gluten - Transcript Annotator 2.0

A Flask-based Python 3 annotation application.

Important Legacy note

We maintain a copy of the "legacy" Java application that predated this current project in the folder legacy. However it is NOT maintained. If you are interested in a Java web application using file-based storage for transcripts, this might be for you. Let us know and if interest is sufficient we might consider spinning out a separate project. Otherwise use at your own peril. See also this PR for why you might want to proceed with caution: memphis-iis#1

Getting started

If you are running on a Ubuntu-like environment, you need to make sure you've installed libyaml-dev. If you have, you can run setup.sh, provide a test.config file.

$ sudo apt-get install libyaml-dev
$ git clone https://github.com/memphis-iis/gluten
$ cd gluten
$ ./setup.sh

After that you can run a local debug server with ./local.sh.

If you aren't running in a Mac/Linux type environment, you can use the bundled Vagrantfile (assuming that you have vagrant and VirtualBox installed). In fact, you can use the Vagrantfile even if you're running on Linux. That way you don't need to insure you have all the correct dependencies:

$ vagrant up
$ vagrant ssh

To run the application locally, you'll need to set up a virtualenv with the correct dependencies. Assuming you're SSH'ed into the VM (using vagrant ssh as above):

$ cd gluten
$ ./setup.sh

Then you need to create a test config file:

$ touch test.config
$ nano test.config

At a minimum, you need to supply the Google OAuth credentials that allow users to log in. A sample test.config file would look something like this:

DEBUG = 1
TEST_EMAIL = '[email protected]'
FLASK_SECRET = 'Some big random string'
GOOGLE_OAUTH_CLIENT_ID = 'get this from google'
GOOGLE_OAUTH_CLIENT_SECRET = 'get this from google'

You'll want to use your gmail address as the TEST_EMAIL so that when you login, a few test transcripts will be waiting for you (they're created every time you start the application). FLASK_SECRET should be a pretty large random string. GOOGLE_OAUTH_CLIENT_ID and GOOGLE_OAUTH_CLIENT_SECRET should be available at https://console.developers.google.com/. Be sure to include http://localhost:5000 under the authorized JavaScript origins, and http://localhost:5000/auth/authorized under the authorized redirect URL's.

Using Vagrant

We provide a Vagrantfile and provisioning scripts for testing/debugging gluten locally. Here are some things to keep in mind...

IMPORTANT - The helper scripts in gluten try to pollute your workstation, so all Python virtualenv's are created in the directory tree at well-known places (which are in the .gitignore file so that they aren't checked in to source code control). This combined with Python's pycache directories and .pyc files means that you should NOT run gluten both with and without Vagrant on local workstation. Choose one or the other. Or if you know what you're doing clean out the virtualenv's and Python bytecode files before you switch. You have been warned.

If you are familiar with vagrant, you need to know that we create a link from /vagrant to ~/gluten. If you are NOT familiar with vagrant, you create our VM, log into it, and then access the code like this:

$ vagrant up
$ vagrant ssh
$ cd gluten

READ the README.md file!!! You need to supply a test.config file before you can really test things out.

You can use your favorite code editor and version control application in the host operating system - you can just use this little login to test, start, or stop the application.

First things first: log in and set up the test environment

$ vagrant ssh $ cd gluten $ ./setup.sh

To run the server in development mode:

$ vagrant ssh
$ cd gluten
$ ./local.sh

To run unit tests:

$ vagrant ssh
$ cd gluten
$ ./run_tests.sh

Connect to gluten from your host operating system at:

http://localhost:5000/

If for some reason you wanted to run mock AWS services in the background (e.g. better simulate AWS testing):

$ vagrant ssh $ cd gluten $ test/local_test_services.sh

Deploying to AWS Elastic Beanstalk

The deployment is fairly easy:

• Make sure your have a prod.config file created with at least these variables defined:
  • DEBUG (should be 0)
  • FLASK_SECRET (use a really good random string)
  • GOOGLE_OAUTH_CLIENT_ID (you get this from Google)
  • GOOGLE_OAUTH_CLIENT_SECRET (you get this from Google)
• Use ./make_aws_upload.sh to create the ZIP'ed application bundle
• When you create your AWS environment/application, be sure to select:
  • The Python application type
  • A single instance web application (instead of the default load balanced)
  • Use t2.micro instead of t1.micro (t1.micro is the default, but t2.micro is cheaper and has more resources)
  • Set the GLUTEN_CONFIG_FILE environment variable to "prod.config"
• Once everything is running, don't forget to add the new URL to the authorized list under your Google Developer's console. If you just use a default URL from EB, it will look something like: http://my-cool-env-name.elasticbeanstalk.com/auth/authorized