From c247774256523d75937c94aa2dade84cee5b29e5 Mon Sep 17 00:00:00 2001
From: Travis Shears <t@travisshears.com>
Date: Fri, 18 Jul 2025 12:40:41 +0200
Subject: [PATCH] init blue sky scraping

---
 config.sample.edn              |  9 +++-
 src/micro_blog/blue_sky.clj    | 83 ++++++++++++++++++++++++++++++++++
 src/micro_blog/pocket_base.clj | 71 +++++++++++++++++++++++++++++
 src/micro_blog/utils.clj       |  9 ++++
 4 files changed, 171 insertions(+), 1 deletion(-)
 create mode 100644 src/micro_blog/blue_sky.clj
 create mode 100644 src/micro_blog/pocket_base.clj
 create mode 100644 src/micro_blog/utils.clj

diff --git a/config.sample.edn b/config.sample.edn
index e4025ef..4aa25de 100644
--- a/config.sample.edn
+++ b/config.sample.edn
@@ -1,3 +1,10 @@
 {:mistral-api-key "xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx"
  :mistral-agent-id "ag:xxxxxxxx:xxxxxxxx:xxxxxxxxxxxxxxxxxxx:xxxxxxxx"
- :mistral-host "https://api.mistral.ai"}
+ :mistral-host "https://api.mistral.ai"
+
+ :pocket-base-pw "xxxxxxxx"
+ :pocket-base-user "xxxxxxxxxxxxxxxxxxx"
+ :pocket-base-host "xxxxxxxxxxxxx"
+
+ :blue-sky-api-key "xxxxxxxxxxxxxxxxxxx"
+ :blue-sky-username "coolguy.bsky.social"}
diff --git a/src/micro_blog/blue_sky.clj b/src/micro_blog/blue_sky.clj
new file mode 100644
index 0000000..6f7bd8f
--- /dev/null
+++ b/src/micro_blog/blue_sky.clj
@@ -0,0 +1,83 @@
+(ns micro-blog.blue-sky
+  (:require
+   [clj-http.client :as http-client]
+   [micro-blog.pocket-base :as pb]
+   [micro-blog.utils :as utils]
+   [malli.core :as m]
+   [clojure.pprint :refer [pprint]]
+   [micro-blog.config :refer [config]]))
+
+(defn create-session []
+  (let [identifier (@config :blue-sky-username)
+        api-key (@config :blue-sky-api-key)
+        body {:identifier identifier :password api-key}
+        url (str (@config :blue-sky-host) "/com.atproto.server.createSession")
+        res-schema [:map
+                    [:did string?]
+                    [:accessJwt string?]]]
+    (-> (http-client/post url
+                          {:form-params body
+                           :content-type :json
+                           :as :json})
+        :body
+        (utils/validate-with-throw res-schema)
+        (#(assoc % :access-jwt (:accessJwt %)))
+        (select-keys  [:did :access-jwt]))))
+
+(def post-res-schema [:map
+                      [:cursor [:maybe :string]]
+                      [:feed [:vector
+                              [:map [:post [:map
+                                            [:cid :string]
+                                            [:author [:map
+                                                      [:handle :string]]]
+
+                                            [:record [:map
+                                                      [:createdAt :string]]]]]]]]])
+
+(defn get-posts-until-id
+  ([session id] (get-posts-until-id session id nil []))
+  ([session id cursor prev-posts]
+   (let [limit 5
+         body
+         (-> (http-client/get (str (@config :blue-sky-host) "/app.bsky.feed.getAuthorFeed")
+                              {:headers {"Authorization" (str "Bearer " (session :access-jwt))}
+                               :query-params  (cond-> {:actor (:did session)
+                                                       :limit limit}
+                                                cursor (assoc :cursor cursor))
+                               :content-type :json
+                               :as :json})
+             :body
+             (utils/validate-with-throw post-res-schema))
+         posts (map :post (:feed body))
+         new-cursor (:cursor body)
+         new-posts (take-while #(not= (:cid %) id) posts)
+         new-and-prev-posts (concat new-posts prev-posts)]
+     (cond
+       ;; end of posts
+       (not= (count posts) limit) new-and-prev-posts
+       ;; found post
+       (some #(= id (:cid %)) posts) new-and-prev-posts
