Merge branch 'main' into feat/yt-dlp-pots

src/auto_archiver/core/module.py

@@ -214,7 +214,7 @@ class LazyBaseModule:
 
         # check external dependencies are installed
         def check_deps(deps, check):
-            for dep in filter(lambda d: len(d.strip()), deps):
+            for dep in filter(lambda d: len(d.strip()) > 0, deps):
                 if not check(dep.strip()):
                     logger.error(
                         f"Module '{self.name}' requires external dependency '{dep}' which is not available/setup. \
@@ -274,6 +274,9 @@ class LazyBaseModule:
         # finally, get the class instance
         instance: BaseModule = getattr(sys.modules[sub_qualname], class_name)()
 
+        # save the instance for future easy loading
+        self._instance = instance
+
         # set the name, display name and module factory
         instance.name = self.name
         instance.display_name = self.display_name
@@ -286,8 +289,6 @@ class LazyBaseModule:
         instance.config_setup(config)
         instance.setup()
 
-        # save the instance for future easy loading
-        self._instance = instance
         return instance
 
     def __repr__(self):

src/auto_archiver/core/orchestrator.py

@@ -387,8 +387,10 @@ Here's how that would look: \n\nsteps:\n  extractors:\n  - [your_extractor_name_
                 except (KeyboardInterrupt, Exception) as e:
                     if not isinstance(e, KeyboardInterrupt) and not isinstance(e, SetupError):
                         logger.error(f"Error during setup of modules: {e}\n{traceback.format_exc()}")
-                    if loaded_module and module_type == "extractor":
-                        loaded_module.cleanup()
+
+                    # access the _instance here because loaded_module may not return if there's an error
+                    if lazy_module._instance and module_type == "extractor":
+                        lazy_module._instance.cleanup()
                     raise e
 
                 if not loaded_module:

src/auto_archiver/core/validators.py

@@ -4,12 +4,6 @@ import argparse
 import json
 
 
-def example_validator(value):
-    if "example" not in value:
-        raise argparse.ArgumentTypeError(f"{value} is not a valid value for this argument")
-    return value
-
-
 def positive_number(value):
     if value < 0:
         raise argparse.ArgumentTypeError(f"{value} is not a positive number")

src/auto_archiver/modules/telethon_extractor/__manifest__.py

@@ -19,7 +19,7 @@
         },
         "session_file": {
             "default": "secrets/anon",
-            "help": "optional, records the telegram login session for future usage, '.session' will be appended to the provided value.",
+            "help": "Path of the file to save the telegram login session for future usage, '.session' will be appended to the provided path.",
         },
         "join_channels": {
             "default": True,

src/auto_archiver/modules/telethon_extractor/telethon_extractor.py

@@ -1,4 +1,10 @@
+import os
 import shutil
+import re
+import time
+from pathlib import Path
+from datetime import date
+
 from telethon.sync import TelegramClient
 from telethon.errors import ChannelInvalidError
 from telethon.tl.functions.messages import ImportChatInviteRequest
@@ -8,11 +14,9 @@ from telethon.errors.rpcerrorlist import (
     InviteRequestSentError,
     InviteHashExpiredError,
 )
-from loguru import logger
+
 from tqdm import tqdm
-import re
-import time
-import os
+from loguru import logger
 
 from auto_archiver.core import Extractor
 from auto_archiver.core import Metadata, Media
@@ -31,10 +35,22 @@ class TelethonExtractor(Extractor):
         """
         logger.info(f"SETUP {self.name} checking login...")
 
+        # in case the user already added '.session' to the session_file
+        base_session_name = self.session_file.removesuffix(".session")
+        base_session_filepath = f"{base_session_name}.session"
+
+        if self.session_file and not os.path.exists(base_session_filepath):
+            logger.warning(
+                f"SETUP - Session file {base_session_filepath} does not exist for {self.name}, creating an empty one."
+            )
+            Path(base_session_filepath).touch()
+
         # make a copy of the session that is used exclusively with this archiver instance
-        new_session_file = os.path.join("secrets/", f"telethon-{time.strftime('%Y-%m-%d')}{random_str(8)}.session")
-        shutil.copy(self.session_file + ".session", new_session_file)
-        self.session_file = new_session_file.replace(".session", "")
+        self.session_file = os.path.join(
+            os.path.dirname(base_session_filepath), f"telethon-{date.today().strftime('%Y-%m-%d')}{random_str(8)}"
+        )
+        logger.debug(f"Making a copy of the session file {base_session_filepath} to {self.session_file}.session")
+        shutil.copy(base_session_filepath, f"{self.session_file}.session")
 
         # initiate the client
         self.client = TelegramClient(self.session_file, self.api_id, self.api_hash)
@@ -87,8 +103,8 @@ class TelethonExtractor(Extractor):
                     pbar.update()
 
     def cleanup(self) -> None:
-        logger.info(f"CLEANUP {self.name}.")
-        session_file_name = self.session_file + ".session"
+        logger.info(f"CLEANUP {self.name} - removing session file {self.session_file}.session")
+        session_file_name = f"{self.session_file}.session"
         if os.path.exists(session_file_name):
             os.remove(session_file_name)