Simpler thumbnail caching

src/documents/caching.py

@@ -8,7 +8,23 @@ CACHE_1_MINUTE: Final[int] = 60
 CACHE_5_MINUTES: Final[int] = 5 * CACHE_1_MINUTE
 CACHE_50_MINUTES: Final[int] = 50 * CACHE_1_MINUTE
 
-DOC_SUGGESTIONS_BASE: Final[str] = "doc_{}_suggest"
-DOC_METADATA_BASE: Final[str] = "doc_{}_metadata"
-DOC_THUMBNAIL_ETAG_BASE: Final[str] = "doc_{}_thumbnail_etag"
-DOC_THUMBNAIL_MODIFIED_BASE: Final[str] = "doc_{}_thumbnail_modified"
+
+def get_suggestion_key(document_id: int) -> str:
+    """
+    Builds the key to store a document's suggestion data in the cache
+    """
+    return f"doc_{document_id}_suggest"
+
+
+def get_metadata_key(document_id: int) -> str:
+    """
+    Builds the key to store a document's metadata data in the cache
+    """
+    return f"doc_{document_id}_metadata"
+
+
+def get_thumbnail_modified_key(document_id: int) -> str:
+    """
+    Builds the key to store a thumbnail's timestamp
+    """
+    return f"doc_{document_id}_thumbnail_modified"

src/documents/classifier.py

@@ -213,6 +213,15 @@ class DocumentClassifier:
             and self.last_doc_change_time >= latest_doc_change
         ) and self.last_auto_type_hash == hasher.digest():
             logger.info("No updates since last training")
+            # Set the classifier information into the cache
+            # Caching for 50 minutes, so slightly less than the normal retrain time
+            cache.set(
+                CLASSIFIER_MODIFIED_KEY,
+                self.last_doc_change_time,
+                CACHE_50_MINUTES,
+            )
+            cache.set(CLASSIFIER_HASH_KEY, hasher.hexdigest(), CACHE_50_MINUTES)
+            cache.set(CLASSIFIER_VERSION_KEY, self.FORMAT_VERSION, CACHE_50_MINUTES)
             return False
 
         # subtract 1 since -1 (null) is also part of the classes.

src/documents/conditionals.py

@@ -1,6 +1,5 @@
 from datetime import datetime
 from datetime import timezone
-from hashlib import sha256
 from typing import Optional
 
 from django.conf import settings
@@ -11,8 +10,7 @@ from documents.caching import CACHE_50_MINUTES
 from documents.caching import CLASSIFIER_HASH_KEY
 from documents.caching import CLASSIFIER_MODIFIED_KEY
 from documents.caching import CLASSIFIER_VERSION_KEY
-from documents.caching import DOC_THUMBNAIL_ETAG_BASE
-from documents.caching import DOC_THUMBNAIL_MODIFIED_BASE
+from documents.caching import get_thumbnail_modified_key
 from documents.classifier import DocumentClassifier
 from documents.models import Document
 
@@ -113,7 +111,10 @@ def preview_etag(request, pk: int) -> Optional[str]:
 
 
 def preview_last_modified(request, pk: int) -> Optional[str]:
-    """ """
+    """
+    Uses the documents modified time to set the Last-Modified header.  Not strictly
+    speaking correct, but close enough and quick
+    """
     try:
         doc = Document.objects.get(pk=pk)
         return doc.modified
@@ -122,26 +123,6 @@ def preview_last_modified(request, pk: int) -> Optional[str]:
     return None
 
