def preload_content_author(article), do: Repo.preload(article, author: :user)
+
   # delete all records before insert_all, this will dynamiclly update
   # those cited info when update article
   # 插入引用记录之前先全部清除，这样可以在更新文章的时候自动计算引用信息
+  defp delete_all_cited_contents(%Comment{} = comment) do
+    query = from(c in CitedContent, where: c.comment_id == ^comment.id)
+    ORM.delete_all(query, :if_exist)
+  end
+
   defp delete_all_cited_contents(article) do
     with {:ok, thread} <- thread_of_article(article),
          {:ok, info} <- match(thread) do
@@ -68,15 +76,18 @@ defmodule GroupherServer.CMS.Delegate.CiteTasks do
     end
   end
 
-  # defp batch_done
-
+  # batch insert CitedContent record and update citing count
   defp update_cited_info(cited_contents) do
-    clean_cited_contents = Enum.map(cited_contents, &Map.delete(&1, :cited_article))
-    # IO.inspect(clean_cited_contents, label: "clean_cited_contents")
-    with true <- {0, nil} !== Repo.insert_all(CitedContent, clean_cited_contents) do
-      update_citing_count(cited_contents)
-    else
-      _ -> {:error, "insert cited content error"}
+    # see: https://github.com/elixir-ecto/ecto/issues/1932#issuecomment-314083252
+    clean_cited_contents =
+      cited_contents
+      |> Enum.map(&(&1 |> Map.merge(%{inserted_at: &1.citing_time, updated_at: &1.citing_time})))
+      |> Enum.map(&Map.delete(&1, :cited_content))
+      |> Enum.map(&Map.delete(&1, :citing_time))
+
+    case {0, nil} !== Repo.insert_all(CitedContent, clean_cited_contents) do
+      true -> update_citing_count(cited_contents)
+      false -> {:error, "insert cited content error"}
     end
   end
 
@@ -85,10 +96,10 @@ defmodule GroupherServer.CMS.Delegate.CiteTasks do
       count_query = from(c in CitedContent, where: c.cited_by_id == ^content.cited_by_id)
       count = Repo.aggregate(count_query, :count)
 
-      cited_article = content.cited_article
-      meta = Map.merge(cited_article.meta, %{citing_count: count})
+      cited_content = content.cited_content
+      meta = Map.merge(cited_content.meta, %{citing_count: count})
 
-      case cited_article |> ORM.update_meta(meta) do
+      case cited_content |> ORM.update_meta(meta) do
         {:ok, _} -> true
         {:error, _} -> false
       end
@@ -140,47 +151,56 @@ defmodule GroupherServer.CMS.Delegate.CiteTasks do
       block_linker: ["block-ZgKJs"],
       cited_by_id: 190057,
       cited_by_type: "POST",
-      cited_article: #loaded,
+      cited_content: #loaded,
       post_id: 190059,
       user_id: 1413053
     }
     ...
   ]
   """
-  defp parse_cited_info_per_block(article, %{"id" => block_id, "data" => %{"text" => text}}) do
-    links_in_block = Floki.find(text, "a[href]")
-
-    Enum.reduce(links_in_block, [], fn link, acc ->
-      with {:ok, cited_article} <- parse_cited_article(link),
-           # do not cite artilce itself
-           true <- article.id !== cited_article.id do
-        List.insert_at(acc, 0, shape_cited_content(article, cited_article, block_id))
-      else
+  defp parse_cited_info_per_block(content, %{"id" => block_id, "data" => %{"text" => text}}) do
+    links = Floki.find(text, "a[href]")
+
+    do_parse_cited_info_per_block(content, block_id, links)
+  end
+
+  # links Floki parsed fmt
+  # content means both article and comment
+  # e.g:
+  # [{"a", [{"href", "https://coderplanets.com/post/195675"}], []},]
+  defp do_parse_cited_info_per_block(content, block_id, links) do
+    Enum.reduce(links, [], fn link, acc ->
+      case parse_valid_cited(content.id, link) do
+        {:ok, cited} -> List.insert_at(acc, 0, shape_cited(content, cited, block_id))
         _ -> acc
       end
     end)
     |> Enum.uniq()
   end
 
