Abalone Analytics

The Abalone project is a data tracking and analytics system aimed at storing and measuring data for population trends, mortality rates, and breeding programs. Designed as a multi-tenant application, Abalone will initially serve two stakeholders, the Bodega Marine Laboratory at UC Davis and the Puget Sound Restoration Fund in Washington State.

The Bodega Marine Laboratory's White Abalone captive breeding program is working to prevent the extinction of the White Abalone (Haliotis sorenseni), an endangered marine snail. White abalone are one of seven species found in California and are culturally significant to the native people of the area. White abalone were perilously overfished throughout the 20th century, resulting in a 99 percent population decrease by the end of the 1970s. This group is working to reverse their decline and have already seen some great success, they currently have more abalone in the lab than exist in the wild!

The Puget Sound Restoration Fund works to raise and outplant hatchery-reared Pinto Abalone (Haliotis kamtschatkana), the only abalone species found in the Washington waters. This species has cultural and ecological significance, grazing rock surfaces and maintaining the health of rocky reef habitat and kelp beds. The Washington Department of Fish & Wildlife (WDFW) documented a ~98% decline from 1992 to 2017, leading the pinto abalone to be listed as a state endangered species in 2019.

This application will enable groups to add data either through CSV upload or through the web interface. Groups can view reports and visual representations of key data. Future plans include giving groups the ability to generate custom reports on the fly.

Getting Started

Prerequisites

This application is built on following and you must have these installed before you begin:

• Ruby (3.0.3)
• Rails (6.1.4.6)
• PostgreSQL (tested on 9.x)
• Yarn

Setup

After forking this repo and cloning your own copy onto your local machine, execute the following commands in your CLI:

gem install bundler
bundle install
yarn install
bin/webpack
rake db:create
rake db:migrate
rake db:seed

Run Test Suite

bundle exec rake

Run Webserver for Abalone

Webpack dependencies can be rebuilt on command with bin/webpack. Alternatively you can run bin/webpack-dev-server in another terminal window. This will effectively run bin/webpack for you whenever files change.

Then, run bundle exec rails s and browse to http://localhost:3000/.

Login information for white abalone:

Email: admin@whiteabalone.com
Password: password

Login information for pinto abalone:

Email: admin@pintoabalone.com
Password: password

Running Background Jobs

The app uses the gem delayed_job for processing CSVs. To run background jobs, run the following command in your CLI:

rake jobs:work

To confirm background jobs are processing, try uploading a CSV at http://localhost:3000/file_uploads/new. You should see the job complete in your CLI and see the file upload results here at http://localhost:3000/file_uploads.

To see detailed logs from background jobs, run:

tail -f log/delayed_job.log

To clear background jobs, run:

rake jobs:clear

Direct SQL Reporting

This application uses a modified implementation of the Blazer gem to provide direct SQL access with data scoped to an organizational level. This requires some setup to use in your development environment. See the instructions for setting this up locally to get started.

Docker

We are currently experimenting with Docker for development. While we would love for more people to try it out, be forewarned - Docker functionality may not be maintained moving forward. You will need Docker and docker-compose.

• Docker Desktop is recommended for Windows and Mac computers.
• The make utility can also make your development life easier. It is usually already installed on Linux and Mac computers. For Windows, an easy way to install it is via Chocolatey, a software package management system similar to Homebrew for Windows. Once Chocolatey is installed, install make with choco install make in a command prompt running as Administrator.
• If you run into issues using Docker Desktop on windows, we recommend you view this page for troubleshooting info.

Starting Fresh

To start the application in development mode:

• docker-compose up --detach db to start the database
• docker-compose run --rm schema_migrate to bring the database schema up-to-date
• docker-compose up --detach web delayed_job to start the web and background job processes

Or, run only this:

• make minty_fresh to do all of the above

The web app will be available on your host at http://localhost:3000. The logs for the web app and delayed_job processes can be seen and followed with the make watch command.

Some Routine Tasks

• make spec will run the RSpec tests
• make lint will run the Rubocop linting and style checks
• make brakeman will run the Brakeman security vulnerabilities checks
• make test will run spec, lint, brakeman
• make build will build the Docker image for the abalone application. You'll need to run this occasionally if the gem libraries for the project are updated.
• make database_seeds will seed the database according to seeds.rb.
• make nuke will stop all Abalone docker services, remove containers, and delete the development and test databases. This is also used in the make minty_fresh command to restart the development and test environment with a clean slate.

