BIOMERO Schema

The BIOMERO schema is defined as a pydantic model, and can be exported as json-schema.

A CLI tool is included to validate and parse JSON files against the schema.

The schema is a mix of:

• the BiaFlows subset of the (now deprecated) original cytomine schema
• the new cytomine App Engine task schema
• the NIST fair-compute schema
• the Bilayers schema
• the BIAFlows problem class list

Installation

pixi is used for installation and running some tasks. It's not strictly required, but recommended to install pixi.

cd biomero_schema
pixi shell

Or use pip install -e . (or equivalent) in a virtual environment.

Usage

help

biomero-schema --help

Predefined Tasks

Available pixi tasks.

pixi task list

JSON Schema

The pydantic model is able to generate a JSON schema, which is printable with:

pixi run json-schema
# OR directly
biomero-schema schema

Validate a JSON file against the schema

pixi run test-validate

Parse a JSON file into a Pydantic representation

# Basic parsing with summary
pixi run test-parse

# Pretty print the full parsed object
pixi run test-pparse

# Output as JSON
pixi run test-jparse

Files

• src/biomero_schema/models.py - Pydantic models for the schema
• src/biomero_schema/cli.py - CLI implementation
• tests/example_workflow.json - Example workflow file for testing

Schema Structure

The schema defines a workflow with the following main components:

• Basic Info: name, description, schema-version
• People & Organizations: authors, institutions
• Citations: Required list of tool citations
• Problem Class: Optional BIAFlows problem class
• Container: container-image specification
• Configuration: Technical settings and resource requirements
• Parameters: inputs and outputs with type definitions
• Command Line: Template for execution

Example

See tests/example_workflow.json for a complete example of a valid workflow definition.

Schema Details

The following is json-ish psuedo-code that describes the schema:

schema = 
{
  "name": "string",                      // Required. GitHub workflow repository name (without prefix). E.g. NucleiTracking-ImageJ
  "description": "string"                // Required. Description of workflow.
  "schema-version": "string"             // Required. Semver of schema version.
  "authors":                             // Optional. Authors list.
    [
      {
        "name": "string",                // Required. Full name of author.
        "email": "optional",             // Optional. Email address of author.
        "affiliations": "string[]"       // Optional. List of affiliations matching "id" of an instituttion in instititions list.
      }
    ],
  "institutions":                        // Optional. Institutions list.
    [
      {
        "id": "string",                  // Required. Unique institute identifier.
        "name": "string"                 // Optional. Name of the institions. Defaults to id.
      }
    ],
  "citations":                           // Optional. List of citations for the tool. Defaults to empty list.
    [
      {
        "name": "string",                // Required. Name of the tool being cited.
        "doi": "string",                 // Optional. DOI number of the tool being cited. Defaults to empty string.
        "license": "string",             // Required. License of the tool being cited.
        "description": "string"          // Optional. Description of the tool being cited. Defaults to empty string.
      }
    ],
  // corresponding to: https://neubias-wg5.github.io/problem_class_ground_truth.html#steps-section
  "problem-class":                       // Optional. Biaflows problem class ("object-segmentation" | "pixel-classification" | "object-counting" | "object-detection" | "filament-tree-tracing" | "filament-networks-tracing" | "landmark-detection" | "particle-tracking" | "object-tracking").
  "container-image":                     // Required. Base cotnainer description.
    {
      "image": "string",                 // Required. Image to match the name of your workflow GitHub repository (lower case only). E.g. neubiaswg5/w_nucleitracking-imagej:1.0.0
      "type": "string",                  // Required. "oci" | "singularity" (lower case only).
      "platforms": "string"[]            // Optional. Build-time multi-platform targets.
    },
  "configuration":                       // Optional. Technical configuration.
  {
    "input_folder": "string",            // Optional. Full path where the input folder must be mounted in the container. Defaults to "/inputs".
    "output_folder": "string",           // Optional. Full path where teh output folder must be mounted in the container. Defaults to "/outputs".
    "resources":                         // Optional.
      {
        "networking": "boolean",         // Optional. Whether internet connection is needed. Defaults to False.
        "ram-min": "number",             // Optional. Minimum RAM in mebibytes (Mi). Defaults to 0.
        "cores-min": "number",           // Optional. Minimum number of CPU cores. Defaults to 1.
        "gpu": "boolean",                // Optional. GPU/accelerator required. Defaults to False.
        "cuda-requirements":             // Optional. GPU Cuda-related requirements.
          {
            "device-memory-min": "number", // Optional. Minimum device memory. Defaults to 0.
            "cuda-compute-capability": "string|string[]", // Optional: The cudaComputeCapability Schema; single min value or list of valid values. Defaults to None.
          },
        "cpuAVX": "boolean",             // Optional. Advanced Vector Extensions (AVX) CPU capability required. Defaults to False.
        "cpuAVX2": "boolean",            // Optional. Advanced Vector Extensions 2 (AVX2) CPU capability required. Defaults to False.
      }
  }
  "inputs":                              // Required. List of parameter descriptors.
    [
      {
        // references to "@id" get the value of "id" in lowercase
        // references to "@ID" get the value of "id" in uppercase
        "id": "string",                  // Required. Unique parameter identifier.
        "type": "string",                // Required. Data type of the parameter (Number|String|ineger|flaot|boolean|string|file|image|array).
        "name": "string",                // Optional. Human-readable display name appearing in BIAFLOWS UI (paramater dialog box). Defaults to "@id".
        "description": "string",         // Optional. Description of paramater. Context help in BIAFLOWS UI (paramater dialog box). Soft Defaults to "".
        "value-key": "string",           // Optional. Substitution key in CLI. Defaults to "[@ID]".
        "command-line-flag": "string",   // Optional. CLI flag. Defaults to "--@id".
        "default-value": "string|number|boolean", // Optional. Default value in BIAFLOWS UI (paramater dialog box). Soft Defaults to empty string.
        "optional": "boolean",           // Optional. If true, parameter not required. Soft Defaults to False.
        "set-by-server": "boolean",      // Optional. If true, parameter is server-assigned. Soft Defaults to False.
      }
    ]
  "outputs":                           // Optional. List of output parameter descriptors.
    [
      {
        // references to "@id" get the value of "id" in lowercase
        // references to "@ID" get the value of "id" in uppercase
        "id": "string",                  // Required. Unique parameter identifier.
        "type": "string",                // Required. Data type of the parameter (Number|String).
        "name": "string",                // Optional. Human-readable display name appearing in BIAFLOWS UI (paramater dialog box). Defaults to "@id".
        "description": "string",         // Optional. Description of paramater. Context help in BIAFLOWS UI (paramater dialog box). Soft Defaults to "".
        "value-key": "string",           // Optional. Substitution key in CLI. Defaults to "[@ID]".
        "command-line-flag": "string",   // Optional. CLI flag. Defaults to "--@id".
        "default-value": "string|number|boolean", // Optional. Default value in BIAFLOWS UI (paramater dialog box). Soft Defaults to empty string.
        "optional": "boolean",           // Optional. If true, parameter not required. Soft Defaults to False.
        "set-by-server": "boolean",      // Optional. If true, parameter is server-assigned. Soft Defaults to False.
      }
    ]
  "command-line": "string"               // Required. e.g. "python wrapper.py CYTOMINE_HOST CYTOMINE_PUBLIC_KEY CYTOMINE_PRIVATE_KEY CYTOMINE_ID_PROJECT CYTOMINE_ID_SOFTWARE IJ_RADIUS IJ_THRESHOLD".
}

file =
{
  "format": "string"                    // Required. Extension of the file type (.csv).
}

image =
{
  "sub-type": "string",                 // Required. Image type (grayscale|color|binary|labeled|class).
  "format": "string"                    // Required. Extension of the image type (tif, png, jpg, jpeg, tiff, ometiff).
}

array =
{
  "format": "string"                    // Required. Extension of the file type (npy, npz)
}