-  defp shape_cited_content(article, cited_article, block_id) do
-    {:ok, thread} = thread_of_article(article)
-    {:ok, info} = match(thread)
-
-    %{
-      cited_by_id: cited_article.id,
-      cited_by_type: cited_article.meta.thread,
-      # used for updating citing_count, avoid load again
-      cited_article: cited_article,
-      block_linker: [block_id],
-      user_id: article.author.user.id
-    }
-    |> Map.put(info.foreign_key, article.id)
+  # parse cited with check if citing link is point to itself
+  defp parse_valid_cited(content_id, link) do
+    with {:ok, cited} <- parse_cited(link),
+         %{content: content} <- cited do
+      case content.id !== content_id do
+        true -> {:ok, cited}
+        false -> {:error, "citing itself"}
+      end
+    end
   end
 
-  defp parse_cited_article({"a", attrs, _}) do
+  # return fmt: %{type: :comment | :article, content: %Comment{} | Article}
+  # 要考虑是否有 comment_id 的情况，如果有，那么 就应该 load comment 而不是 article
+  defp parse_cited({"a", attrs, _}) do
     with {:ok, link} <- parse_link(attrs),
          true <- is_site_article_link?(link) do
-      load_cited_article_from_url(link)
+      # IO.inspect(link, label: "parse link")
+      # IO.inspect(is_comment_link?(link), label: "is_comment_link")
+
+      case is_comment_link?(link) do
+        true -> load_cited_comment_from_url(link)
+        false -> load_cited_article_from_url(link)
+      end
     end
   end
 
@@ -204,6 +224,26 @@ defmodule GroupherServer.CMS.Delegate.CiteTasks do
     Enum.any?(@valid_article_prefix, &String.starts_with?(url, &1))
   end
 
+  defp is_comment_link?(url) do
+    with %{query: query} <- URI.parse(url) do
+      not is_nil(query) and String.starts_with?(query, "comment_id=")
+    end
+  end
+
+  defp load_cited_comment_from_url(url) do
+    %{query: query} = URI.parse(url)
+
+    try do
+      comment_id = URI.decode_query(query) |> Map.get("comment_id")
+
+      with {:ok, comment} <- ORM.find(Comment, comment_id) do
+        {:ok, %{type: :comment, content: comment}}
+      end
+    rescue
+      _ -> {:error, "load comment error"}
+    end
+  end
+
   # get cited article from url
   # e.g: https://coderplanets.com/post/189993 -> ORM.find(Post, 189993)
   defp load_cited_article_from_url(url) do
@@ -212,11 +252,86 @@ defmodule GroupherServer.CMS.Delegate.CiteTasks do
     thread = path_list |> Enum.at(1) |> String.downcase() |> String.to_atom()
     article_id = path_list |> Enum.at(2)
 
-    with {:ok, info} <- match(thread) do
-      ORM.find(info.model, article_id)
+    with {:ok, info} <- match(thread),
+         {:ok, article} <- ORM.find(info.model, article_id) do
+      {:ok, %{type: :article, content: article}}
     end
   end
 
+  # cite article in comment
+  # 在评论中引用文章
+  defp shape_cited(%Comment{} = comment, %{type: :article, content: cited}, block_id) do
+    %{
+      cited_by_id: cited.id,
+      cited_by_type: cited.meta.thread,
+      comment_id: comment.id,
+      block_linker: [block_id],
+      user_id: comment.author_id,
+      # extra fields for next-step usage
+      # used for updating citing_count, avoid load again
+      cited_content: cited,
+      # for later insert all
+      citing_time: comment.updated_at |> DateTime.truncate(:second)
+    }
+  end
+
+  # cite comment in comment
+  # 评论中引用评论
+  defp shape_cited(%Comment{} = comment, %{type: :comment, content: cited}, block_id) do
+    %{
+      cited_by_id: cited.id,
+      cited_by_type: "COMMENT",
+      comment_id: comment.id,
+      block_linker: [block_id],
+      user_id: comment.author_id,
+      # extra fields for next-step usage
+      # used for updating citing_count, avoid load again
+      cited_content: cited,
+      # for later insert all
+      citing_time: comment.updated_at |> DateTime.truncate(:second)
+    }
+  end
+
+  # cite article in article
+  # 文章之间相互引用
+  defp shape_cited(article, %{type: :article, content: cited}, block_id) do
+    {:ok, thread} = thread_of_article(article)
+    {:ok, info} = match(thread)
+
+    %{
+      cited_by_id: cited.id,
+      cited_by_type: cited.meta.thread,
+      block_linker: [block_id],
+      user_id: article.author.user.id,
+      # extra fields for next-step usage
+      # used for updating citing_count, avoid load again
+      cited_content: cited,
+      # for later insert all
+      citing_time: article.updated_at |> DateTime.truncate(:second)
+    }
+    |> Map.put(info.foreign_key, article.id)
+  end
+
+  # cite comment in article
+  # 文章中引用评论
+  defp shape_cited(article, %{type: :comment, content: cited}, block_id) do
+    {:ok, thread} = thread_of_article(article)
+    {:ok, info} = match(thread)
+
+    %{
+      cited_by_id: cited.id,
+      cited_by_type: "COMMENT",
+      block_linker: [block_id],
+      user_id: article.author.user.id,
+      # extra fields for next-step usage
+      # used for updating citing_count, avoid load again
+      cited_content: cited,
+      # for later insert all
+      citing_time: article.updated_at |> DateTime.truncate(:second)
+    }
+    |> Map.put(info.foreign_key, article.id)
+  end
+
   defp result({:ok, %{update_cited_info: result}}), do: {:ok, result}
 
