session file logic

.gitignore

@@ -22,4 +22,5 @@ credentials.json
 secrets/*
 browsertrix/*
 browsertrix-tmp/*
-instaloader/*
+instaloader/*
+instaloader.session

archivers/instagram_archiver.py

@@ -26,8 +26,16 @@ class InstagramArchiver(Archiver):
             try:
                 self.insta.login(config.instagram_config.username, config.instagram_config.
                 password)
+                #TODO: wait for this issue to be fixed https://github.com/instaloader/instaloader/issues/1758
+                self.insta.save_session_to_file(config.instagram_config.session_file)
             except Exception as e:
-                logger.error(f"Unable to finish login: {e}\n{traceback.format_exc()}")
+                logger.error(f"Unable to finish login (retrying from file): {e}\n{traceback.format_exc()}")
+                try:
+                    self.insta.load_session_from_file(config.instagram_config.username, config.instagram_config.session_file)
+                except Exception as e2:
+                    logger.error(f"Unable to login from session file: {e2}\n{traceback.format_exc()}")
+
+
 
     def download(self, url, check_if_exists=False):
         post_matches = self.post_pattern.findall(url)

configs/config.py

@@ -185,7 +185,8 @@ class Config:
         if "instagram" in secrets:
             self.instagram_config = InstagramConfig(
                 username=secrets["instagram"]["username"],
-                password=secrets["instagram"]["password"]
+                password=secrets["instagram"]["password"],
+                session_file=secrets["instagram"].get("session_file", "instaloader.session")
             )
         else:
             self.instagram_config = None

configs/instagram_config.py

@@ -6,3 +6,4 @@ from dataclasses import dataclass
 class InstagramConfig:
     username: str
     password: str
+    session_file: str

example.config.yaml

@@ -71,6 +71,7 @@ secrets:
   instagram:
     username: "username"
     password: "password"
+    session_file: "instaloader.session" # <- default value
 
   google_sheets:
     # local filename: defaults to service_account.json, see https://gspread.readthedocs.io/en/latest/oauth2.html#for-bots-using-service-account