Merge pull request #210 from bellingcat/logger_fix

Fix issue #200 + Refactor _LAZY_LOADED_MODULES

Author: pjrobertson

docs/scripts/scripts.py

@@ -1,6 +1,6 @@
 # iterate through all the modules in auto_archiver.modules and turn the __manifest__.py file into a markdown table
 from pathlib import Path
-from auto_archiver.core.module import available_modules
+from auto_archiver.core.module import ModuleFactory
 from auto_archiver.core.base_module import BaseModule
 from ruamel.yaml import YAML
 import io
@@ -41,7 +41,7 @@ def generate_module_docs():
     configs_cheatsheet = "\n## Configuration Options\n"
     configs_cheatsheet += header_row
 
-    for module in sorted(available_modules(with_manifest=True), key=lambda x: (x.requires_setup, x.name)):
+    for module in sorted(ModuleFactory().available_modules(), key=lambda x: (x.requires_setup, x.name)):
         # generate the markdown file from the __manifest__.py file.
 
         manifest = module.manifest

pyproject.toml

@@ -4,7 +4,7 @@ build-backend = "poetry.core.masonry.api"
 
 [project]
 name = "auto-archiver"
-version = "0.13.3"
+version = "0.13.4"
 description = "Automatically archive links to videos, images, and social media content from Google Sheets (and more)."
 
 requires-python = ">=3.10,<3.13"

src/auto_archiver/core/__init__.py

@@ -3,7 +3,7 @@
 """
 from .metadata import Metadata
 from .media import Media
-from .module import BaseModule
+from .base_module import BaseModule
 
 # cannot import ArchivingOrchestrator/Config to avoid circular dep
 # from .orchestrator import ArchivingOrchestrator

src/auto_archiver/core/base_module.py

@@ -1,13 +1,18 @@
 
-from urllib.parse import urlparse
-from typing import  Mapping, Any
+from __future__ import annotations
+
+from typing import  Mapping, Any, Type, TYPE_CHECKING
 from abc import ABC
 from copy import deepcopy, copy
 from tempfile import TemporaryDirectory
 from auto_archiver.utils import url as UrlUtil
+from auto_archiver.core.consts import MODULE_TYPES as CONF_MODULE_TYPES
 
 from loguru import logger
 
+if TYPE_CHECKING:
+    from .module import ModuleFactory
+
 class BaseModule(ABC):
 
     """
@@ -17,41 +22,24 @@ class BaseModule(ABC):
     however modules can have a .setup() method to run any setup code
     (e.g. logging in to a site, spinning up a browser etc.)
 
-    See BaseModule.MODULE_TYPES for the types of modules you can create, noting that
+    See consts.MODULE_TYPES for the types of modules you can create, noting that
     a subclass can be of multiple types. For example, a module that extracts data from
     a website and stores it in a database would be both an 'extractor' and a 'database' module.
 
     Each module is a python package, and should have a __manifest__.py file in the
     same directory as the module file. The __manifest__.py specifies the module information
-    like name, author, version, dependencies etc. See BaseModule._DEFAULT_MANIFEST for the
+    like name, author, version, dependencies etc. See DEFAULT_MANIFEST for the
     default manifest structure.
 
     """
 
-    MODULE_TYPES = [
-        'feeder',
-        'extractor',
-        'enricher',
-        'database',
-        'storage',
-        'formatter'
-    ]
-
-    _DEFAULT_MANIFEST = {
-    'name': '', # the display name of the module
-    'author': 'Bellingcat', # creator of the module, leave this as Bellingcat or set your own name!
-    'type': [], # the type of the module, can be one or more of BaseModule.MODULE_TYPES
-    'requires_setup': True, # whether or not this module requires additional setup such as setting API Keys or installing additional softare
-    'description': '', # a description of the module
-    'dependencies': {}, # external dependencies, e.g. python packages or binaries, in dictionary format
-    'entry_point': '', # the entry point for the module, in the format 'module_name::ClassName'. This can be left blank to use the default entry point of module_name::ModuleName
-    'version': '1.0', # the version of the module
-    'configs': {} # any configuration options this module has, these will be exposed to the user in the config file or via the command line
-}
+    MODULE_TYPES = CONF_MODULE_TYPES
 
+    # NOTE: these here are declard as class variables, but they are overridden by the instance variables in the __init__ method
     config: Mapping[str, Any]
     authentication: Mapping[str, Mapping[str, str]]
     name: str
+    module_factory: ModuleFactory
 
     # this is set by the orchestrator prior to archiving
     tmp_dir: TemporaryDirectory = None

src/auto_archiver/core/config.py

@@ -11,7 +11,7 @@
 from loguru import logger
 
 from copy import deepcopy
-from .module import BaseModule
+from auto_archiver.core.consts import MODULE_TYPES
 
 from typing import Any, List, Type, Tuple
 
@@ -21,7 +21,7 @@
 # Auto Archiver Configuration
 # Steps are the modules that will be run in the order they are defined
 
-steps:""" + "".join([f"\n   {module}s: []" for module in BaseModule.MODULE_TYPES]) + \
+steps:""" + "".join([f"\n   {module}s: []" for module in MODULE_TYPES]) + \
 """
 
 # Global configuration

src/auto_archiver/core/consts.py

@@ -0,0 +1,23 @@
+
+MODULE_TYPES = [
+    'feeder',
+    'extractor',
+    'enricher',
+    'database',
+    'storage',
+    'formatter'
+]
+
+MANIFEST_FILE = "__manifest__.py"
+
+DEFAULT_MANIFEST = {
+    'name': '', # the display name of the module
+    'author': 'Bellingcat', # creator of the module, leave this as Bellingcat or set your own name!
+    'type': [], # the type of the module, can be one or more of MODULE_TYPES
+    'requires_setup': True, # whether or not this module requires additional setup such as setting API Keys or installing additional softare
+    'description': '', # a description of the module
+    'dependencies': {}, # external dependencies, e.g. python packages or binaries, in dictionary format
+    'entry_point': '', # the entry point for the module, in the format 'module_name::ClassName'. This can be left blank to use the default entry point of module_name::ModuleName
+    'version': '1.0', # the version of the module
+    'configs': {} # any configuration options this module has, these will be exposed to the user in the config file or via the command line
+}