Build Process & Deployment

Relevant source files

The following files were used as context for generating this wiki page:

• buildClient.sh
• package.json
• public/js/p3.profile.js
• public/js/release.profile.js
• CONTRIBUTING-old.md

The build process for the DXKB web application is centered around a JavaScript-heavy client-side architecture using the Dojo Toolkit. It utilizes a shell-based build script to orchestrate the compilation, optimization, and packaging of frontend assets, transforming source modules into a production-ready distribution. Deployment is managed via a branch-based strategy where different Git branches correspond to specific environments like production, beta, and internal preview.

Sources: buildClient.sh, CONTRIBUTING-old.md:31-40

Client Build Orchestration

The application uses buildClient.sh to automate the creation of optimized JavaScript bundles. This process involves the Dojo build system, which reads "profiles" to determine how to group modules, minify code, and handle dependencies.

Build Pipeline Flow

The build script performs several sequential steps to ensure a clean and optimized environment for the frontend assets.

This diagram illustrates the automated sequence from directory cleanup to the generation of optimized layers.
Sources: buildClient.sh:1-10

Build Script Configuration

The buildClient.sh script invokes the Dojo build tool with specific flags to control the output.

Flag/Parameter	Description
load=build	Loads the Dojo build system entry point.
profile=p3.profile.js	Defines the specific layers and modules for the DXKB application.
releaseDir	Specifies the target directory for compiled assets (public/js/release).
action=release	Executes the full build and optimization cycle.

Sources: buildClient.sh:10

Build Profiles and Optimization

The build process relies on Dojo Build Profiles to define how the application is structured for production. These profiles determine "layers"—bundles of multiple JavaScript files merged into a single file to reduce HTTP requests.

Layering and Minimization

Profiles like p3.profile.js and release.profile.js define the entry points for the application. The build system processes these to:

1. Resolve Dependencies: Trace all define and require calls to include necessary code.
2. Code Minimization: Strips whitespace and renames variables to reduce file size.
3. Dead Code Removal: Filters out modules that are not explicitly required by the defined layers.

Sources: public/js/p3.profile.js, public/js/release.profile.js

Branching and Environment Deployment

The project follows a structured branching model where specific branches are mapped to deployment URLs. This allows for concurrent development, internal testing, and stable production releases.

The relationship between Git branches and target deployment environments.
Sources: CONTRIBUTING-old.md:31-45

Environment Mapping

Branch	Environment	Deployment URL	Purpose
master	Production	https://www.bv-brc.org	Stable production releases.
develop	Beta	https://beta.bv-brc.org	Release-ready commits and bug fixes.
preview	Alpha	https://alpha.bv-brc.org	Work-in-progress and internal feature reviews.

Sources: CONTRIBUTING-old.md:31-40

Release Preparation

Preparing a release involves incrementing the application version and committing the final build artifacts. The version number in package.json serves as the primary identifier for the deployment.

Release Workflow

1. Version Increment: Update the version field in package.json using semantic versioning.
2. Build Execution: Run the build scripts to generate updated assets in the release folder.
3. Commit and Push: Stage all changes, including the updated package.json and build artifacts, and push to the develop or master branch.

# Example release commands from documentation
edit ./package.json   # increment "version"
git add --all
git commit -m 'release js'
git push

Sources: CONTRIBUTING-old.md:71-80, package.json:3

Deployment Infrastructure

The application is a Node.js-based web server (p3-web) that serves the compiled frontend assets. The package.json defines the entry point and necessary scripts for starting the service.

Server Execution

• Start Script: node ./bin/p3-web initializes the server.
• Dependencies: The server relies on express for routing and various clients like bvbrc_js_client for data interaction.
• Post-Installation: A postinstall script (./post_install.sh) is automatically executed after dependency installation to perform environment-specific setup.

Sources: package.json:5-13

Conclusion

The build and deployment process for DXKB-web ensures that complex frontend dependencies are efficiently bundled and optimized through the Dojo Build System. By maintaining a strict mapping between Git branches and deployment environments (Alpha, Beta, Production), the project facilitates a stable release lifecycle while allowing for continuous integration of new features.