#227 Gen filename from it's contents hash (#234)

* #227  Gen filename from it's contents hash

* #227  Fix comments and import style

images/migrations/0007_custom_image_field.py

@@ -0,0 +1,21 @@
+# -*- coding: utf-8 -*-
+# Generated by Django 1.10 on 2018-12-27 10:04
+from __future__ import unicode_literals
+
+from django.db import migrations
+import images.models
+
+
+class Migration(migrations.Migration):
+
+    dependencies = [
+        ('images', '0006_auto_20170418_1615'),
+    ]
+
+    operations = [
+        migrations.AlterField(
+            model_name='image',
+            name='image',
+            field=images.models.ImageField(upload_to='', verbose_name='image'),
+        ),
+    ]

images/models.py

@@ -1,5 +1,5 @@
+import hashlib
 import os
-import time
 
 from django.db import models
 from django.contrib.contenttypes.fields import GenericForeignKey, GenericRelation
@@ -9,20 +9,25 @@
 from sorl.thumbnail import delete
 
 
-# @todo #STB296:60m Generate image file name with hash.
-#  Now generation uses timestamp.
-#  It breaks media<->fixtures consistency.
-#  See comment for details:
-#  https://github.com/fidals/stroyprombeton/issues/296#issuecomment-447770194
-#  And test absolute filename generation.
-def model_directory_path(instance, filename):
-    def generate_filename(old_name: str, slug: str) -> str:
-        _, file_extension = os.path.splitext(old_name)
-        return (slug or str(time.time())) + file_extension
+# stayed here for migrations compatibility
+def model_directory_path(*args, **kwargs):
+    return ''
 
-    result_filename = generate_filename(filename, instance.slug)
-    models_folder_name = type(instance.model).__name__.lower()
-    return '{}/{}/{}'.format(models_folder_name, instance.model.pk, result_filename)
+
+class ImageField(thumbnail.ImageField):
+
+    def generate_filename(self, instance, filename):
+        """Reloaded this method. See `django.db.models.fields.files.FileField`."""
+        f = getattr(instance, self.name).file
+        _, extension = os.path.splitext(filename)
+        file_hash = hashlib.md5(f.read()).hexdigest()
+        models_folder_name = type(instance.model).__name__.lower()
+        filename = '/'.join([
+            models_folder_name,
+            str(instance.model.pk),
+            file_hash + extension
+        ])
+        return self.storage.generate_filename(filename)
 
 
 class ImageManager(models.Manager):
@@ -67,7 +72,7 @@ class Meta:
     slug = models.SlugField(max_length=400, blank=True, db_index=True, verbose_name=_('slug'))
     description = models.TextField(default='', blank=True, verbose_name=_('description'))
     created = models.DateField(auto_now_add=True, verbose_name=_('created'))
-    image = thumbnail.ImageField(upload_to=model_directory_path, verbose_name=_('image'))
+    image = ImageField(verbose_name=_('image'))
     is_main = models.BooleanField(default=False, db_index=True, verbose_name=_('is main'))
 
     def __str__(self):

tests/images/tests.py

@@ -1,4 +1,5 @@
 import os.path
+import string
 
 from django.core.files.images import ImageFile
 from django.db import models
@@ -94,3 +95,21 @@ def test_update_main_image(self):
         self.assertTrue(another_image_model.is_main)
         self.assertEquals(another_image_model.image, self.page.main_image)
         another_image_model.delete()
+
+    def test_file_name(self):
+        """ImageField should generate filename based on file's content hash."""
+
+        def is_md5(data: str) -> bool:
+            return (
+                all(s in (string.ascii_letters + string.digits) for s in data)
+                and len(data) == 32
+            )
+
+        model = create_image_model(self.page, self.IMG_PATH, slug='one-another')
+        # @todo #227:60m Explore additional hash.
+        #  Now autosaved file has this pattern `<md5>_dTy2ykr.jpg`
+        #  but should have this one: `<md5>.jpg`.
+        #  ManifestStaticFilesStorage adds redundant hash after md5.
+        #  Make the storage to remove this hash.
+        hash, _ = os.path.basename(model.image.file.file.name).split('_')
+        self.assertTrue(is_md5(hash))