$schema: "https://json-schema.org/draft/2020-12/schema"
type: object
description: Configuration schema for the unsupervised_analysis workflow.
additionalProperties: false
properties:
  mem:
    type: integer
    minimum: 1
    default: 32000
    description: Memory limit in MB used by workflow rules.
  threads:
    type: integer
    minimum: 1
    default: 2
    description: Default thread count used by workflow rules.
  annotation:
    type: string
    minLength: 1
    description: Path to the sample annotation CSV file.
  result_path:
    type: string
    minLength: 1
    description: Base output directory for workflow results.
  project_name:
    type: string
    minLength: 1
    description: Short project identifier used in reports and exported config copies.
  pca:
    type: object
    additionalProperties: false
    properties:
      n_components:
        anyOf:
          - type: integer
            minimum: 1
          - type: number
            exclusiveMinimum: 0
            maximum: 1
          - const: "mle"
        description: Number of components, explained variance fraction, or "mle".
      svd_solver:
        type: string
        enum:
          - auto
          - full
          - covariance_eigh
          - arpack
          - randomized
    required:
      - n_components
      - svd_solver
  umap:
    type: object
    additionalProperties: false
    properties:
      metrics:
        type: array
        items:
          type: string
          minLength: 1
        minItems: 0
      n_neighbors:
        type: array
        items:
          type: integer
          minimum: 2
        minItems: 0
      min_dist:
        type: array
        items:
          type: number
          minimum: 0
        minItems: 0
      n_components:
        type: array
        items:
          type: integer
          minimum: 2
        minItems: 0
      densmap:
        type: integer
        enum: [0, 1]
      connectivity:
        type: integer
        enum: [0, 1]
      diagnostics:
        type: integer
        enum: [0, 1]
    required:
      - metrics
      - n_neighbors
      - min_dist
      - n_components
      - densmap
      - connectivity
      - diagnostics
  heatmap:
    type: object
    additionalProperties: false
    properties:
      metrics:
        type: array
        items:
          type: string
          minLength: 1
      hclust_methods:
        type: array
        items:
          type: string
          minLength: 1
      n_observations:
        anyOf:
          - type: integer
            minimum: 1
          - type: number
            exclusiveMinimum: 0
            maximum: 1
      n_features:
        anyOf:
          - type: integer
            minimum: 1
          - type: number
            exclusiveMinimum: 0
            maximum: 1
    required:
      - metrics
      - hclust_methods
      - n_observations
      - n_features
  leiden:
    type: object
    additionalProperties: false
    properties:
      metrics:
        type: array
        items:
          type: string
          minLength: 1
      n_neighbors:
        type: array
        items:
          type: integer
          minimum: 2
      partition_types:
        type: array
        items:
          type: string
          minLength: 1
        minItems: 1
      resolutions:
        type: array
        items:
          type: number
          exclusiveMinimum: 0
        minItems: 1
      n_iterations:
        type: integer
        minimum: -1
        not:
          const: 0
    required:
      - metrics
      - n_neighbors
      - partition_types
      - resolutions
      - n_iterations
  clustree:
    type: object
    additionalProperties: false
    properties:
      count_filter:
        type: integer
        minimum: 0
      prop_filter:
        type: number
        minimum: 0
        maximum: 1
      layout:
        type: string
        enum:
          - tree
          - sugiyama
      categorical_label_option:
        type: string
        enum:
          - pure
          - majority
      numerical_aggregation_option:
        type: string
        enum:
          - mean
          - median
          - max
          - min
    required:
      - count_filter
      - prop_filter
      - layout
      - categorical_label_option
      - numerical_aggregation_option
  sample_proportion:
    type: number
    minimum: 0
    maximum: 1
    description: Set to 0 to skip internal cluster validation.
  metadata_of_interest:
    type: array
    items:
      type: string
      minLength: 1
    default: []
  coord_fixed:
    type: integer
    enum: [0, 1]
  scatterplot2d:
    type: object
    additionalProperties: false
    properties:
      size:
        type: number
        exclusiveMinimum: 0
      alpha:
        type: number
        minimum: 0
        maximum: 1
    required:
      - size
      - alpha
  features_to_plot:
    type: array
    items:
      type: string
      minLength: 1
    default: []
required:
  - annotation
  - result_path
  - project_name
  - pca
  - umap
  - heatmap
  - leiden
  - clustree
  - sample_proportion
  - metadata_of_interest
  - coord_fixed
  - scatterplot2d
  - features_to_plot