From a08510f21efa7ec3a32082f70d4420b64b73c156 Mon Sep 17 00:00:00 2001
From: Yay295 <Yay295@users.noreply.github.com>
Date: Mon, 28 Oct 2024 15:13:42 -0500
Subject: [PATCH] Add Bluesky scraper (#290)

* add bluesky scraper

* use com.atproto.sync.getBlob to get original bluesky image

* ignore data after bluesky post id

* fix json access and missing function

* fix bluesky fullsize image regex
---
 lib/philomena_proxy/scrapers.ex         |  1 +
 lib/philomena_proxy/scrapers/bluesky.ex | 48 +++++++++++++++++++++++++
 2 files changed, 49 insertions(+)
 create mode 100644 lib/philomena_proxy/scrapers/bluesky.ex

diff --git a/lib/philomena_proxy/scrapers.ex b/lib/philomena_proxy/scrapers.ex
index a96f0817..08674d44 100644
--- a/lib/philomena_proxy/scrapers.ex
+++ b/lib/philomena_proxy/scrapers.ex
@@ -21,6 +21,7 @@ defmodule PhilomenaProxy.Scrapers do
         }
 
   @scrapers [
+    PhilomenaProxy.Scrapers.Bluesky,
     PhilomenaProxy.Scrapers.Deviantart,
     PhilomenaProxy.Scrapers.Pillowfort,
     PhilomenaProxy.Scrapers.Twitter,
diff --git a/lib/philomena_proxy/scrapers/bluesky.ex b/lib/philomena_proxy/scrapers/bluesky.ex
new file mode 100644
index 00000000..598d1470
--- /dev/null
+++ b/lib/philomena_proxy/scrapers/bluesky.ex
@@ -0,0 +1,48 @@
+defmodule PhilomenaProxy.Scrapers.Bluesky do
+  @moduledoc false
+
+  alias PhilomenaProxy.Scrapers.Scraper
+  alias PhilomenaProxy.Scrapers
+
+  @behaviour Scraper
+
+  @url_regex ~r|https://bsky\.app/profile/([^/]+)/post/([^/?#]+)|
+  @fullsize_image_regex ~r|.*/img/feed_fullsize/plain/([^/]+)/([^@]+).*|
+  @blob_image_url_pattern "https://bsky.social/xrpc/com.atproto.sync.getBlob/?did=\\1&cid=\\2"
+
+  @spec can_handle?(URI.t(), String.t()) :: boolean()
+  def can_handle?(_uri, url) do
+    String.match?(url, @url_regex)
+  end
+
+  @spec scrape(URI.t(), Scrapers.url()) :: Scrapers.scrape_result()
+  def scrape(_uri, url) do
+    [handle, id] = Regex.run(@url_regex, url, capture: :all_but_first)
+
+    api_url_resolve_handle =
+      "https://public.api.bsky.app/xrpc/com.atproto.identity.resolveHandle?handle=#{handle}"
+
+    did = PhilomenaProxy.Http.get(api_url_resolve_handle) |> json!() |> Map.fetch!(:did)
+
+    api_url_get_posts =
+      "https://public.api.bsky.app/xrpc/app.bsky.feed.getPosts?uris=at://#{did}/app.bsky.feed.post/#{id}"
+
+    post_json = PhilomenaProxy.Http.get(api_url_get_posts) |> json!() |> Map.fetch!(:posts) |> hd
+
+    %{
+      source_url: url,
+      author_name: post_json["author"]["handle"],
+      description: post_json["record"]["text"],
+      images:
+        post_json["embed"]["images"]
+        |> Enum.map(
+          &%{
+            url: String.replace(&1["fullsize"], @fullsize_image_regex, @blob_image_url_pattern),
+            camo_url: PhilomenaProxy.Camo.image_url(&1["thumb"])
+          }
+        )
+    }
+  end
+
+  defp json!({:ok, %{body: body, status: 200}}), do: Jason.decode!(body)
+end