DHTI CLI

DHTI CLI

The DHTI CLI (dhti-cli) is your command-line swiss-army knife for composing, installing, developing, packaging, and managing DHTI modules.

Installation

You don't need to install globally. Use npx so you always run the latest published version:

npx dhti-cli help

Core Concepts

• Compose: Build a docker-compose file with selected modules.
• Elixir: A LangServe app providing GenAI functionality.
• Conch: An OpenMRS O3 ESM (or CDS Hooks client) providing UI & interaction.
• Bootstrap: Runtime configuration (models, hyperparameters, etc.) injected without rebuilding.

Commands Overview

For detailed usage, see the specific pages:

• Compose: dhti-cli compose
• Synthea: dhti-cli synthea

Compose Commands

• npx dhti-cli compose add -m openmrs -m langserve [ ... ] Adds modules to ~/dhti/docker-compose.yml.
• npx dhti-cli compose read Displays the generated compose file.

Elixir Commands

• npx dhti-cli elixir install -g <git_url> -n <name> Install an elixir from Git.
• npx dhti-cli elixir dev -d <path_to_elixir_src> -n <name> -c <container> Sync code (hot-reload like) into running container.
• npx dhti-cli elixir init Initialize new project.

Conch Commands

• npx dhti-cli conch install -g <git_url> -n <name> Install a conch (OpenMRS O3 ESM or CDS client).
• npx dhti-cli conch dev -d <path_to_conch_dist> -n <name> -c <container> Copy built dist into running frontend container.
• npx dhti-cli conch start -n <name> Start a conch.

Docker Commands

• npx dhti-cli docker -u Start (up) all services in compose.
• npx dhti-cli docker -d Stop & remove containers.
• npx dhti-cli docker -n <image_name> -t elixir|conch Build an image from installed module.
• npx dhti-cli docker bootstrap -f <bootstrap.py> -c <container> Inject new bootstrap config.

Other Commands

• docktor: Manage ML inference containers. See DOCKTOR.
• synthetic: Generate synthetic data using LLM prompts. See Synthetic Data.
• mimic: Submit FHIR requests/tests.
• plugins: Manage CLI plugins.

Typical Workflow

1. Create compose: npx dhti-cli compose add -m openmrs -m langserve -m redis -m ollama
2. Install elixir template: npx dhti-cli elixir install -g https://github.com/dermatologist/dhti-elixir-template.git -n dhti-elixir-template
3. Install conch template: npx dhti-cli conch install -g https://github.com/dermatologist/openmrs-esm-dhti-template.git -n openmrs-esm-dhti-template
4. Start stack: npx dhti-cli docker -u
5. Develop elixir with hot reload: npx dhti-cli elixir dev -d ../dhti-elixir-template -n dhti-elixir-template -c dhti-langserve-1
6. Develop conch UI: npx dhti-cli conch dev -d ../openmrs-esm-dhti-template -n openmrs-esm-dhti-template -c dhti-frontend-1
7. Build images to share: npx dhti-cli docker -n yourhandle/genai-elixir:1.0 -t elixir and corresponding conch.

Modules Catalog

Module	Purpose
openmrs	EMR (patient records, UI host)
langserve	Elixir runtime (serves LangChain apps)
redis	Vector store for RAG
ollama	Local LLM hosting
langfuse	Observability & tracing
neo4j	Graph analytics & relationships
cqlFhir	HAPI FHIR server with CQL support
mcpFhir	MCP tool server exposing FHIR utilities
mcpx	MCP Gateway for connecting tools
docktor	Traditional ML model containers

Tips

• Use small, focused elixirs—each with a single responsibility.
• Keep bootstrap overrides versioned so experiments are reproducible.
• Prefer semantic version tags for built images (1.0.0, 1.1.0, etc.).
• Share elixirs & conches via Git to promote reuse.

Troubleshooting

Issue	Suggestion
Container fails to start	Run docker logs <container>; check port conflicts.
Elixir not updated after dev sync	Confirm path correctness & no dependency changes requiring rebuild.
Conch assets not visible	Clear browser cache / ensure correct container name.
Ollama model missing	docker exec -it dhti-ollama-1 ollama pull <model> then update bootstrap.

Next Steps

See Getting Started for a guided quickstart and Architecture for system diagram.

---

CLI accelerates ideas → validated prototypes → production handoff.