separate filename and filehash (#21)

* separate filename and filehash

* rename migration

* test url

* bump version

* update changelog

* test fail url

* update coveragerc

* add coveragerc

* EOFNL

* update covrc

* filename

.coveragerc

@@ -11,3 +11,4 @@ exclude_lines =
 ignore_errors = True
 omit =
     tests/*
+    dbfilestorage/migrations/*

README.md

@@ -54,6 +54,7 @@ I will sign everything with 0x769A1BC78A2DDEE2
 
 ## CHANGELOG
 
+- 2017-01-20 [Tyrel Souza] Split filename and filehash.
 - 2016-12-09 [Tyrel Souza] Add signing key to readme.
 - 2016-12-09 [Tyrel Souza] Update Tests, add some cleanup.
 - 2016-12-08 [Tyrel Souza] Add more documentation.

dbfilestorage/migrations/0002_add_filehash_rename_files.py

@@ -0,0 +1,43 @@
+# -*- coding: utf-8 -*-
+# Generated by Django 1.10.4 on 2017-01-20 18:44
+from __future__ import unicode_literals
+
+from django.db import migrations, models
+
+def copy_filehash(apps, schema_editor):
+    DBFile = apps.get_model('dbfilestorage', 'DBFile')
+    for dbf in DBFile.objects.all():
+        dbf.filehash = dbf.name
+        dbf.save()
+
+def fix_filename(apps, schema_editor):
+    DBFile = apps.get_model('dbfilestorage', 'DBFile')
+    for dbf in DBFile.objects.all():
+        ext = dbf.content_type.split("/")[0]
+        dbf.name = "{name}.{ext}".format(
+            name=dbf.name,
+            ext=ext)
+        dbf.save()
+
+
+class Migration(migrations.Migration):
+
+    dependencies = [
+        ('dbfilestorage', '0001_initial'),
+    ]
+
+    operations = [
+        migrations.AddField(
+            model_name='dbfile',
+            name='filehash',
+            field=models.CharField(default='filehash', max_length=32, primary_key=True, serialize=False),
+            preserve_default=False,
+        ),
+        migrations.AlterField(
+            model_name='dbfile',
+            name='name',
+            field=models.CharField(max_length=100),
+        ),
+        migrations.RunPython(copy_filehash),
+        migrations.RunPython(fix_filename),
+    ]

dbfilestorage/models.py

@@ -7,15 +7,16 @@ class DBFile(models.Model):
     # This is kept as `name` and not something like `md5` because the file
     # operations pass around `name` as the identifier, so it's kept the same
     # to make sense.
-    name = models.CharField(max_length=100, primary_key=True)
+    name = models.CharField(max_length=100)
+    filehash = models.CharField(max_length=32, primary_key=True)
 
     # file data
     content_type = models.CharField(max_length=100)
     b64 = models.TextField()
 
     def __unicode__(self):
-        return u"{name} <{content_type}>".format(
-            name=self.name, content_type=self.content_type)
+        return u"{filehash} <{content_type}>".format(
+            filehash=self.filehash, content_type=self.content_type)
 
     def save(self, **kwargs):
         if self.content_type is None:

dbfilestorage/storage.py

@@ -1,6 +1,7 @@
 import mimetypes
 import logging
 import hashlib
+import os
 
 from django.db.transaction import atomic
 from django.core.files.base import ContentFile
@@ -50,16 +51,17 @@ class DBFileStorage(Storage):
         ct = mimetypes.guess_type(name)[0]
 
         # After we get the mimetype by name potentially, mangle it.
-        name = hashlib.md5(read_data).hexdigest()
+        filehash = hashlib.md5(read_data).hexdigest()
 
         # create the file, or just return name if the exact file already exists
         if not DBFile.objects.filter(pk=name).exists():
             the_file = DBFile(
-                name=name,
+                name=os.path.basename(name),
+                filehash=filehash,
                 content_type=ct,
                 b64=b64)
             the_file.save()
-        return name
+        return filehash
 
     def get_available_name(self, name, max_length=None):
         return name

dbfilestorage/views.py

@@ -1,4 +1,5 @@
-from django.http import HttpResponse
+from django.http import HttpResponse, Http404
+from django.db.models import Q
 from django.shortcuts import get_object_or_404
 
 from .models import DBFile
@@ -12,7 +13,12 @@ def show_file(request, name):
 