 
-def thumbnail_etag(request, pk: int) -> Optional[str]:
-    """
-    Returns the SHA256 of a thumbnail, either from cache or calculated
-    """
-    try:
-        doc = Document.objects.get(pk=pk)
-        if not doc.thumbnail_path.exists():
-            return None
-        doc_key = DOC_THUMBNAIL_ETAG_BASE.format(pk)
-        cache_hit = cache.get(doc_key)
-        if cache_hit is not None:
-            return cache_hit
-        hasher = sha256()
-        hasher.update(doc.thumbnail_path.read_bytes())
-        thumb_checksum = hasher.hexdigest()
-        cache.set(doc_key, thumb_checksum, CACHE_50_MINUTES)
-    except Document.DoesNotExist:  # pragma: no cover
-        return None
-
-
 def thumbnail_last_modified(request, pk: int) -> Optional[int]:
     """
     Returns the filesystem last modified either from cache or from filesystem
@@ -150,10 +131,14 @@ def thumbnail_last_modified(request, pk: int) -> Optional[int]:
         doc = Document.objects.get(pk=pk)
         if not doc.thumbnail_path.exists():
             return None
-        doc_key = DOC_THUMBNAIL_MODIFIED_BASE.format(pk)
+        doc_key = get_thumbnail_modified_key(pk)
+
         cache_hit = cache.get(doc_key)
         if cache_hit is not None:
+            cache.touch(doc_key, CACHE_50_MINUTES)
             return cache_hit
+
+        # No cache, get the timestamp and cache the datetime
         last_modified = datetime.fromtimestamp(
             doc.thumbnail_path.stat().st_mtime,
             tz=timezone.utc,

src/documents/views.py

@@ -36,6 +36,7 @@ from django.utils.translation import get_language
 from django.views import View
 from django.views.decorators.cache import cache_control
 from django.views.decorators.http import condition
+from django.views.decorators.http import last_modified
 from django.views.generic import TemplateView
 from django_filters.rest_framework import DjangoFilterBackend
 from langdetect import detect
@@ -65,8 +66,8 @@ from documents.bulk_download import OriginalAndArchiveStrategy
 from documents.bulk_download import OriginalsOnlyStrategy
 from documents.caching import CACHE_5_MINUTES
 from documents.caching import CACHE_50_MINUTES
-from documents.caching import DOC_METADATA_BASE
-from documents.caching import DOC_SUGGESTIONS_BASE
+from documents.caching import get_metadata_key
+from documents.caching import get_suggestion_key
 from documents.classifier import load_classifier
 from documents.conditionals import metadata_etag
 from documents.conditionals import metadata_last_modified
@@ -74,7 +75,6 @@ from documents.conditionals import preview_etag
 from documents.conditionals import preview_last_modified
 from documents.conditionals import suggestions_etag
 from documents.conditionals import suggestions_last_modified
-from documents.conditionals import thumbnail_etag
 from documents.conditionals import thumbnail_last_modified
 from documents.data_models import ConsumableDocument
 from documents.data_models import DocumentMetadataOverrides
@@ -415,7 +415,7 @@ class DocumentViewSet(
         except Document.DoesNotExist:
             raise Http404
 
-        doc_key = DOC_METADATA_BASE.format(doc.pk)
+        doc_key = get_metadata_key(doc.pk)
 
         cache_hit = cache.get(doc_key)
 
@@ -472,7 +472,7 @@ class DocumentViewSet(
         ):
             return HttpResponseForbidden("Insufficient permissions")
 
-        doc_key = DOC_SUGGESTIONS_BASE.format(doc.pk)
+        doc_key = get_suggestion_key(doc.pk)
 
         cache_hit = cache.get(doc_key)
 
@@ -521,9 +521,7 @@ class DocumentViewSet(
 
     @action(methods=["get"], detail=True)
     @method_decorator(cache_control(public=False, max_age=CACHE_50_MINUTES))
-    @method_decorator(
-        condition(etag_func=thumbnail_etag, last_modified_func=thumbnail_last_modified),
-    )
+    @method_decorator(last_modified(thumbnail_last_modified))
     def thumb(self, request, pk=None):
         try:
             doc = Document.objects.get(id=pk)

src/paperless/settings.py

@@ -767,6 +767,7 @@ CACHES = {
             "django.core.cache.backends.redis.RedisCache",
         ),
         "LOCATION": _CHANNELS_REDIS_URL,
+        "KEY_PREFIX": os.getenv("PAPERLESS_REDIS_PREFIX", ""),
     },
 }