+       ;; recur
+       :else (recur session id new-cursor new-and-prev-posts)))))
+
+;; TODO: create post
+;; const data = {
+;;     "remoteId": "test",
+;;     "authorId": "test",
+;;     "posted": "2022-01-01 10:00:00.123Z",
+;;     "source": "pleroma",
+;;     "tags": [
+;;         "RELATION_RECORD_ID"
+;;     ],
+;;     "fullPost": "JSON",
+;;     "images": [
+;;         "RELATION_RECORD_ID"
+;;     ]
+;; };
+
+(defn run []
+  (let [session (create-session)
+        last-saved-id (pb/get-latest-post-remote-id-by-source :blue_sky)
+        new-posts (get-posts-until-id session last-saved-id)]
+    {:session session :last-saved-id last-saved-id :new-posts new-posts}))
diff --git a/src/micro_blog/pocket_base.clj b/src/micro_blog/pocket_base.clj
new file mode 100644
index 0000000..5a027ae
--- /dev/null
+++ b/src/micro_blog/pocket_base.clj
@@ -0,0 +1,71 @@
+(ns micro-blog.pocket-base
+  (:require
+   [clojure.pprint :refer [pprint]]
+   [clojure.string :as str]
+   [clj-http.client :as http-client]
+   [malli.core :as m]
+   [micro-blog.config :refer [config]]))
+
+(defonce token-cache ^:private (atom {:token nil :fetched-at nil}))
+(defn- now []
+  (java.time.Instant/now))
+
+(defn older-then-a-day? [fetched-at]
+  (when fetched-at
+    (let [duration (java.time.Duration/between fetched-at (now))]
+      (> (.toHours duration) 23)))) ; 23 to be safe, or use 24
+
+(defn get-login-token []
+  (let [user-name (@config :pocket-base-user)
+        pw (@config :pocket-base-pw)
+        body {:identity user-name :password pw}
+        url (str (@config :pocket-base-host) "/api/collections/users/auth-with-password")]
+    (->
+     (http-client/post url
+                       {:form-params body
+                        :content-type :json
+                        :as :json})
+     :body
+     :token)))
+
+(defn get-login-token-with-cache []
+  (let [{:keys [token fetched-at]} @token-cache]
+    (if (and token (not (older-then-a-day? fetched-at)))
+      token
+      (let [new-token (get-login-token)]
+        (println "Getting new login token")
+        (reset! token-cache {:token new-token :fetched-at (now)})
+        new-token))))
+
+(def sources #{:pleroma :blue_sky :mastodon :pixelfed :nostr})
+(defn valid-source? [source]
+  (contains? sources source))
+
+(defn get-latest-post-remote-id-by-source [source]
+  (let [res-schema
+        [:map
+         [:items
+          [:vector
+           [:map
+            [:id string?]
+            [:remoteId string?]]]]]]
+    (when (not (valid-source? source))
+      (throw (ex-info "Invalid source" {:source source})))
+    (as->
+     (http-client/get  (str (@config :pocket-base-host) "/api/collections/micro_blog_posts/records")
+                       {:headers {"Authorization" (get-login-token-with-cache)}
+                        :query-params  {:page 1
+                                        "perPage" 1
+                                        :sort "-posted"
+                                        :filter (str "source = '" (name source) "'")
+                                        :fields (str/join "," ["remoteId" "id"])
+                                        "skipTotal" true}
+                        :content-type :json
+                        :as :json}) x
+      (:body x)
+      (if (m/validate res-schema x)
+        x
+        (do
+          (m/explain res-schema x)
+          (throw (ex-info "Res does not follow schema" {:res x}))))
+      (-> x :items first :remoteId))))
diff --git a/src/micro_blog/utils.clj b/src/micro_blog/utils.clj
new file mode 100644
index 0000000..2d68e46
--- /dev/null
+++ b/src/micro_blog/utils.clj
@@ -0,0 +1,9 @@
+(ns micro-blog.utils
+  (:require [malli.core :as m]))
+
+(defn validate-with-throw [value schema]
+  (if (m/validate schema value)
+    value
+    (do
+      (m/explain schema value)
+      (throw (ex-info "Res does not follow schema" {:value value})))))