generated: '2026-08-15'
method: searched
source: >-
  https://github.com/tuva-health/FHIR_inferno#usage +
  https://tuva-health.github.io/tuva_empi/docs/getting-started/production-environment +
  https://github.com/tuva-health/tuva-core/blob/main/AGENTS.md
note: >-
  Tuva ships no installable, versioned CLI binary and no package-manager-installed
  command (nothing on PyPI, npm, Homebrew or crates.io). What it does publish are
  three documented command surfaces, recorded here honestly for what they are:
  a Python utility script in the FHIR Inferno repo, container entrypoint commands
  for the Tuva EMPI backend image, and a repo-local dbt wrapper script. Users
  otherwise drive Tuva through dbt's own CLI, which is not Tuva's software.
surfaces:
  - id: fhir-inferno
    kind: python-utility-script
    repo: https://github.com/tuva-health/FHIR_inferno
    install: git clone https://github.com/tuva-health/FHIR_inferno
    published_to_registry: false
    commands:
      - command: python utils/run_parser.py --config <config.ini> --input <input.json> --output <output.csv>
        description: Parse a FHIR JSON/NDJSON file (or a directory of them) into CSV via a mapping config.
        flags:
          - flag: --config
            description: 'Path to the .ini mapping config (requires [GenConfig] and [Struct] sections).'
          - flag: --input
            description: Input FHIR .json or .ndjson file, or a directory of them.
          - flag: --output
            description: Output CSV path.
    library_equivalent: parseFhir.parse(configPath=..., inputPath=..., outputPath=...)
    helper_scripts: https://github.com/tuva-health/FHIR_inferno/tree/main/helper_scripts
  - id: tuva-empi-backend
    kind: container-entrypoint-commands
    image: ghcr.io/tuva-health/tuva-empi-backend
    docs: https://tuva-health.github.io/tuva_empi/docs/getting-started/production-environment
    commands:
      - {command: '(none)', description: Runs the Tuva EMPI API on port 8000.}
      - {command: migrate, description: Run database migrations; must be run once per new version before deploying containers. Only one instance at a time.}
      - {command: bootstrap, description: Register the initial admin user from initial_setup.admin_email. Run once on first launch.}
      - {command: matching-service, description: Run the matching service that dispatches Splink matching jobs.}
    env:
      - {name: TUVA_EMPI_CONFIG_FILE, description: Path to the backend JSON config file (takes priority).}
      - {name: TUVA_EMPI_CONFIG_AWS_SECRET_ARN, description: ARN of an AWS Secrets Manager secret holding the config file.}
      - {name: TUVA_EMPI_MATCHING_SERVICE__K8S_JOB_RUNNER__JOB_IMAGE, description: Image used for the matching job under the K8s job runner.}
  - id: dbt-local
    kind: repo-local-wrapper-script
    repo: https://github.com/tuva-health/tuva-core
    commands:
      - {command: scripts/dbt-local deps, description: Install package dependencies locally.}
      - {command: scripts/dbt-local debug, description: Validate the local warehouse connection.}
      - {command: 'scripts/dbt-local build --select <selector>', description: Build a narrow slice of the Tuva models.}
      - {command: scripts/dbt-local build --full-refresh, description: Full-refresh build, expected before opening a PR.}
    note: >-
      Contributor tooling described in the repository's AGENTS.md, not an end-user
      product CLI.