palewire / django-bakery Goto Github PK

A set of helpers for baking your Django site out as flat files

Home Page: https://palewi.re/docs/django-bakery/

License: MIT License

Python 98.83% JavaScript 0.01% CSS 0.09% HTML 1.07%

class-based-views django python static-site static-site-generator

django-bakery's Introduction

                   
                # # # # #
              __#_#_#_#_#__
             {_` ` ` ` ` `_}
            _{_._._._._._._}_
           {_  D J A N G O  _}
          _{_._._._._._._._._}_
         {_    B A K E R Y    _}
     .---{_._._._._._._._._._._}---.
    (   `"""""""""""""""""""""""`   )
     `~~~~~~~~~~~~~~~~~~~~~~~~~~~~~`

A set of helpers for baking your Django site out as flat files

Documentation: palewi.re/docs/django-bakery
Issues: https://github.com/datadesk/django-bakery/issues
Packaging: https://pypi.python.org/pypi/django-bakery
The dream, in Powerpoint form: palewi.re/docs/django-bakery/_static/the-dream.pdf

Features

Models, views and management commands that will build your site as flat files.
Management commands to sync your flat files with a bucket on Amazon S3.
Optional integration of a Celery job queue to automatically build and publish model objects when they are saved

django-bakery's People

Contributors

Stargazers

Watchers

django-bakery's Issues

ImportError: No module named urls

Hi there,

Firstly well done on this awesome app. Loving it.

I've run into a small issue with /bakery/management/commands/buildserver.py line 1 specifies "import urls" which results in ImportError: No module named urls.

The bakery app does not have a urls.py and I was wondering which urls the app was referring to?

all the best
Ross

Refactor sync_s3() command so it actually makes sense to people

The publish command could use a --force option to republish everything, maybe

Add a --config kwarg option for the publish command for custom .s3cfg file paths

should the models have "unbuild" methods?

How can we set the s3 headers for just gzipped files?

Class-based views should not gzip everything, but check against the same whitelist of filetypes as build

Now that we have boto in there, could we thread parallel uploads at the same time?

Something like this

tests for the publish and unpublish commands

Maybe we could set up a test S3 bucket with encrypted credentials?

documentation on what license this is published under is missing?

At a glance, I can't see a LICENSE or mention of the licensing in setup.py, nor is it specified in the LA Times datadesk tutorial, which only mentions it being open source.

Add docs for how the robots.txt and favicon.ico bits work

Automate coverage report with coveralls.io

If we have GZIP_FILE_MATCH in publish, shouldn't we also use it in build for the static files?

cc @emamd

PEP8 and pyflakes compliance and testing

test the publish command, somehow

Do we actually neeeeeed .gz and regular .html/css/js files?

retrofit heatmeter.latimes.com

test_build_cmd should check to see if files actually make it over

build cmd doesn't work with 1.4's new directory structure

six.text_type fails to write some stuff

Error

Traceback (most recent call last):
  File "manage.py", line 14, in <module>
    execute_manager(settings)
  File "/home/ben/Code/datadesk/sites/documents.latimes.com/local/lib/python2.7/site-packages/django/core/management/__init__.py", line 438, in execute_manager
    utility.execute()
  File "/home/ben/Code/datadesk/sites/documents.latimes.com/local/lib/python2.7/site-packages/django/core/management/__init__.py", line 379, in execute
    self.fetch_command(subcommand).run_from_argv(self.argv)
  File "/home/ben/Code/datadesk/sites/documents.latimes.com/local/lib/python2.7/site-packages/django/core/management/base.py", line 191, in run_from_argv
    self.execute(*args, **options.__dict__)
  File "/home/ben/Code/datadesk/sites/documents.latimes.com/local/lib/python2.7/site-packages/django/core/management/base.py", line 220, in execute
    output = self.handle(*args, **options)
  File "/home/ben/Code/datadesk/sites/documents.latimes.com/repo/documents/management/commands/build.py", line 10, in handle
    super(Command, self).handle(*args, **options)
  File "/home/ben/Code/datadesk/sites/documents.latimes.com/src/bakery/bakery/management/commands/build.py", line 109, in handle
    view().build_method()
  File "/home/ben/Code/datadesk/sites/documents.latimes.com/src/bakery/bakery/views.py", line 89, in build_queryset
    outfile.write(six.text_type(html))
UnicodeDecodeError: 'ascii' codec can't decode byte 0xe2 in position 75580: ordinal not in range(128)

Add docs for some of the new options

can we make gzipping optional?

