revert a faulty change that caused memory usage to explode #537

src/documents/classifier.py

@@ -5,7 +5,6 @@ import pickle
 import re
 
 from django.conf import settings
-from django.core.cache import cache
 
 from documents.models import Document, MatchingModel
 
@@ -31,29 +30,23 @@ def load_classifier():
         )
         return None
 
-    version = os.stat(settings.MODEL_FILE).st_mtime
+    classifier = DocumentClassifier()
+    try:
+        classifier.load()
 
-    classifier = cache.get("paperless-classifier", version=version)
-
-    if not classifier:
-        classifier = DocumentClassifier()
-        try:
-            classifier.load()
-            cache.set("paperless-classifier", classifier,
-                      version=version, timeout=86400)
-        except (EOFError, IncompatibleClassifierVersionError) as e:
-            # there's something wrong with the model file.
-            logger.exception(
-                f"Unrecoverable error while loading document "
-                f"classification model, deleting model file."
-            )
-            os.unlink(settings.MODEL_FILE)
-            classifier = None
-        except OSError as e:
-            logger.error(
-                f"Error while loading document classification model: {str(e)}"
-            )
-            classifier = None
+    except (EOFError, IncompatibleClassifierVersionError) as e:
+        # there's something wrong with the model file.
+        logger.exception(
+            f"Unrecoverable error while loading document "
+            f"classification model, deleting model file."
+        )
+        os.unlink(settings.MODEL_FILE)
+        classifier = None
+    except OSError as e:
+        logger.error(
+            f"Error while loading document classification model: {str(e)}"
+        )
+        classifier = None
 
     return classifier
 

src/documents/tests/test_classifier.py

@@ -3,6 +3,7 @@ import tempfile
 from pathlib import Path
 from unittest import mock
 
+import pytest
 from django.conf import settings
 from django.test import TestCase, override_settings
 
@@ -233,7 +234,6 @@ class TestClassifier(DirectoriesMixin, TestCase):
         self.assertFalse(os.path.exists(settings.MODEL_FILE))
         self.assertIsNone(load_classifier())
 
-    @override_settings(CACHES={'default': {'BACKEND': 'django.core.cache.backends.dummy.DummyCache'}})
     @mock.patch("documents.classifier.DocumentClassifier.load")
     def test_load_classifier(self, load):
         Path(settings.MODEL_FILE).touch()
@@ -242,6 +242,7 @@ class TestClassifier(DirectoriesMixin, TestCase):
 
     @override_settings(CACHES={'default': {'BACKEND': 'django.core.cache.backends.locmem.LocMemCache'}})
     @override_settings(MODEL_FILE=os.path.join(os.path.dirname(__file__), "data", "model.pickle"))
+    @pytest.mark.skip(reason="Disabled caching due to high memory usage - need to investigate.")
     def test_load_classifier_cached(self):
         classifier = load_classifier()
         self.assertIsNotNone(classifier)
@@ -250,7 +251,6 @@ class TestClassifier(DirectoriesMixin, TestCase):
             classifier2 = load_classifier()
             load.assert_not_called()
 
-    @override_settings(CACHES={'default': {'BACKEND': 'django.core.cache.backends.dummy.DummyCache'}})
     @mock.patch("documents.classifier.DocumentClassifier.load")
     def test_load_classifier_incompatible_version(self, load):
         Path(settings.MODEL_FILE).touch()
@@ -260,7 +260,6 @@ class TestClassifier(DirectoriesMixin, TestCase):
         self.assertIsNone(load_classifier())
         self.assertFalse(os.path.exists(settings.MODEL_FILE))
 
-    @override_settings(CACHES={'default': {'BACKEND': 'django.core.cache.backends.dummy.DummyCache'}})
     @mock.patch("documents.classifier.DocumentClassifier.load")
     def test_load_classifier_os_error(self, load):
         Path(settings.MODEL_FILE).touch()

src/documents/tests/test_tasks.py

@@ -52,7 +52,6 @@ class TestTasks(DirectoriesMixin, TestCase):
         load_classifier.assert_called_once()
         self.assertFalse(os.path.isfile(settings.MODEL_FILE))
 
-    @override_settings(CACHES={'default': {'BACKEND': 'django.core.cache.backends.dummy.DummyCache'}})
     def test_train_classifier(self):
         c = Correspondent.objects.create(matching_algorithm=Tag.MATCH_AUTO, name="test")
         doc = Document.objects.create(correspondent=c, content="test", title="test")

src/paperless/settings.py

@@ -169,16 +169,6 @@ CHANNEL_LAYERS = {
     },
 }
 
-CACHES = {
-    "default": {
-        "BACKEND": "django_redis.cache.RedisCache",
-        "LOCATION": os.getenv("PAPERLESS_REDIS", "redis://localhost:6379"),
-        "OPTIONS": {
-            "CLIENT_CLASS": "django_redis.client.DefaultClient",
-        }
-    }
-}
-
 ###############################################################################
 # Security                                                                    #
 ###############################################################################