Project Structure

Description of the folder structure expected by SUSHI for a FSH project

Minimal Project

The simplest FSH project contains only a configuration file and an input/fsh subdirectory of FSH files containing FHIR Shorthand definitions. It might look like this:

├── input
|    └── fsh
|        ├── file1.fsh
|        ├── file2.fsh
|        └── file3.fsh
└── sushi-config.yaml

Each FSH file can contain multiple FSH definitions of varying types. FSH file names are not significant, but must end with the .fsh extension. In addition, FSH files can be organized into subdirectories. This provides authors the flexibility to organize their FSH definitions in whatever way makes sense to then.

The sushi-config.yaml file provides project configuration data to SUSHI. It is described further in the Configuration documentation.

Initializing a SUSHI Project

Setting up a fully-featured FSH project can be complex, so SUSHI provides an init command to do it automatically. When sushi init is run, SUSHI will request project information from the user:

Name (Default: ExampleIG): my-project
Id (Default: fhir.example):
Canonical (Default:
Status (Default: draft): active
Version (Default: 0.1.0): 2.0.0
Release Label (Default: ci-build): trial-use
Publisher Name (Default: Example Publisher): MyPublisher
Publisher Url (Default:
Initialize SUSHI project in C:\Users\shorty\dev\my-project? [y/n]: y

These values are then used to generate a project structure compatible with the FHIR IG Publisher:

├── .gitignore
├── _genonce.bat
├── _updatePublisher.bat
├── ig.ini
├── input
|   ├── ignoreWarnings.txt
|   ├── fsh
|   |   └── patient.fsh
|   └── pagecontent
|       └──
└── sushi-config.yaml

In addition to the input/fsh folder, init creates an input/pagecontent folder containing a dummy home page for your IG. The files ig.ini and ignoreWarnings.txt are required by the template-based IG Publisher. The .bat and .sh scripts which allow you to run the IG Publisher from your command line. Finally, a default .gitignore file for integration with GitHub is provided. From this point on, you can modify the configuration and definitions as necessary.

Using the HL7 IG Publisher and Auto-Builder

This project structure integrates with the HL7 IG Publisher Auto-Builder. When the IG Publisher detects an input/fsh subdirectory, it will automatically run SUSHI on the project directory and output the SUSHI results to a fsh-generated directory (e.g., my-project/fsh-generated in the example above). It will then continue with the normal IG Publisher process.

This approach allows a GitHub repository to be configured such that whenever changes to FSH files are pushed to GitHub, the Auto-Builder will pick them up, run the SUSHI/IG Publisher process, and publish the resulting IG to

SUSHI provides support for several of the files and directories required by the template-based IG Publisher for building Implementation Guides. Some IG customizations can be configured using additional properties in the sushi-config.yaml file. A FSH project integrated into the template-based IG Publisher may look like this:

├── ig.ini
├── input
|   ├── fsh
|   |   ├── file1.fsh
|   |   ├── file2.fsh
|   |   └── file3.fsh
│   ├── ignoreWarnings.txt
│   ├── images
│   │   ├── myDocument.pdf
│   │   ├── myGraphic.png
│   │   └── mySpreadsheet.xlsx
│   ├── includes
│   │   └── menu.xml
│   └── pagecontent
│       ├──
│       ├──
│       ├──
│       └──
├── package-list.json
├── sushi-ignoreWarnings.txt
└── sushi-config.yaml

You can populate your project as follows:

  • sushi-config.yaml: This file provides configuration data to SUSHI. It is described further in the Configuration documentation.

  • input/fsh/*.fsh: FSH files contain the FHIR Shorthand definitions for all the resources and examples in your IG.

  • ig.ini: Configuration file required for the FHIR IG Publication process. NOTE: As of the SUSHI 1.0 release, this file MUST use a template based on fhir.base.template#current. Specific template versions (i.e., other than #current) are expected to work in the future. For now, any of the following should work:

    • template = fhir.base.template#current
    • template = hl7.base.template#current
    • template = hl7.fhir.template#current
    • template = hl7.davinci.template#current
    • template = hl7.cda.template#current
  • input/ignoreWarnings.txt: This file is used to suppress specific QA warnings and information messages produced by the FHIR IG Publisher (as opposed to SUSHI).

  • input/images/*: Put anything that is not a page in the IG, such as images, spreadsheets or zip files, in the input/images subdirectory. These files can be referenced by user-provided pages or menus.

  • input/includes/menu.xml: If present, this file will be used for the IG’s main menu layout. Note that the presence of this file will block usage of the menu property in sushi-config.yaml.

  • input/pagecontent/*: Put either markup (.xml) or markdown (.md) files with the narrative content of your IG in the input/pagecontent/ subdirectory. These files are the sources for the html pages that accompany the automatically-generated pages of your IG. The header and footer of these pages are automatically generated, so your content should not include these elements. Any number of pages can be added. In addition to stand-alone pages, you can provide additional text for generated artifact pages. The naming of these files is significant:

    • index.xml|md: This file provides the content for the IG’s main page.
    • N_pagename.xml|md: If present, these files will be generated as individual pages in the IG. The leading integer (N) determines the order of the pages in the table of contents. Adding a leading integer is optional, and in the absence of a leading integer, SUSHI will sort the pages alphabetically. The order of the pages can also be explitly specified with the pages property in sushi-config.yaml.
    • {artifact-file-name}-intro.xml|md: If present, the contents of the file will be placed on the relevant page before the artifact’s definition.
    • {artifact-file-name}-notes.xml|md: If present, the contents of the file will be placed on the relevant page after the artifact’s definition.
  • input/{supported-resource-input-directory}/* (not shown above): JSON or XML files in supported resource directories (e.g., profiles, extensions, examples, etc.) can be referenced by FHIR artifacts defined in FSH, and will be added to the generated ImplementationGuide.json file. If there are additional subfolders (e.g., input/resources/nested), use the path-resource parameter in sushi-config.yaml to tell the IG Publisher which additional input paths to process (see Specifying Additional Resource Paths for details).

  • package-list.json: This optional file, described here, should contain the version history of your IG.

  • sushi-ignoreWarnings.txt: This optional file can be used to suppress warnings logged by SUSHI. Errors and informational logs from SUSHI cannot be ignored. This file should be placed either at the root of the project (e.g., my-project/sushi-ignoreWarnings.txt in the example above), or within the input directory (e.g., my-project/input/sushi-ignoreWarnings.txt). Warnings will be ignored if they completely match the contents of any line of the file (one line per warning, case-sensitive). Regular expressions can also be specified, one per line, indicated by starting and ending the line with /. For example:

    Instance PatientExample1 is not an instance of a resource, so it should only be used inline on other instances, and it will not be exported to a standalone file. Specify "Usage: #inline" to remove this warning.
    /Detected the following non-conformant Resource definitions.*/

    Any warning which exactly matches the contents of the first line will be ignored. The second line specifies that any warning beginning with Detected the following non-conformant Resource definitions will be ignored.