Open Recipes is an open database of recipe bookmarks.

Our goals are simple:

1. Help publishers make their recipes as discoverable and consumable (get it?) as possible.
2. Prevent good recipes from disappearing when a publisher goes away.

That's pretty much it. We're not trying to save the world. We're just trying to save some recipes.

The recipes in Open Recipes do not include preparation instructions. This is why we like to think of Open Recipes as a database of recipe bookmarks. We think this database should provide everything you need to find a great recipe, but not everything you need to prepare a great recipe. For preparation instructions, please link to the source.

Regular snapshots of the database will be provided as JSON. The format will mirror the schema.org Recipe format. We've posted an example dump of data so you can get a feel for it.

We're not a bunch of chefs. We're not even good cooks.

When we read about the acquisition shutdown of Punchfork, we just shook our heads. It was the same ol' story:

We're excited to share the news that we're gonna be rich! To celebrate, we're shutting down the site and taking all your data down with it. So long, suckers!

This part of the story isn't unique, but it continues. When one of our Studiomates spoke up about her disappointment, we listened. Then, we acted. What happens next surprised us. The CEO of Punchfork took issue with our good deed and demanded that we not save any data, even the data (likes) of users who asked us to save their data.

Here's the thing. None of the recipes belonged to Punchfork. They were scraped from various publishers to begin with. But, we don't wanna ruffle any feathers, so we're starting over.

Use the force; seek the source?

Wanna help? Fantastic. We knew we liked you.

We're gonna be using the wiki to help organize this effort. Right now, there are two simple ways to help:

1. Add a publisher. We wanna have the most complete list of recipe publishers. This is the easiest way to contribute. Please also add an issue and tag it publisher. If you don't have a github account you can also email us suggestions at openrecipes@fictivekin.com
2. Claim a publisher.

Claiming a publisher means you are taking responsibility for writing a simple parser for the recipes from this particular publisher. Our tech (see below) will store this in an object type based on the schema.org Recipe format, and can convert it into other formats for easy storage and discovery.

Each publisher is a GitHub issue, so you can claim a publisher by claiming an issue. Just like a bug, and just as delicious. Just leave a comment on the issue claiming it, and it's all yours.

When you have a working parser (what we call "spiders" below), you contribute it to this project by submitting a Github pull request. We'll use it to periodically bring recipe data into our database. The database will be available intially as data dumps.

To gather data for Open Recipes, we are building spiders based on Scrapy, a web scraping framework written in Python. We are using Scrapy v0.16 at the moment. To contribute spiders for sites, you should have basic familiarity with:

• Python
• Git
• HTML and/or XML

Note: this is strongly biased towards OS X. Feel free to contribute instructions for other operating systems.

To get things going, you will need the following tools:

1. Python 2.7 (including headers)
2. Git
3. pip
4. virtualenv

You will probably already have the first two, although you may need to install Python headers on Linux with something like apt-get install python-dev.

If you don't have pip, follow the installation instructions in the pip docs. Then you can install virtualenv using pip.

Once you have pip and virtualenv, you can clone our repo and install requirements with the following steps:

1. Open a terminal and cd to the directory that will contain your repo clone. For these instructions, we'll assume you cd ~/src.

2. git clone https://github.com/fictivekin/openrecipes.git to clone the repo. This will make a ~/src/openrecipes directory that contains your local repo.

3. cd ./openrecipes to move into the newly-cloned repo.

4. virtualenv --no-site-packages venv to create a Python virtual environment inside ~/src/openrecipes/venv.

5. source venv/bin/activate to activate your new Python virtual environment.

6. pip install -r requirements.txt to install the required Python libraries, including Scrapy.

7. scrapy -h to confirm that the scrapy command was installed. You should get a dump of the help docs.

8. cd scrapy_proj/openrecipes to move into the Scrapy project directory

9. cp settings.py.default settings.py to set up a working settings module for the project

10. scrapy crawl thepioneerwoman.feed to test the feed spider written for thepioneerwoman.com. You should get output like the following:

    2013-03-30 14:35:37-0400 [scrapy] INFO: Scrapy 0.16.4 started (bot: openrecipes)
    2013-03-30 14:35:37-0400 [scrapy] DEBUG: Enabled extensions: LogStats, TelnetConsole, CloseSpider, WebService, CoreStats, SpiderState
    2013-03-30 14:35:37-0400 [scrapy] DEBUG: Enabled downloader middlewares: HttpAuthMiddleware, DownloadTimeoutMiddleware, UserAgentMiddleware, RetryMiddleware, DefaultHeadersMiddleware, RedirectMiddleware, CookiesMiddleware, HttpCompressionMiddleware, ChunkedTransferMiddleware, DownloaderStats
    2013-03-30 14:35:37-0400 [scrapy] DEBUG: Enabled spider middlewares: HttpErrorMiddleware, OffsiteMiddleware, RefererMiddleware, UrlLengthMiddleware, DepthMiddleware
    2013-03-30 14:35:37-0400 [scrapy] DEBUG: Enabled item pipelines: MakestringsPipeline, DuplicaterecipePipeline
    2013-03-30 14:35:37-0400 [thepioneerwoman.feed] INFO: Spider opened
    2013-03-30 14:35:37-0400 [thepioneerwoman.feed] INFO: Crawled 0 pages (at 0 pages/min), scraped 0 items (at 0 items/min)
    2013-03-30 14:35:37-0400 [scrapy] DEBUG: Telnet console listening on 0.0.0.0:6023
    2013-03-30 14:35:37-0400 [scrapy] DEBUG: Web service listening on 0.0.0.0:6080
    2013-03-30 14:35:38-0400 [thepioneerwoman.feed] DEBUG: Crawled (200)  (referer: None)
    2013-03-30 14:35:38-0400 [thepioneerwoman.feed] DEBUG: Crawled (200)  (referer: http://feeds.feedburner.com/pwcooks)
    ...
    

    If you do, baby you got a stew going!

For now, we recommend looking at the following spider definitions to get a feel for writing them:

• spiders/thepioneerwoman_spider.py
• spiders/thepioneerwoman_feedspider.py

Both files are extensively documented, and should give you an idea of what's involved. If you have questions, check the Feedback section and hit us up.

To generate your own spider, use the included generate.py program. From the scrapy_proj directory, run the following (make sure you are in the correct virtualenv:

python generate.py SPIDER_NAME START_URL

This will generate a basic spider for you named SPIDER_NAME that starts crawling at START_URL. All that remains for you to do is to fill in the correct info for scraping the name, image, etc. See `python generate.py --help' for other command line options.

We'll use the "fork & pull" development model for collaboration, so if you plan to contribute, make sure to fork your own repo off of ours. Then you can send us a pull request when you have something to contribute. Please follow "PEP 8 - Style Guide for Python Code" for code you write.

We're just trying to do the right thing, so we value your feedback as we go. You can ping Ed, Chris, Andreas, or anyone from Fictive Kin. General suggestions and feedback to openrecipes@fictivekin.com are welcome, too.

We're also gonna be on IRC, so please feel free to join us if you have any questions or comments. We'll be hanging out in #openrecipes on Freenode. See you there!