Initial commit.

brejoc · brejoc · commit 199ddae5b846 · 2013-12-27T00:19:42.000+01:00
diff --git a/.hgignore b/.hgignore
@@ -0,0 +1,30 @@
+# use glob syntax.
+syntax: glob
+
+.DS_Store
+.svn
+*.svn-base
+.cvs
+*.elc
+*.pyc
+*~
+.*.swp
+nbproject
+build/
+*.mode1v3
+*.pbxuser
+*.db
+*.orig
+
+# switch to regexp syntax.
+syntax: regexp
+^.pc/
+^sizr/uploads
+^sizr/bucket
+^sizr/db.sql
+^sizr/out.html
+^sizr/sizr/settings_local.py
+^sizr/.idea/
+^sizr/out.png
+^.idea/
+^sizr/static/
diff --git a/README.md b/README.md
@@ -0,0 +1,20 @@
+# sizr - Fast scaling of images for web and mobile with caching
+
+sizr is a django app that can resize images on the fly with just a GET request. sizr then fetches the image, resizes it caches the image in a "bucket" and redirects you to the local copy. The next request then gets redirected directly.
+
+At the moment sizr has the status "prove of concept", so don't expect it to work with every image format or think of it as unbreakable.
+
+# Dependencies
+
+* Django 1.4.10
+* Grapelli 2.4.8
+* Python Requests (>= 0.8.2-1 | shipped with Ubuntu 12.04)
+
+
+# Some notes on base64
+
+sizr uses base64 to encode URLs. But it has to be URL-safe, which standard base64 isn't.
+
+iOS: http://stackoverflow.com/questions/11106393/url-safe-base64-in-objective-c
+Java: http://commons.apache.org/proper/commons-codec/apidocs/org/apache/commons/codec/binary/Base64.html#Base64%28boolean%29
+Perl: http://search.cpan.org/~kazuho/MIME-Base64-URLSafe-0.01/lib/MIME/Base64/URLSafe.pm
diff --git a/examples/UrlsafeBase64.java b/examples/UrlsafeBase64.java
@@ -0,0 +1,23 @@
+public static byte[] encodeUrlSafe(byte[] data) {
+    byte[] encode = Base64.encode(data);
+    for (int i = 0; i < encode.length; i++) {
+        if (encode[i] == '+') {
+            encode[i] = '-';
+        } else if (encode[i] == '/') {
+            encode[i] = '_';
+        }
+    }
+    return encode;
+}
+
+public static byte[] decodeUrlSafe(byte[] data) {
+    byte[] encode = Arrays.copyOf(data, data.length);
+    for (int i = 0; i < encode.length; i++) {
+        if (encode[i] == '-') {
+            encode[i] = '+';
+        } else if (encode[i] == '_') {
+            encode[i] = '/';
+        }
+    }
+    return Base64.decode(encode);
+}
diff --git a/examples/get_image.py b/examples/get_image.py
@@ -0,0 +1,14 @@
+#!/usr/bin/env python
+# -*- coding: utf-8 -*-
+
+
+import requests
+from urllib2 import quote
+from base64 import urlsafe_b64encode
+
+
+#url = "http://localhost:8000/resize/?url=%s&x=50&y=50" % (quote("http://dajool.com/public/images/dajool_badge.png"), )
+url = "http://localhost:8000/thumbnail/%s/50/50/" % (urlsafe_b64encode("http://dajool.com/public/images/dajool_badge.png"), )
+
+request = requests.get(url)
+print request.content
diff --git a/examples/test.html b/examples/test.html
@@ -0,0 +1,9 @@
+<html>
+<body>
+
+<img src="http://sizr.pelicanbay.de/thumbnail/aHR0cDovL2Rham9vbC5jb20vcHVibGljL2ltYWdlcy9kYWpvb2xfYmFkZ2UucG5n/500/500/" />
+<img src="http://sizr.pelicanbay.de/thumbnail/aHR0cHM6Ly93d3cuZ29vZ2xlLmRlL2ltYWdlcy9zcnByL2xvZ280dy5wbmc=/90/90/" />
+<img src="http://sizr.pelicanbay.de/thumbnail/aHR0cDovL2Rham9vbC5jb20vcHVibGljL2ltYWdlcy9wb3N0cy9tb2JpbGVfc29sdXRpb25fZW5lcmd5LnBuZw==/80/80/" />
+
+</body
+</html>
diff --git a/examples/urlsafe_b64encode.js b/examples/urlsafe_b64encode.js
@@ -0,0 +1,4 @@
+function urlsafe_b64encode(url){
+    var b64url = btoa(url);
+    return b64url.replace(/\+/g, '-').replace(/\//g, '_');
+}
diff --git a/sizr/image_sizr/__init__.py b/sizr/image_sizr/__init__.py
diff --git a/sizr/image_sizr/admin.py b/sizr/image_sizr/admin.py
@@ -0,0 +1,9 @@
+# -*- coding: utf-8 -*-
+
+from django.contrib import admin
+from image_sizr.models import Image
+
+class ImageAdmin(admin.ModelAdmin):
+    pass
+
+admin.site.register(Image, ImageAdmin)
diff --git a/sizr/image_sizr/models.py b/sizr/image_sizr/models.py
@@ -0,0 +1,83 @@
+# -*- coding: utf-8 -*-
+
