naftiko: 1.0.0-alpha2
info:
  label: CodeProject.AI Server API — Training
  description: 'CodeProject.AI Server API — Training. 6 operations. Lead operation: Stop a running training job. Self-contained
    Naftiko capability covering one Codeproject business surface.'
  tags:
  - Codeproject
  - Training
  created: '2026-05-19'
  modified: '2026-05-19'
binds:
- namespace: env
  keys:
    CODEPROJECT_API_KEY: CODEPROJECT_API_KEY
capability:
  consumes:
  - type: http
    namespace: ai-server-training
    baseUri: http://localhost:32168
    description: CodeProject.AI Server API — Training business capability. Self-contained, no shared references.
    resources:
    - name: v1-train-cancel
      path: /v1/train/cancel
      operations:
      - name: canceltraining
        method: POST
        description: Stop a running training job
        outputRawFormat: json
        outputParameters:
        - name: result
          type: object
          value: $.
    - name: v1-train-create_dataset
      path: /v1/train/create_dataset
      operations:
      - name: createtrainingdataset
        method: POST
        description: Create a custom training dataset
        outputRawFormat: json
        outputParameters:
        - name: result
          type: object
          value: $.
    - name: v1-train-get_dataset
      path: /v1/train/get_dataset
      operations:
      - name: gettrainingdataset
        method: POST
        description: Retrieve a training dataset
        outputRawFormat: json
        outputParameters:
        - name: result
          type: object
          value: $.
    - name: v1-train-get_model
      path: /v1/train/get_model
      operations:
      - name: gettrainedmodel
        method: POST
        description: Retrieve a trained model
        outputRawFormat: json
        outputParameters:
        - name: result
          type: object
          value: $.
    - name: v1-train-status
      path: /v1/train/status
      operations:
      - name: gettrainingstatus
        method: POST
        description: Check training progress
        outputRawFormat: json
        outputParameters:
        - name: result
          type: object
          value: $.
    - name: v1-train-train_model
      path: /v1/train/train_model
      operations:
      - name: trainmodel
        method: POST
        description: Start YOLOv5 model training
        outputRawFormat: json
        outputParameters:
        - name: result
          type: object
          value: $.
    authentication:
      type: apikey
      key: X-Disabled
      value: '{{env.CODEPROJECT_API_KEY}}'
      placement: header
  exposes:
  - type: rest
    namespace: ai-server-training-rest
    port: 8080
    description: REST adapter for CodeProject.AI Server API — Training. One Spectral-compliant resource per consumed operation,
      prefixed with /v1.
    resources:
    - path: /v1/v1/train/cancel
      name: v1-train-cancel
      description: REST surface for v1-train-cancel.
      operations:
      - method: POST
        name: canceltraining
        description: Stop a running training job
        call: ai-server-training.canceltraining
        outputParameters:
        - type: object
          mapping: $.
    - path: /v1/v1/train/create-dataset
      name: v1-train-create-dataset
      description: REST surface for v1-train-create_dataset.
      operations:
      - method: POST
        name: createtrainingdataset
        description: Create a custom training dataset
        call: ai-server-training.createtrainingdataset
        outputParameters:
        - type: object
          mapping: $.
    - path: /v1/v1/train/get-dataset
      name: v1-train-get-dataset
      description: REST surface for v1-train-get_dataset.
      operations:
      - method: POST
        name: gettrainingdataset
        description: Retrieve a training dataset
        call: ai-server-training.gettrainingdataset
        outputParameters:
        - type: object
          mapping: $.
    - path: /v1/v1/train/get-model
      name: v1-train-get-model
      description: REST surface for v1-train-get_model.
      operations:
      - method: POST
        name: gettrainedmodel
        description: Retrieve a trained model
        call: ai-server-training.gettrainedmodel
        outputParameters:
        - type: object
          mapping: $.
    - path: /v1/v1/train/status
      name: v1-train-status
      description: REST surface for v1-train-status.
      operations:
      - method: POST
        name: gettrainingstatus
        description: Check training progress
        call: ai-server-training.gettrainingstatus
        outputParameters:
        - type: object
          mapping: $.
    - path: /v1/v1/train/train-model
      name: v1-train-train-model
      description: REST surface for v1-train-train_model.
      operations:
      - method: POST
        name: trainmodel
        description: Start YOLOv5 model training
        call: ai-server-training.trainmodel
        outputParameters:
        - type: object
          mapping: $.
  - type: mcp
    namespace: ai-server-training-mcp
    port: 9090
    transport: http
    description: MCP adapter for CodeProject.AI Server API — Training. One tool per consumed operation, routed inline through
      this capability's consumes block.
    tools:
    - name: stop-running-training-job
      description: Stop a running training job
      hints:
        readOnly: false
        destructive: false
        idempotent: false
      call: ai-server-training.canceltraining
      outputParameters:
      - type: object
        mapping: $.
    - name: create-custom-training-dataset
      description: Create a custom training dataset
      hints:
        readOnly: false
        destructive: false
        idempotent: false
      call: ai-server-training.createtrainingdataset
      outputParameters:
      - type: object
        mapping: $.
    - name: retrieve-training-dataset
      description: Retrieve a training dataset
      hints:
        readOnly: true
        destructive: false
        idempotent: false
      call: ai-server-training.gettrainingdataset
      outputParameters:
      - type: object
        mapping: $.
    - name: retrieve-trained-model
      description: Retrieve a trained model
      hints:
        readOnly: true
        destructive: false
        idempotent: false
      call: ai-server-training.gettrainedmodel
      outputParameters:
      - type: object
        mapping: $.
    - name: check-training-progress
      description: Check training progress
      hints:
        readOnly: true
        destructive: false
        idempotent: false
      call: ai-server-training.gettrainingstatus
      outputParameters:
      - type: object
        mapping: $.
    - name: start-yolov5-model-training
      description: Start YOLOv5 model training
      hints:
        readOnly: false
        destructive: false
        idempotent: false
      call: ai-server-training.trainmodel
      outputParameters:
      - type: object
        mapping: $.