Add .devcontainer to template to facilitate using the repo in codespaces

[cmungall]

Something like this:

{
  "name": "odkfull-codespace",
  // consider also: Codespaces Prebuilds
  "image": "obolibrary/odkfull:latest",

  // Mount your repo where ODK expects it
  "workspaceMount": "source=${localWorkspaceFolder},target=/work,type=bind,consistency=cached",
  "workspaceFolder": "/work",

  // Keep VS Code in control of the container command (safer than using image ENTRYPOINT)
  "overrideCommand": true,

  // Be root so you can add small tools if you need them
  "remoteUser": "root",    
  //"features": {
  //  "ghcr.io/devcontainers/features/node:1": { "version": "lts" }
  //},
  "extensions": [
      "anthropic.claude-code",
      "ms-python.python",
      "ms-azuretools.vscode-docker"
  ],
  "remoteEnv": {
    "TEST_ENV": "123"
  },
  "postCreateCommand": "bash .devcontainer/post_create.sh"
}

[github-actions]

Issue Summary

This enhancement request proposes adding a .devcontainer configuration to the copier template to facilitate development in GitHub Codespaces and VS Code Dev Containers.

Context

The github-ai-integrations repository is a copier template that applies AI integrations (GitHub Actions workflows for Claude, Dragon AI, etc.) to existing projects. Per the CLAUDE.md:

• Template files are in template/ (copied to target repos)
• This repo's own files are at the root level
• The repo "eats its own dogfood" by applying the template to itself

Current State

The template currently includes:

• GitHub Actions workflows (template/.github/workflows/)
• AI agent configuration (template/AGENTS.md.jinja)
• Setup utilities (template/ai.just)
• No devcontainer configuration

Proposed Enhancement

Add a .devcontainer/devcontainer.json to the template that would provide users with:

1. Pre-configured development environment for repos using this template
2. ODK (Ontology Development Kit) integration via the obolibrary/odkfull:latest Docker image
3. Essential VS Code extensions pre-installed:
   • anthropic.claude-code - For Claude Code integration
   • ms-python.python - For Python development
   • ms-azuretools.vscode-docker - For Docker management
4. Post-creation setup via post_create.sh script

Technical Considerations

The provided example configuration includes some ODK-specific settings:

"workspaceMount": "source=${localWorkspaceFolder},target=/work,type=bind,consistency=cached",
"workspaceFolder": "/work",

Recommendation: This should be made optional/conditional based on the copier questions, since:

• is_ontology_project question already exists in copier.yaml
• Not all users need ODK (only those working with ontologies)
• Python-only and LinkML-only projects may need different base images

Implementation Suggestions

1. Add copier question: use_devcontainer (bool, default: true)
2. Conditional base image based on project type:
   • Ontology projects → obolibrary/odkfull:latest
   • Python projects → Standard Python devcontainer image
   • LinkML projects → LinkML-specific image or Python base
3. Template the configuration using Jinja2 conditionals
4. Include post-create script that runs just -f ai.just setup-ai automatically

Related Best Practices

Dev Container configurations align well with the AI-integration goals:

• Enables consistent development environments across contributors
• Facilitates onboarding new contributors
• Supports Codespaces for cloud-based development
• Pre-configures Claude Code extension for immediate AI assistance

Benefits

• Lower barrier to entry for contributors
• Consistent environments across team members
• Cloud development ready via GitHub Codespaces
• AI tooling pre-configured (Claude Code extension included)

Next Steps

Would you like me to create an implementation PR that:

1. Adds devcontainer configuration to the template
2. Makes it conditional based on project type (ontology/python/linkml)
3. Includes appropriate base images for each project type?

Or would you prefer to assign this to the dragon-ai-agent for implementation?

[cmungall]

@dragon-ai-agent please do it.

[cmungall]

@claude you take over

[claude]

Claude finished @cmungall's task —— View job • claude/issue-15-20251113-0049 • Create PR ➔

---

Implementing .devcontainer for template ✅

Taking over from dragon-ai-agent to implement devcontainer configuration for the copier template.

Todo List:

• Analyze current template structure and copier configuration
• Create conditional .devcontainer/devcontainer.json template
• Create post-create script template
• Update copier questions if needed
• Test and commit changes

Implementation Summary:

✅ Created conditional devcontainer configuration based on project type:

• Ontology projects: obolibrary/odkfull:latest image with ODK-specific workspace mount to /work
• Python projects: mcr.microsoft.com/devcontainers/python:3.11 image
• Other projects: Base Ubuntu image with Node.js feature

✅ Added automated post-create setup script that:

• Installs just command runner using appropriate method for each container type
• Sets up git configuration for new users
• Installs Python dependencies (requirements.txt or pyproject.toml)
• Configures LinkML tools for schema projects
• Automatically runs just -f ai.just setup-ai if GitHub token is available

✅ Enhanced copier configuration:

• Added include_devcontainer question (default: true) for optional inclusion
• Maintains compatibility with existing project types

✅ Pre-configured essential VS Code extensions:

• anthropic.claude-code - Claude Code integration
• ms-python.python + ms-python.pylint - Python development (conditional)
• redhat.vscode-yaml - YAML support for LinkML schemas (conditional)
• ms-azuretools.vscode-docker - Docker management
• GitHub.copilot - AI pair programming

Technical Features:

• Conditional templating: Only generates files when include_devcontainer: true
• Smart base image selection: Automatically chooses appropriate container based on project flags
• Environment variable passthrough: GitHub token from local environment
• Project-specific setup: Different installation paths for ODK vs standard containers
• Graceful fallbacks: Setup continues even if AI integration fails

Benefits:

• One-click development setup via GitHub Codespaces
• Consistent environments across team members
• Pre-configured AI tooling with Claude Code extension
• Automatic dependency installation for Python/LinkML projects
• ODK integration for ontology development workflows

---