diff --git a/kodak/database/__init__.py b/kodak/database/__init__.py
index bbfc7a5..54cde94 100644
--- a/kodak/database/__init__.py
+++ b/kodak/database/__init__.py
@@ -8,6 +8,7 @@ import peewee
 from kodak import constants
 from kodak import exceptions
 from kodak.configuration import KodakConfig
+from kodak.database._shared import Checksum
 from kodak.database._shared import INTERFACE as interface
 from kodak.database._shared import KodakModel
 from kodak.database.access import AccessRecord
diff --git a/kodak/database/_shared.py b/kodak/database/_shared.py
index dcef188..036abfc 100644
--- a/kodak/database/_shared.py
+++ b/kodak/database/_shared.py
@@ -1,11 +1,14 @@
 import datetime
 import enum
+import hashlib
 import typing
 import uuid
 from pathlib import Path
+from typing import Callable
 from typing import Dict
 from typing import NamedTuple
 from typing import Type
+from typing import Union
 
 import peewee
 
@@ -31,6 +34,24 @@ class Checksum(NamedTuple):
         """Construct from a hashlib object"""
         return cls(algorithm=data.name, digest=data.hexdigest())
 
+    @classmethod
+    def from_path(cls, path: Union[str, Path], constructor: Callable = hashlib.sha256):
+        """Construct from a file path, generating the hash of the file
+
+        .. note:: This method attempts to _efficiently_ compute a hash of large image files. The
+                  hashing code was adapted from here:
+
+                  https://stackoverflow.com/a/44873382/5361209
+        """
+
+        hasher = constructor()
+        view = memoryview(bytearray(1024 * 1024))
+        with path.open("rb", buffering=0) as infile:
+            for chunk in iter(lambda: infile.readinto(view), 0):  # type: ignore
+                hasher.update(view[:chunk])
+
+        return cls.from_hash(hasher)
+
     def as_header(self) -> str:
         """Format the checksum for the Content-Digest HTTP header"""
         if self.algorithm.startswith("sha"):
diff --git a/kodak/database/image.py b/kodak/database/image.py
index 3a173e6..cc28fb3 100644
--- a/kodak/database/image.py
+++ b/kodak/database/image.py
@@ -1,4 +1,3 @@
-import hashlib
 import os
 from pathlib import Path
 
@@ -29,18 +28,7 @@ class ImageRecord(KodakModel):
         :param config: Populated application configuration object
         :param path: Full path to the image file to process. The file path provided is expected to
                      already be absolute, with all symlinks and aliases resolved.
-
-        .. note:: This method attempts to _efficiently_ compute a hash of large image files. The
-                  hashing code was adapted from here:
-
-                  https://stackoverflow.com/a/44873382/5361209
         """
-        hasher = hashlib.sha256()
-        view = memoryview(bytearray(1024 * 1024))
-        with path.open("rb", buffering=0) as infile:
-            for chunk in iter(lambda: infile.readinto(view), 0):  # type: ignore
-                hasher.update(view[:chunk])
-
         name = path.stem
         extension = path.suffix
 
@@ -56,7 +44,7 @@ class ImageRecord(KodakModel):
         )
 
         return cls(
-            name=name, source=path, format_=format_, checksum=Checksum.from_hash(hasher)
+            name=name, source=path, format_=format_, checksum=Checksum.from_path(path)
         )
 
     def create_link(self, config: configuration.KodakConfig) -> Path:
@@ -65,7 +53,8 @@ class ImageRecord(KodakModel):
         :param config: Populated application configuration object
         :returns: Path to the created symbolic link back to the source file
         """
-        link = Path(config.content_dir, self.name)
+        Path(config.content_dir, self.name).mkdir(exist_ok=True)
+        link = Path(config.content_dir, self.name, "original")
         try:
             link.symlink_to(self.source)
         except FileExistsError:
@@ -77,4 +66,4 @@ class ImageRecord(KodakModel):
 
         :param config: Populated application configuration object
         """
-        Path(config.content_dir, self.name).unlink(missing_ok=True)
+        Path(config.content_dir, self.name, "original").unlink(missing_ok=True)