Only the Database

Some developers prefer to run the Ruby and Rails processes directly on their host computers instead of running everything in containers. It might still be convenient for those developers to run the database in a container and not deal with the installation of yet another server on their computer. To do so:

• set an environment variable on your host: export DATABASE_URL="postgres://dockerpg:supersecret@localhost:54321"
• start the database with make database_started

Development

We have included the Annotate gem in this project, for better development experience. It annotates (table attributes) models, model specs, and factories.

The annotate task will run automatically when running migrations. Please see lib/tasks/auto_annotate_models.rake for configuration details.

If it does not run automatically, you can run it manually, on the project root dir, with:

annotate

Check out their Github page for more running options.

Architectural Constraints

In submitting features or bug fixes, please do not add new infrastructure components — e.g. databases, message queues, external caches — that might increase operational hosting costs. We are hosting on a free Heroku instance and need to keep it this way for the foreseeable future. Come talk to us if you have questions by posting in the Ruby for Good #abalone slack channel or creating an issue.

Other Considerations

We want it to be easy to understand and contribute to this app, which means we like comments in our code! We also want to keep the codebase beginner-friendly. Please keep this in mind when you are tempted to refactor that abstraction into an additional abstraction.

Get Familiar with the App

Application Overview

The Problem

Our stakeholders, the Bodega Marine Laboratory and the Puget Sound Restoration Fund work with large amounts of data collected as part of their abalone captive breeding programs. They need a system that can act as a central data repository for all of this data and provide robust reporting capabilities to help them examine trends and combine data collected across their research efforts.

The Solution

We are building a multi-tenant application which has the following capabilities:

1. Store Data: There are several types of measurement data collected that should be stored in the system and retrievable by each organization.
2. Import CSVs: Users are able to import single and bulk CSVs. Users should generally submit cleaned CSVs, but the app should alert users if there are parsing problems and which row(s) need to be fixed.
3. Display Charts and Analytics: Display charts and analytics to meet the reporting needs of each organization. Allow organizations to directly query their data.
4. Export CSVs: TBD.

Key Definitions

• Tag number(s), date = e.g. Green_389 from 3/4/08 to 4/6/15 We sometimes tag individuals; however, not all individuals have tags. We can't tag individuals until they are older than one year old because they are too small. Generally a color, a 3-digit number and dates that tag was on. Sometimes tags fall off. It can be logistically challenging to give them the same tag, so they sometimes get assigned new tags. Also, occasionally tags have another form besides color_### (e.g., they have 2 or 4 digits and/or have no color associated with them), and sometimes they are something crazy like, "no tag" or "no tag red zip tie" for animals that lived long ago ... though I suppose we could re-code those into something more tractable.
• Shellfish Health Lab Case Number (shl number) = SF##-## Animals from each spawning date and from each wild collection have a unique case number created by California's state Shellfish Health Laboratory (SHL). Sometimes animals from a single spawning date have more than one SHL number.
• Cohort = place_YYYY This is how the lab coloquilly refers to each of their populations spawned on a certain date. It's bascially a note/nickname for each group of animals with a particular SHL #/spawning date.
• Enclosures = e.g. Juvenile Rack 1 Column A Trough 3 from 3/4/15 - 6/2/16 This is the tank space by date. This is a note. The types of input will vary significantly within a facility and over time.
• Locations = facility_name - location_name Animals may be located in different location within a single facility
• Facilities = e.g. BML from 6/5/13 - 11/20/14 Animals move around among a finite number of partner institutions (it is possible for new facilities to be added, but it only happens about once every few years).
• Organizations e.g. Bodega Marine Laboratory Organizations act as the tenants within the application for the purpose of walling off data
• MortalityEvents this is a way to track the mortality event of either a specific animal or a cohort. If the mortality event is related to a specific animal, the mortality_count is expected to be nil; if the mortality event is related to a cohort, the mortality_count is expected to be present, but the animal id is expected to be nil.

See a full data dictionary here.

And Don't Forget...

...that Gary needs you.

Photo credit: John Burgess/The Press Democrat