Merge pull request #76 from james-acsc/fix/collector-compat

fix: resolve runtime collector failure

maco/collector.py

@@ -5,11 +5,11 @@
 import logging.handlers
 import os
 import sys
-from multiprocessing import Manager, Process, Queue
 from tempfile import NamedTemporaryFile
 from types import ModuleType
-from typing import Any, BinaryIO, Dict, List, Union
+from typing import Any, BinaryIO, Dict, List, TypedDict, Union
 
+from multiprocess import Manager, Process, Queue
 from pydantic import BaseModel
 
 from maco import extractor, model, utils, yara
@@ -40,6 +40,26 @@ def _verify_response(resp: Union[BaseModel, dict]) -> Dict:
     return resp.model_dump(exclude_defaults=True)
 
 
+class ExtractorMetadata(TypedDict):
+    """Extractor-supplied metadata."""
+
+    author: str
+    family: str
+    last_modified: str
+    sharing: str
+    description: str
+
+
+class ExtractorRegistration(TypedDict):
+    """Registration collected by the collector for a single extractor."""
+
+    venv: str
+    module_path: str
+    module_name: str
+    extractor_class: str
+    metadata: ExtractorMetadata
+
+
 class Collector:
     def __init__(
         self,
@@ -60,7 +80,7 @@ def __init__(
 
         path_extractors = os.path.realpath(path_extractors)
         self.path: str = path_extractors
-        self.extractors: Dict[str, Dict[str, str]] = {}
+        self.extractors: Dict[str, ExtractorRegistration] = {}
 
         with Manager() as manager:
             extractors = manager.dict()

maco/utils.py

@@ -6,14 +6,15 @@
 import json
 import logging
 import logging.handlers
-import multiprocessing
 import os
 import re
 import shutil
 import subprocess
 import sys
 import tempfile
 
+from multiprocess import Queue
+
 from maco import yara
 
 if sys.version_info >= (3, 11):
@@ -390,7 +391,7 @@ def register_extractors(
                 break
 
 
-def proxy_logging(queue: multiprocessing.Queue, callback: Callable[[ModuleType, str], None], *args, **kwargs):
+def proxy_logging(queue: Queue, callback: Callable[[ModuleType, str], None], *args, **kwargs):
     """Ensures logging is set up correctly for a child process and then executes the callback."""
     logger = logging.getLogger()
     qh = logging.handlers.QueueHandler(queue)

requirements.txt

@@ -4,3 +4,4 @@ tomli >= 1.1.0 ; python_version < "3.11"
 uv
 yara-python
 yara-x==0.11.0
+multiprocess>=0.70.17