workforwardnola

Job seeker focused website for NOLA. There is a version live at http://careerpathnola.com (hosted by the City of New Orleans).

Dev setup

Requires Ruby 2.5.3 and Postgres (9.5+ preferred). You may need to use a differing version of Ruby, depending on deploy method (see the AWS section below)

After cloning the repository for the first time, run bundle install in the workforwardnola directory.

Copy .env.example to a new file called .env. You can leave the admin password as-is for development, or change it.

Database Setup

Make sure Postgres is installed: which psql should give a reasonable answer. Also create a file called .env with the line DATABASE_URL=postgres://localhost:5432/workforwardnola. (You may need to create a new DB in postgres called workforwardnola.)

Run rake db:migrate and rake db:seed to set up the database structure and fill it with sample data. If the data doesn't show up, try running rake db:reset and rake db:seed. Career data can also be loaded via http://localhost:9292/manage. and a spreadsheet.

Running app & deployment

Run the app by running bin/start, with an optional command line argument "$IP:$PORT", all it does is call rerun -p "**/*.{rb,js,scss,mustache,ru,jpg,jpeg,png}" rackup. The site will be available at http://localhost:9292. If there are errors when you try to refresh to see changes, try again - you may have been faster than the app regenerated.

The Code for America NOLA fellows team is no longer actively developing this. We had set up a deployment pipeline via heroku, syncing the staging site with the master branch and enabling review apps for pull requests. It worked well!

The "Email to yourself" career assessment feature requires the EMAIL_xxxx config variables to be set. We use the pony gem to send emails, please see their documentation for more details.

Some changes have included various AWS services. In addition to SMTP from the pony gem, several config variables must be set for AWS services (S3 and SES). These are new additions added below the EMAIL_xxxx config variables.

Setting up SES for Job System: SENDER_EMAIL, OWNER_EMAIL, AWS_ACCESS, AWS_SECRET must all be configured in .env for the SES to work through the job form.

Setting up S3: Configure AWS_BUCKET in .env

Job System Results

There is additionally an optional feature to write to GoogleSsheets using a file client_secret.json. An example is included in client_secret.json.example. If you do not wish to use this feature, do not create or use client_secret.json! You can find information on the google-drive-ruby GitHub. This feature requires setting up a Google Service Account. You can find more information in Google's documentation

Updating content

For details on updating content see other files under the docs/ folder. Details on updating career info via spreadsheet specifically is in docs/career_assessment_how_to.md. Updating Opportunity Centers is now done through the Management page instead of changing through text.

Deploying to production

The ADMIN_xxxx and EMAIL_xxxx config variables need to be set regardless of the deployment. ADMIN_USER and ADMIN_PASSWORD protect the /manage part of the site, and the EMAIL_xxxx variables are for the pony gem. After the app is up, make sure to load the career data via http://your_url_here/manage.

Redeploying for another city

Most of the content is specific to New Orleans and the Opportunity Centers, including which careers are highlighted.

Heroku

Setting up a Heroku pipeline is relatively straightforward. We set up a pipeline with a staging app (with automatic deploys from the master branch), production app (we periodically promote the app from staging to production), and review apps enabled. We used the Postgres add-on.

AWS

We are not AWS experts, so if you have recommendations to improve the following, please make a PR! We found that this requires at least a t2.micro instance to avoid out of memory errors during deployment.

1. Create an IAM user (as recommended by Amazon) with appropriate permissions for deployment credentials. @antislice has no idea what specific permissions are needed for EB deployment, so we tested with admin.
2. Install the Elastic Beanstalk client. This is only be necessary if you are deploying to AWS from the command line. You can do this from the AWS Elastic Beanstalk UI.

• NOTE: When setting up authentication for the CLI tools (so that you can later run eb init or eb deploy), you'll want a the user with deployment credentials -- more permissions than the service user that you'll use to run the service itself (which will only need access to the S3 bucket, elastic beanstalk components, and RDS instance).

3. Follow the Create an Application steps in this documentation.

• For steps 5/6, select Ruby and Ruby 2.5 (Puma)
• NOTE: AWS config only works on specific versions of Ruby. We went with 2.5.3.
• SSH login is optional, but convenient

4. At this point, you'll want to set up the DB. We created an integrated Postgres database instance (v. 9.5.2) as described in here.

• You will need to configure Amazon RDS. Create a database and use the information from the DB to connect (this is important to do first before you upload and deploy)
• An AWS RDS database will follow the same Database URL as the given format in .ENV as if you were running Postgres locally. You can find the necessary credentials in the RDS dashboard for your database instance.
• When you are using AWS RDS, you might have to connect from outside to run migrations. To do this, go to your database instance, add a new inbound security rule for "Anywhere" IP, and then run your migrations locally/remotely, then remove the inbound security role.

5. Create an S3 Bucket for use in storing uploaded Resume files from the Job System form. Copying bucket policies from the Elastic Beanstalk auto-generated bucket works fine from what we've seen, or you can set specific permissions as needed.
6. Walk through Create an Environment
7. Create a source bundle if uploading through the UI. zip ../filename.zip -r * .[^.]*
8. ‼️ At this point, stop and check on the instance type. You may need to configure a VPC.
9. Try deploying: eb deploy or use the Upload and Deploy feature integrated into EB.

• If the app is not working, try using a different Ruby version (e.g. 2.5.1). Note that this would mean testing to ensure all the Ruby gems work, especially if downgrading.

Configuring the "email to yourself" feature

After someone completes the career assessment and sees potential careers, they can use a function through the Pony gem to email the results to themselves. We recommend setting up SES SMTP through AWS in addition to SES API for the Job System emails (a central hub for all email sending!). You can find more info on their docs for SMTP EMAIL_SERVER, EMAIL_DOMAIN, EMAIL_PORT, EMAIL_USER, and EMAIL_PASSWORD are all used for SMTP purposes for the Pony gem. A description of how the SMTP works can be found in app.rb, line 195, or a brief description in .env.example

Further notes on our initial test deployment and changes that were made:

• PR #108 contains the code and configuration changes we made to the app to get it to work with AWS (Elastic Beanstalk)
• Some of the process is described in #106