+__author__ = 'Jochen Breuer <brejoc@gmail.com>'
+__copyright__ = 'Jochen Breuer <brejoc@gmail.com>'
+__license__ = 'BSD 3-Clause License'
+
+import os
+import random
+import string
+import logging
+
+from shutil import rmtree
+
+from django.db import models
+from django.conf import settings
+from django.dispatch import receiver
+from django.db.models.signals import pre_delete
+from django.utils.translation import ugettext_lazy as _
+
+
+# returns the absolute path to this file extended by the parameter
+abs_path = lambda parameter: os.path.join(
+    os.path.dirname(os.path.abspath(__file__)),
+    parameter)
+
+# Get an instance of a logger
+logger = logging.getLogger(__name__)
+
+class Image(models.Model):
+    """\
+    …
+    """
+
+    url_hash = models.CharField(max_length=32, primary_key=True)
+    url = models.URLField()
+    bucket = models.CharField(max_length=2, blank=True, null=True)
+    suffix = models.CharField(max_length=8)
+    # regarding the length of the mime type field:
+    # http://tools.ietf.org/html/rfc4288#section-4.2  |  127 + 1 + 127 = 255
+    mime_type = models.CharField(max_length=255)
+    img_hash = models.CharField(max_length=255,
+                                help_text=_(u"""This is either the etag or a md5sum
+                                generated from last-modified and content-length"""))
+
+    class Meta:
+        verbose_name = _(u"Image")
+        verbose_name_plural = _(u"Images")
+
+    def __unicode__(self):
+        return self.url_hash
+
+    def save(self, *args, **kwargs):
+        """\
+        Randomly choose bucket and save.
+        """
+        model = self.__class__
+        if self.bucket is None or self.bucket == "":
+            self.bucket = ''.join(random.choice(string.digits) for x in range(2))
+        mkdir_p(self.path_to_images())
+        return super(Image, self).save(*args, **kwargs)
+
+    def path_to_images(self):
+        return os.path.join(settings.BUCKET_ROOT, self.bucket, self.url_hash, self.img_hash)
+
+    def path_to_original_image(self):
+        return os.path.join(self.path_to_images(), self.img_hash + '.' + self.suffix)
+
+
+@receiver(pre_delete)
+def del_images(sender, instance, **kwargs):
+    """\
+    Also delete images from Bucket when dataset is deleted.
+    """
+    if sender==Image:
+        try:
+            rmtree(instance.path_to_images())
+        except:
+            logger.error("Couldn't remove bucket: %s" % instance.path_to_images())
+
+
+def mkdir_p(path):
+    if not os.path.isdir(path):
+        os.makedirs(path)
diff --git a/sizr/image_sizr/tests.py b/sizr/image_sizr/tests.py
@@ -0,0 +1,16 @@
+"""
+This file demonstrates writing tests using the unittest module. These will pass
+when you run "manage.py test".
+
+Replace this with more appropriate tests for your application.
+"""
+
+from django.test import TestCase
+
+
+class SimpleTest(TestCase):
+    def test_basic_addition(self):
+        """
+        Tests that 1 + 1 always equals 2.
+        """
+        self.assertEqual(1 + 1, 2)
diff --git a/sizr/image_sizr/views.py b/sizr/image_sizr/views.py
@@ -0,0 +1,168 @@
+# -*- coding: utf-8 -*-
+
+__author__ = 'Jochen Breuer <brejoc@gmail.com>'
+__copyright__ = 'Jochen Breuer <brejoc@gmail.com>'
+__license__ = 'BSD 3-Clause License'
+
+
+import os
+import requests
+import hashlib
+import logging
+
+from base64 import urlsafe_b64decode
+from PIL import Image as Pil
+from StringIO import StringIO
+
+from django.conf import settings
+from django.shortcuts import redirect
+from django.views.decorators.cache import cache_page
+
+from image_sizr.models import Image
+
+# Get an instance of a logger
+logger = logging.getLogger(__name__)
+
+DEBUG = settings.DEBUG
+
+class MissingContentTypeError(Exception):
+    """\
+    We need our own Error, so we can catch it.
+    """
+    pass
+
+class ContentTypeError(Exception):
+    """\
+    We need our own Error, so we can catch it.
