From 9befb9776c3fb311d570d03140de3ab4e22878ef Mon Sep 17 00:00:00 2001
From: Patrick Robertson <robertson.patrick@gmail.com>
Date: Thu, 23 Jan 2025 21:08:54 +0100
Subject: [PATCH] Fix loading modules when entry_point isn't set

---
 src/auto_archiver/core/config.py              | 11 +++++-----
 src/auto_archiver/core/loader.py              | 21 ++++++++++++++++---
 src/auto_archiver/core/orchestrator.py        |  2 --
 .../modules/generic_extractor/__manifest__.py |  1 -
 tests/test_config.py                          |  4 ++++
 5 files changed, 28 insertions(+), 11 deletions(-)

diff --git a/src/auto_archiver/core/config.py b/src/auto_archiver/core/config.py
index 3811e2b..1c19ae2 100644
--- a/src/auto_archiver/core/config.py
+++ b/src/auto_archiver/core/config.py
@@ -100,16 +100,17 @@ def merge_dicts(dotdict: dict, yaml_dict: CommentedMap) -> CommentedMap:
 
     # first deal with lists, since 'update' replaces lists from a in b, but we want to extend
     def update_dict(subdict, yaml_subdict):
-        for key, value in yaml_subdict.items():
-            if not subdict.get(key):
+        for key, value in subdict.items():
+            if not yaml_subdict.get(key):
+                yaml_subdict[key] = value
                 continue
 
             if is_dict_type(value):
-                update_dict(subdict[key], value)
+                update_dict(value, yaml_subdict[key])
             elif is_list_type(value):
-                yaml_subdict[key].extend(s for s in subdict[key] if s not in yaml_subdict[key])
+                yaml_subdict[key].extend(s for s in value if s not in yaml_subdict[key])
             else:
-                yaml_subdict[key] = subdict[key]
+                yaml_subdict[key] = value
 
     update_dict(from_dot_notation(dotdict), yaml_dict)
 
diff --git a/src/auto_archiver/core/loader.py b/src/auto_archiver/core/loader.py
index 310e0e6..bbd686e 100644
--- a/src/auto_archiver/core/loader.py
+++ b/src/auto_archiver/core/loader.py
@@ -25,6 +25,7 @@ MANIFEST_FILE = "__manifest__.py"
 _DEFAULT_MANIFEST = {
     'name': '',
     'author': 'Bellingcat',
+    'type': [],
     'requires_setup': True,
     'description': '',
     'dependencies': {},
@@ -90,8 +91,18 @@ def load_module(module: str) -> object: # TODO: change return type to Step
     qualname = f'auto_archiver.modules.{module.name}'
 
     logger.info(f"Loading module '{module.display_name}'...")
-    loaded_module = __import__(qualname)
-    instance = getattr(sys.modules[qualname], module.entry_point)()
+    # first import the whole module, to make sure it's working properly
+    __import__(qualname)
+
+    
+    # then import the file for the entry point
+    file_name, class_name = module.entry_point.split('::')
+    sub_qualname = f'{qualname}.{file_name}'
+
+    __import__(f'{qualname}.{file_name}', fromlist=[module.entry_point])
+    
+    # finally, get the class instance
+    instance = getattr(sys.modules[sub_qualname], class_name)()
     if not getattr(instance, 'name', None):
         instance.name = module.name
 
@@ -107,7 +118,11 @@ def load_manifest(module_path):
     manifest = copy.deepcopy(_DEFAULT_MANIFEST)
 
     with open(join(module_path, MANIFEST_FILE)) as f:
-        manifest.update(ast.literal_eval(f.read()))
+        try:
+            manifest.update(ast.literal_eval(f.read()))
+        except ( ValueError, TypeError, SyntaxError, MemoryError, RecursionError) as e:
+            logger.error(f"Error loading manifest from file {module_path}/{MANIFEST_FILE}: {e}")
+            return manifest
     return manifest
 
 def get_module(module_name):
diff --git a/src/auto_archiver/core/orchestrator.py b/src/auto_archiver/core/orchestrator.py
index b17dcec..2c9841e 100644
--- a/src/auto_archiver/core/orchestrator.py
+++ b/src/auto_archiver/core/orchestrator.py
@@ -109,7 +109,6 @@ class ArchivingOrchestrator:
         
         parser.set_defaults(**to_dot_notation(yaml_config))
 
-        breakpoint()
         # reload the parser with the new arguments, now that we have them
         parsed, unknown = parser.parse_known_args(unused_args)
 
@@ -180,7 +179,6 @@ class ArchivingOrchestrator:
     def setup_logging(self):
         # setup loguru logging
         logger.remove() # remove the default logger
-
         logging_config = self.config['logging']
         logger.add(sys.stderr, level=logging_config['level'])
         if log_file := logging_config['file']:
diff --git a/src/auto_archiver/modules/generic_extractor/__manifest__.py b/src/auto_archiver/modules/generic_extractor/__manifest__.py
index 6f469c9..f46c13c 100644
--- a/src/auto_archiver/modules/generic_extractor/__manifest__.py
+++ b/src/auto_archiver/modules/generic_extractor/__manifest__.py
@@ -3,7 +3,6 @@
     'version': '0.1.0',
     'author': 'Bellingcat',
     'type': ['extractor', 'feeder', 'enricher'],
-    'entry_point': 'GenericExtractor', # this class should be present in the __init__.py
     'requires_setup': False,
     'dependencies': {
         'python': ['yt_dlp', 'requests', 'loguru', 'slugify'],
diff --git a/tests/test_config.py b/tests/test_config.py
index 97793a0..75fe515 100644
--- a/tests/test_config.py
+++ b/tests/test_config.py
@@ -49,17 +49,21 @@ def test_merge_dicts():
             "key1": ["a"],
             "key2": "old_value",
             "key3": ["a", "b", "c"],
+            "key5": "value5",
         })
 
     dotdict = {
         "settings.key1": ["b", "c"],
         "settings.key2": "new_value",
         "settings.key3": ["b", "c", "d"],
+        "settings.key4": "value4",
     }
     merged = config.merge_dicts(dotdict, yaml_dict)
     assert merged["settings"]["key1"] == ["a", "b", "c"]
     assert merged["settings"]["key2"] == "new_value"
     assert merged["settings"]["key3"] == ["a", "b", "c", "d"]
+    assert merged["settings"]["key4"] == "value4"
+    assert merged["settings"]["key5"] == "value5"
 
 
 def test_check_types():