refactor bookmarks code - separate from main rafael loop

src/rafael/__init__.py

@@ -1,69 +1,36 @@
-import time
 import dotenv
 import os
-import click
 import dokuwiki
 import requests
 
-from bs4 import BeautifulSoup
 from flask import Flask
 
 from telegram import Update
 from telegram.ext import Updater, MessageHandler, CommandHandler, Filters, CallbackContext
 
-from datetime import datetime
 
 app = Flask("rafael")
-
 RAFAEL_UA = "RAFAEL/0.1"
 
+
+from rafael.bookmarks import RafaelBookmarkPlugin
+
+
 class RafaelBot:
     
     def __init__(self):
         self.updater = Updater(os.getenv("TELEGRAM_API_KEY"))
 
-
         self.dispatcher = self.updater.dispatcher
 
-        self.dispatcher.add_handler(MessageHandler(Filters.entity('url'), self.handle_url))
+        bh = RafaelBookmarkPlugin()
+        bh.register(self.dispatcher)
 
         try:
             self.wiki = dokuwiki.DokuWiki(os.getenv('DOKU_URL'), os.getenv('DOKU_USER'), os.getenv('DOKU_PASS'), cookieAuth=True)
         except (dokuwiki.DokuWikiError, Exception) as err:
             print('unable to connect: %s' % err)
 
-
-    def get_title(self, url):
-        
-        r = requests.get(url, headers={"User-Agent": RAFAEL_UA})
-        bs = BeautifulSoup(r.content)
-        return bs.title.string
-
-    def handle_url(self, update: Update, context: CallbackContext) -> None:
-        """Echo the user message."""
-        update.message.reply_text("Storing URL...")
-        
-        update_body = ""
-
-        for ent in update.message.entities:
-            if ent.type == "url":
-                url = update.message.text[ent.offset:ent.length]
-                try:
-                    title = self.get_title(url)
-                    link = f"[[{url}|{title}]]"
-                except:
-                    link = "{url}"
-
-                added_ds = f"(retrieved at: {datetime.now().strftime('%d/%m/%Y, %H:%M:%S')})"
-
-                update_body += f"  * {link} {added_ds}"
-
-        try:
-            self.wiki.pages.append("bookmarks/inbox", f"\n\n{update_body}")
-            update.message.reply_text(f"Updated bookmarks page: {os.getenv('DOKU_URL')}bookmarks/inbox")
-        except (dokuwiki.DokuWikiError) as err:
-            update.message.reply_text(f"Failed to store url: {err}")
-
     def run(self):
         self.updater.start_polling()
 

src/rafael/bookmarks.py

@@ -0,0 +1,65 @@
+import time
+import dotenv
+import os
+import click
+import dokuwiki
+import requests
+
+from bs4 import BeautifulSoup
+from flask import Flask
+
+from telegram import Update
+import telegram
+from telegram.ext import Updater, MessageHandler, CommandHandler, Filters, CallbackContext, Dispatcher
+
+from datetime import datetime
+
+from rafael import RAFAEL_UA
+
+
+class RafaelBookmarkPlugin:
+    
+    def __init__(self):
+        self.updater = Updater(os.getenv("TELEGRAM_API_KEY"))
+
+        try:
+            self.wiki = dokuwiki.DokuWiki(os.getenv('DOKU_URL'), os.getenv('DOKU_USER'), os.getenv('DOKU_PASS'), cookieAuth=True)
+        except (dokuwiki.DokuWikiError, Exception) as err:
+            print('unable to connect: %s' % err)
+
+
+    def register(self, dispatcher: Dispatcher):
+        dispatcher.add_handler(MessageHandler(Filters.entity('url'), self.handle_url))
+
+    def get_title(self, url):
+        
+        r = requests.get(url, headers={"User-Agent": RAFAEL_UA})
+        bs = BeautifulSoup(r.content, 'html5lib')
+        return bs.title.string
+
+
+    def handle_url(self, update: Update, context: CallbackContext) -> None:
+        """Echo the user message."""
+        update.message.reply_text("Storing URL...")
+        
+        update_body = ""
+
+        for ent in update.message.entities:
+            if ent.type == "url":
+                url = update.message.text[ent.offset:ent.length]
+                try:
+                    title = self.get_title(url)
+                    link = f"[[{url}|{title}]]"
+                except Exception as err:
+                    update.message.reply_text(f"Failed to retrieve title for page {url}: {err}. Storing url only")
+                    link = f"{url}"
+
+                added_ds = f"(retrieved at: {datetime.now().strftime('%d/%m/%Y, %H:%M:%S')})"
+
+                update_body += f"  * {link} {added_ds}"
+
+        try:
+            self.wiki.pages.append("bookmarks/inbox", f"\n\n{update_body}")
+            update.message.reply_text(f"Updated bookmarks page: {os.getenv('DOKU_URL')}bookmarks/inbox")
+        except (dokuwiki.DokuWikiError) as err:
+            update.message.reply_text(f"Failed to store url: {err}")