Nanonets · Arazzo Workflow

Nanonets Upload Training URLs and Train

Version 1.0.0

Add training images to an OCR model from public URLs, then start training.

1 workflow 1 source API 1 provider
View Spec View on GitHub AIArtificial IntelligenceOCRDocument AIIntelligent Document ProcessingData ExtractionWorkflow AutomationComputer VisionNo-CodeArazzoWorkflows

Provider

nanonets

Workflows

upload-training-urls-and-train
Upload training images by URL and start a training run.
Registers public training image URLs with a Nanonets OCR model and then queues a training job.
2 steps inputs: authorization, data, modelId, urls outputs: trainResponse, uploadResponse
1
uploadTrainingUrls
ocrModelUploadUrlsByModelIdPost
Register the public training image URLs against the model.
2
trainModel
ocrModelTrainByModelIdPost
Trigger training (or retraining) on the model after the URLs are registered.

Source API Descriptions

Arazzo Workflow Specification

nanonets-upload-training-urls-and-train-workflow.yml Raw ↑
arazzo: 1.0.1
info:
  title: Nanonets Upload Training URLs and Train
  summary: Add training images to an OCR model from public URLs, then start training.
  description: >-
    A URL-driven variant of the training-prep flow. The workflow registers one or
    more publicly accessible training image URLs against the target Nanonets OCR
    model and then triggers a training run on the model. Every step spells out its
    request inline so the flow can be read and executed without opening the
    underlying OpenAPI description.
  version: 1.0.0
sourceDescriptions:
- name: ocrApi
  url: ../openapi/nanonets-ocr-api-openapi.yml
  type: openapi
workflows:
- workflowId: upload-training-urls-and-train
  summary: Upload training images by URL and start a training run.
  description: >-
    Registers public training image URLs with a Nanonets OCR model and then
    queues a training job.
  inputs:
    type: object
    required:
    - authorization
    - modelId
    - urls
    properties:
      authorization:
        type: string
        description: HTTP Basic credential header value (Basic <base64 of apiKey:>).
      modelId:
        type: string
        description: Unique identifier for the Nanonets OCR model to train.
      urls:
        type: array
        description: Publicly accessible URLs of the training images.
        items:
          type: string
      data:
        type: string
        description: Optional JSON string describing the annotation for each URL.
  steps:
  - stepId: uploadTrainingUrls
    description: Register the public training image URLs against the model.
    operationId: ocrModelUploadUrlsByModelIdPost
    parameters:
    - name: Authorization
      in: header
      value: $inputs.authorization
    - name: model_id
      in: path
      value: $inputs.modelId
    requestBody:
      contentType: application/json
      payload:
        urls: $inputs.urls
        data: $inputs.data
    successCriteria:
    - condition: $statusCode == 200
    outputs:
      uploadResponse: $response.body
  - stepId: trainModel
    description: Trigger training (or retraining) on the model after the URLs are registered.
    operationId: ocrModelTrainByModelIdPost
    parameters:
    - name: Authorization
      in: header
      value: $inputs.authorization
    - name: model_id
      in: path
      value: $inputs.modelId
    successCriteria:
    - condition: $statusCode == 200
    outputs:
      trainResponse: $response.body
  outputs:
    uploadResponse: $steps.uploadTrainingUrls.outputs.uploadResponse
    trainResponse: $steps.trainModel.outputs.trainResponse