         :return HttpResponse: Rendered file
     """
-    dbf = get_object_or_404(DBFile, pk=name)
-    return HttpResponse(
-        dbf.b64.decode('base64'),
-        content_type=dbf.content_type)
+    dbf = DBFile.objects.filter(Q(name=name)|Q(filehash=name))
+    if dbf.exists():
+        response = HttpResponse(
+            dbf[0].b64.decode('base64'),
+            content_type=dbf[0].content_type)
+        response['Content-Disposition'] = 'attachment; filename="{}"'.format(
+            dbf[0].name)
+        return response
+    raise Http404

docs/conf.py

@@ -54,9 +54,9 @@ author = u'Tyrel Souza'
 # built documents.
 #
 # The short X.Y version.
-version = u'0.1.3'
+version = u'0.1.4'
 # The full version, including alpha/beta/rc tags.
-release = u'0.1.3'
+release = u'0.1.4'
 
 # The language for content autogenerated by Sphinx. Refer to documentation
 # for a list of supported languages.

setup.py

@@ -24,7 +24,7 @@ class CleanCommand(Command):
 
 setup(
     name="django-dbfilestorage",
-    version="0.1.3",
+    version="0.1.4",
     description="Database backed file storage for testing.",
     long_description="Database backed file storage for testing. Stores files as base64 encoded textfields.",
     author="Tyrel Souza",

tests/tests.py

@@ -30,14 +30,14 @@ class DBFileTest(TestCase):
 
     def test_upload(self):
         """ Test that the file storage uploads and puts in DB Properly """
-        self.assertTrue(DBFile.objects.filter(name=self.md5).exists())
+        self.assertTrue(DBFile.objects.filter(filehash=self.md5).exists())
 
     def test_content_file(self):
         """ Test that this code works with ContentFile as well """
         content_file = ContentFile(u"ΑΔΔGΕΝΕ")
         content_file_md5 = hashlib.md5(u"ΑΔΔGΕΝΕ".encode('utf8')).hexdigest()
         default_storage.save("unicode", content_file)
-        unicode_file = DBFile.objects.get(name=content_file_md5)
+        unicode_file = DBFile.objects.get(filehash=content_file_md5)
         self.assertEqual(unicode(unicode_file),
             "{} <application/octet-stream>".format(content_file_md5))
 
@@ -50,7 +50,7 @@ class DBFileTest(TestCase):
     def test_equality(self):
         """ Test that the DB entry matches what is expected from the file """
         with open(self.filepath, 'rb') as f:
-            dbf = DBFile.objects.get(name=self.md5)
+            dbf = DBFile.objects.get(filehash=self.md5)
             self.assertEqual(dbf.b64.decode("base64"), f.read())
             self.assertEqual(dbf.content_type, 'image/jpeg')
 
@@ -66,9 +66,9 @@ class DBFileTest(TestCase):
 
     def test_delete(self):
         """ Test Deletion """
-        self.assertTrue(DBFile.objects.filter(name=self.md5).exists())
+        self.assertTrue(DBFile.objects.filter(filehash=self.md5).exists())
         default_storage.delete(self.md5)
-        self.assertFalse(DBFile.objects.filter(name=self.md5).exists())
+        self.assertFalse(DBFile.objects.filter(filehash=self.md5).exists())
         # Also test that calling delete on something that doesn't exist,
         # errors silently
         self.assertFalse(DBFile.objects.filter(name="Nothing").exists())
@@ -91,10 +91,19 @@ class DBFileTest(TestCase):
 
     def test_view(self):
         client = Client()
-        url = default_storage.url(self.md5)
+        # check it works for both md5 and filename
+        for param in (self.md5, self.filename):
+            url = default_storage.url(param)
             resp = client.get(url)
             self.assertEqual(resp.status_code, 200)
             
+    def test_view_fails(self):
+        client = Client()
+        url = default_storage.url("failure")
+        resp = client.get(url)
+        self.assertEqual(resp.status_code, 404)
+
+
     def test_admin(self):
         my_admin = User.objects.create_superuser(
             username='tester',