+    """
+    pass
+
+
+@cache_page(settings.CACHE_TIME)
+def resize(request, url, x, y):
+    """\
+    Add desciption
+    """
+    # todo: check parameter
+    # todo: also add parameter for quality and file type
+    #url = urlsafe_b64decode(request.GET.get('url', None))
+    url = urlsafe_b64decode(str(url))
+    #new_x = request.GET.get('x', None)
+    #new_y = request.GET.get('y', None)
+    new_x = int(x)
+    new_y = int(y)
+    r = requests.head(url)
+
+    if DEBUG is True:
+        logger.info(r.headers)
+
+    # create url hash and look it up in database
+    url_m = hashlib.md5()
+    url_m.update(url)
+    img_url_hash = url_m.hexdigest()
+
+    if DEBUG is True:
+        logger.info("image url: %s" % url)
+        logger.info("image url hash: %s" % img_url_hash)
+
+    images = Image.objects.filter(url_hash=img_url_hash)
+    db_image = None
+    if images:
+        if len(images) > 1:
+            logger.warning("Found too many images. Deleting all and starting over.")
+            # todo: something is wrong here. there should be only one. delete all and start over
+            pass
+        else:
+            logger.info("Found image.")
+            print("Found image.")
+            db_image = images[0]
+            if not os.path.exists(os.path.join(db_image.path_to_images(), "%s_%sx%s.%s" % \
+                    (db_image.url_hash, new_x, new_y, db_image.suffix))):
+                print "resized image not found"
+                _create_thumb(db_image, new_x, new_y)
+    else:
+        db_image = _fetch_image(url, img_url_hash, r)
+        _create_thumb(db_image, new_x, new_y)
+
+    return redirect("http://localhost:8000/bucket/%(bucket)s/%(url_hash)s/%(img_hash)s/%(url_hash)s_%(x)sx%(y)s.%(suffix)s" % \
+                    {
+                        "bucket": db_image.bucket,
+                        "url_hash": db_image.url_hash,
+                        "img_hash": db_image.img_hash,
+                        "x": new_x,
+                        "y": new_y,
+                        "suffix": db_image.suffix,
+                    })
+
+def _fetch_image(url, img_url_hash, r):
+    """\
+    add description!!!
+    """
+
+    content_types = settings.CONTENT_TYPES.keys()
+
+    if "content-type" not in r.headers:
+        raise MissingContentTypeError("content-type is missing")
+    if r.headers["content-type"] not in content_types:
+        raise ContentTypeError("content-type '%s' is not in list" % \
+            r.headers["content-type"])
+
+
+    # etag or hash
+    img_hash = None
+    if r.headers.has_key("etag"):
+        img_hash = r.headers["etag"]
+        # some servers (like the one play! uses) include quotation marks
+        img_hash = img_hash.strip('"')
+    elif r.headers.has_key("last-modified"):
+        m = hashlib.md5()
+        m.update(r.headers["last-modified"])
+        # also use content-length if it exists
+        if r.headers.has_key("content-length"):
+            m.update(str(r.headers.has_key("content-length")))
+        img_hash = m.hexdigest()
+
+
+    if DEBUG is True:
+        logger.info("loading image…")
+        logger.info("image hash: %s" % img_hash)
+
+    db_image = Image(url=url,
+                     url_hash=img_url_hash,
+                     img_hash=img_hash,
+                     mime_type=r.headers["content-type"],
+                     suffix=settings.CONTENT_TYPES[r.headers["content-type"]])
+    db_image.save()
+
+    r = requests.get(url)
+    image = Pil.open(StringIO(r.content))
+    logger.info(image)
+    image.save(os.path.join(db_image.path_to_images(), img_hash + "." + db_image.suffix),
+               image.format,
+               quality=90)
+    logger.info("saved image to %s" % os.path.join(db_image.path_to_images(), img_hash))
+    return db_image
+
+def _create_thumb(db_image, x, y):
+    """\
+    add description!!!
+    """
+    original_image = db_image.path_to_original_image()
+    resize_filename = os.path.join(db_image.path_to_images(), "%s_%sx%s.%s" % \
+                                            (db_image.url_hash, x, y, db_image.suffix))
+
+    image = Pil.open(original_image)
+    width, height = image.size
+    print "original size: %s:%s" % (width, height)
+
+    if height > y:
+        # portrait images get blurry otherwise
+        print "height > y"
+        y = height
+    new_size = int(x), int(y)
+    image.thumbnail(new_size, Pil.ANTIALIAS)
+    print "new width and height: %sx%s" % (x, y)
+    try:
+        image.save(resize_filename, image.format, quality=90, optimize=1)
+    except:
+        image.save(resize_filename, image.format, quality=90)
diff --git a/sizr/manage.py b/sizr/manage.py
@@ -0,0 +1,10 @@
+#!/usr/bin/env python
+import os
+import sys
+
+if __name__ == "__main__":
+    os.environ.setdefault("DJANGO_SETTINGS_MODULE", "sizr.settings")
+
+    from django.core.management import execute_from_command_line
+
+    execute_from_command_line(sys.argv)
diff --git a/sizr/sizr/__init__.py b/sizr/sizr/__init__.py
diff --git a/sizr/sizr/settings.py b/sizr/sizr/settings.py
diff --git a/sizr/sizr/urls.py b/sizr/sizr/urls.py
diff --git a/sizr/sizr/wsgi.py b/sizr/sizr/wsgi.py
