lgcosta/paperless-ngx
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

better metadata caching

Browse Source
This commit is contained in:
Trenton H 2024-01-30 11:02:13 -08:00
parent 70a2dde757
commit 961c97771d
2 changed files with 29 additions and 20 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

8
src/documents/caching.py
View File

@ -16,11 +16,15 @@ def get_suggestion_key(document_id: int) -> str:
return f"doc_{document_id}_suggest" return f"doc_{document_id}_suggest"
def get_metadata_key(document_id: int) -> str: def get_metadata_key(document_id: int, is_archive: bool) -> str:
""" """
Builds the key to store a document's metadata data in the cache Builds the key to store a document's metadata data in the cache
""" """
return f"doc_{document_id}_metadata" return (
f"doc_{document_id}_archive_metadata"
if is_archive
else f"doc_{document_id}_original_metadata"
)
def get_thumbnail_modified_key(document_id: int) -> str: def get_thumbnail_modified_key(document_id: int) -> str:

41
src/documents/views.py
View File

@ -415,13 +415,28 @@ class DocumentViewSet(
except Document.DoesNotExist: except Document.DoesNotExist:
raise Http404 raise Http404
doc_key = get_metadata_key(doc.pk) doc_original_key = get_metadata_key(doc.pk, is_archive=False)
doc_archive_key = get_metadata_key(doc.pk, is_archive=True)
cache_hit = cache.get(doc_key) cache_hits = cache.get_many([doc_original_key, doc_archive_key])
if cache_hit is not None: # use cached original file metadata if possible, else gather then cache
cache.touch(doc_key, CACHE_5_MINUTES) if doc_original_key in cache_hits:
return Response(cache_hit) cache.touch(doc_original_key, CACHE_5_MINUTES)
original_metadata = cache_hits[doc_original_key]
else:
original_metadata = self.get_metadata(doc.source_path, doc.mime_type)
cache.set(doc_original_key, original_metadata, CACHE_5_MINUTES)
# use cached archive file metadata, if applicable, then cache if it wasn't
archive_metadata = None
archive_filesize = None
if doc.has_archive_version and doc_archive_key in cache_hits:
archive_metadata = cache_hits[doc_archive_key]
elif doc.has_archive_version:
archive_filesize = self.get_filesize(doc.archive_path)
archive_metadata = self.get_metadata(doc.archive_path, "application/pdf")
cache.set(doc_archive_key, archive_metadata, CACHE_5_MINUTES)
meta = { meta = {
"original_checksum": doc.checksum, "original_checksum": doc.checksum,
@ -429,10 +444,12 @@ class DocumentViewSet(
"original_mime_type": doc.mime_type, "original_mime_type": doc.mime_type,
"media_filename": doc.filename, "media_filename": doc.filename,
"has_archive_version": doc.has_archive_version, "has_archive_version": doc.has_archive_version,
"original_metadata": self.get_metadata(doc.source_path, doc.mime_type), "original_metadata": original_metadata,
"archive_checksum": doc.archive_checksum, "archive_checksum": doc.archive_checksum,
"archive_media_filename": doc.archive_filename, "archive_media_filename": doc.archive_filename,
"original_filename": doc.original_filename, "original_filename": doc.original_filename,
"archive_size": archive_filesize,
"archive_metadata": archive_metadata,
} }
lang = "en" lang = "en"
@ -442,18 +459,6 @@ class DocumentViewSet(
pass pass
meta["lang"] = lang meta["lang"] = lang
if doc.has_archive_version:
meta["archive_size"] = self.get_filesize(doc.archive_path)
meta["archive_metadata"] = self.get_metadata(
doc.archive_path,
"application/pdf",
)
else:
meta["archive_size"] = None
meta["archive_metadata"] = None
cache.set(doc_key, meta, CACHE_5_MINUTES)
return Response(meta) return Response(meta)
@action(methods=["get"], detail=True) @action(methods=["get"], detail=True)