I imagine this would be an option on the build management command.

Figure out how we can compress static files

--dry-run option on publish command

Gzipping based on mimetype, not file extension

autoinspect for robots.txt and favicon.ico and if they exist copy them to the root of the BUILD_DIR

During a GZIP'd publish, the second "sync_all" command should be less greedy and only attempt to sync those files that DO NOT match the gzipping file pattern

Can we refactor the publish routine to use boto, instead of s3cmd?

it'd probably give us more control, and a possible solution to why content-encoding: gzip headers aren't always being appended.

Crashes when you try to bake something with a Spanish language accent

Review when we print and went we log in build command

models should use get_callable instead of _view_from_string

Could the class-based views somehow automagically build out a gzipped version of each file?

Do we want to add Cache-Control headers on our pages?

Needs further research, but this could be another big speed and cost improvement. Would need to make sure there are ways to invalidate the cache, and would want a relatively short one for news pages.

I imagine some static resources - images, css, js, could have farther future headers though.

What if you are building your site to be hosted in a subdirectory on a CDN? How do you prefix the URLs during builds?

This question came up this morning during a Twitter conversation with @aplayford.

Here's an initial idea I have. What if:

Bakery was expanded to include a set of context processors that would have to be added to the user's ``settings.py.

TEMPLATE_CONTEXT_PROCESSORS = (
 #... all the usual crap
 'bakery.context_processors.base_url',
 'bakery.context_processors.static_url',
)

Which would create new variables that could be used in each template that looked something like:

<html>
<head>
  <script src="{{ STATIC_URL }}jquery.js"></script>
<body>
  <a href="{{ BASE_URL }}{% url 'homepage' %}">Back to the homepage</a>
</body>
</html>

Then when the template was rendered, the context_processors would analyze the request object they receive and be able to substitute in different values in the baked files by drawing on values in settings.py like:

BUILD_BASE_URL = 'http://www.foobar.com/subdirectory/'
BUILD_STATIC_URL = 'http://www.foobar.com/static/subdirectory/'

I'm not sure how the context processor would know when to make the switch but I suspect it could work something like this one except it would sniff the request to verify if it came from the RequestFactory we use in our buildable views. If that doesn't already include a reliable flag to execute that test, we could expand the views to plant something on the request or in the context dictionary.

Crazy?

I think the DetailView should also _build_extra

https://github.com/datadesk/django-bakery/blob/master/bakery/views.py#L152

Can we add an option to only publish a subdirectory?

six.print_() switches to logger calls

Is there a way to abstract back the task selection code in our custom save override?

It would be nice if we didn't have to do all this every time we implement a Celery based deployment.

Not sure how this happened, but it's bad

cc @emamd

I don't think buildserver works anymore

Gzip flag should direct to either write() or gzip() in the views.py

Should be some way to selectively rebuild/publish views

It can take a long time to rebuild some views, and often we know that only one view has changed. However, in order to publish just the one changed view, we still have to rebuild everything, since the build command blows away the build directory and the publish command deletes anything from the target bucket that's not currently in the build directory.

Either I'm missing something straightforward, or there needs to be some mechanism by which we can tell bakery to only build/sync a specific view/directory.

palewire / django-bakery Goto Github PK

django-bakery's Introduction

Features

django-bakery's People

Contributors

Stargazers

Watchers

Forkers

django-bakery's Issues

Recommend Projects

Recommend Topics

Recommend Org