SignLab

Introduction

SignLab is a tool for tagging short video clips and images with various information depending on the user's needs. The target use case is for a research lab to have the ability to upload videos for many lab researchers to assist in tagging. Those tags can then be exported as a CSV for further analysis.

⚠️ Further documentation on usage to come

Code Organization

The code base is made up of an Angular front end (stored in client/ folder) and a NestJS backend (stored in server/ folder). A Dockerfile and Docker compose file is provided for packing up the application, an official docker image is also stored in DockerHub. A break down of the top level folders is shown below.

• client/: Angular client side code
• cypress/: E2E Testing
• docs/: Sphinx documentation (to be completed on a later date)
• server/: NestJS server side code
• shared/: Code shared between client and server, currently only DTOs

The Angular front end is heavily coupled to the NestJS backend. Most of the logic is contained within the backend with the Angular front end acting just as an interface to the backend. The system works by having the NestJS backend serve the compiled Angular front end as a static webpage.

Environment Variables

The SignLab application can be configured via environment variables. Those environment variables can be supplied in a number of different ways.

For a complete list of configrations used by the server, refer to the fileserver/src/config/configuration.ts. That file has the SignLab wide settings and where they come from.

Locally

Locally, the easiest way to pass in the configuration is via .env files. Below is an example of the usage.

cd server/
export NODE_ENV=local
npm run start:watch

The file .env.local in the top level will then be used for configruation. The server looks for a file .env.<NODE_ENV> so additional environments can be made based on testing needs.

By running npm run start:watch at the top level, the application will use the.env.local file.

Docker

The provided docker-compose file looks for a deployment.env file. The deployment.env file is setup identically to a .env and is the preferred method to insert configuration into the dockerized instance.

Object Storage ("Bucket Storage")

SignLab makes use of object storage for storing the response videos that are uploaded to the platform. Currently two platforms are supported.

First is an interface for storing the responses in a local folder. This is handy for testing locally quickly without the need to a dedicated bucket setup.

Second is S3 compliant object storage endpoints. NERC's object storage is S3 compliant so can be used as well.

Each object storage has its own set of required configuration, you can refer to the bucket section of the configuration contained in server/src/config/configuration.ts. The user selects which bucket storage type they are using via the type value (which is evaluated from the environment variable BUCKET_TYPE. Setting BUCKET_TYPE to S3 will have SignLab look for S3 specific settings and use an S3 object container. Using LOCAL as the BUCKET_TYPE will result in a local folder being user for the object container.

GCP is not yet offically supported, but the functionality is stubbed out.

Running the Code for Development

The following steps will get you setup so you can test SignLab locally and test out changes in real time.

1. (One time) Install dependencies and Generate GraphQL

npm i
npm run introspection

2. Setup Object Storage

For local testing, a folder can be used to emulate an actual object storage container. By default the provided .env.local file expects a folder server/bucket/Responses to exist. Make sure that folder exists when testing locally using a folder for object storage emulation.

3. (Each time) Start up a MongoDB instance using the mongod command. Below is an example.

mongod --dbpath ~/data/signlab

4. Start up the frontend and backend

npm run start:watch

The command above will build the Angular and NestJS components and watch for changes. The commands are run with the "concurrent" package so if reading the output is difficult, you can run the same command separately for the server and client package in different terminals from the respective folders.

Note: For the server, you need to specify an env file to load. The command above will run the server with NODE_ENV set to local.