   defp result({:error, :update_cited_info, _result, _steps}) do

lib/groupher_server/cms/delegates/comment_curd.ex

@@ -11,7 +11,7 @@ defmodule GroupherServer.CMS.Delegate.CommentCurd do
   import ShortMaps
 
   alias Helper.Types, as: T
-  alias Helper.{ORM, QueryBuilder, Converter}
+  alias Helper.{Later, ORM, QueryBuilder, Converter}
   alias GroupherServer.{Accounts, CMS, Repo}
   alias CMS.Model.Post
 
@@ -113,6 +113,9 @@ defmodule GroupherServer.CMS.Delegate.CommentCurd do
           false -> CMS.update_active_timestamp(thread, article)
         end
       end)
+      |> Multi.run(:block_tasks, fn _, %{create_comment: create_comment} ->
+        Later.run({CiteTasks, :handle, [create_comment]})
+      end)
       |> Repo.transaction()
       |> result()
     else

lib/groupher_server/cms/models/cited_content.ex

@@ -13,7 +13,7 @@ defmodule GroupherServer.CMS.Model.CitedContent do
 
   alias CMS.Model.Comment
 
-  @timestamps_opts [type: :utc_datetime_usec]
+  @timestamps_opts [type: :utc_datetime]
 
   @required_fields ~w(cited_by_type cited_by_id user_id)a
   @article_cast_fields general_article_fields(:cast)

lib/groupher_server/cms/models/embeds/comment_meta.ex

@@ -7,7 +7,7 @@ defmodule GroupherServer.CMS.Model.Embeds.CommentMeta do
 
   import Ecto.Changeset
 
-  @optional_fields ~w(is_article_author_upvoted report_count is_reply_to_others reported_count reported_user_ids)a
+  @optional_fields ~w(is_article_author_upvoted report_count is_reply_to_others reported_count reported_user_ids citing_count)a
 
   @doc "for test usage"
   def default_meta() do
@@ -17,7 +17,8 @@ defmodule GroupherServer.CMS.Model.Embeds.CommentMeta do
       report_count: 0,
       upvoted_user_ids: [],
       reported_user_ids: [],
-      reported_count: 0
+      reported_count: 0,
+      citing_count: 0
     }
   end
 
@@ -31,6 +32,7 @@ defmodule GroupherServer.CMS.Model.Embeds.CommentMeta do
     field(:upvoted_user_ids, {:array, :integer}, default: [])
     field(:reported_user_ids, {:array, :integer}, default: [])
     field(:reported_count, :integer, default: 0)
+    field(:citing_count, :integer, default: 0)
   end
 
   def changeset(struct, params) do

priv/repo/migrations/20210614145939_missing_timestamp_for_cite_contents.exs

@@ -0,0 +1,9 @@
+defmodule GroupherServer.Repo.Migrations.MissingTimestampForCiteContents do
+  use Ecto.Migration
+
+  def change do
+    alter table(:cited_contents) do
+      timestamps()
+    end
+  end
+end