Scorer API

• Scorer
  • postPlace a request to score a document
  • postScore a document stored in a remote location
• Job
  • getInformation about a job
  • getInformation about a finished job
  • getObtain logs from scoring
• Pipeline
  • getReturn list of registered pipelines in given project
  • postPublishes a pipeline
  • getGet a config of a pipeline
  • putUpdates an existing pipeline
  • delDeletes a published pipeline
  • getGet logs of a failing pipeline
• Telemetry
  • getTotal number of scored pages
  • getNumber of scored pages by given pipeline

API docs by Redocly

API endpoints, version 0 (version not set)

Download OpenAPI specification:Download

Endpoints previously implemented in docai-scorer.

Find out more about Document AI

Scorer

Place a request to score a document

Authorizations:

access_token

path Parameters

id required

string Model ID. Currently only 1 model with ID "main" is supported

Request Body schema: multipart/form-data

documentGuid	string Document guid
document required	string <binary> Document to score

Responses

https://h2oai.github.io/async/model/{id}/score

Response samples

• 202

Content type

application/json

{

• "jobId": "38ce1872-30dc-11ec-a351-acde48001122",
• "jobUri": "http://localhost:8080/job/38ce1872-30dc-11ec-a351-acde48001122"

}

Score a document stored in a remote location

Authorizations:

access_token

path Parameters

id required

string Model ID

Request Body schema: application/json

inputKey required	string
inputBucket required	string
outputKey required	string
outputBucket required	string
imageKey required	string
imageBucket required	string
logKey required	string
logBucket required	string
extra	string
documentGuid	string

Responses

https://h2oai.github.io/async/model/{id}/remote-document-score

Request samples

• Payload

Content type

application/json

{

• "inputKey": "string",
• "inputBucket": "string",
• "outputKey": "string",
• "outputBucket": "string",
• "imageKey": "string",
• "imageBucket": "string",
• "logKey": "string",
• "logBucket": "string",
• "extra": "string",
• "documentGuid": "string"

}

Job

Information about a job

Authorizations:

access_token

path Parameters

id required

string Job ID obtained from an async call

Responses

https://h2oai.github.io/job/{id}

Response samples

• 200

Content type

application/json

{

• "id": "38ce1872-30dc-11ec-a351-acde48001122",
• "status": "succeeded",
• "resultUri": "http://localhost:8080/job/38ce1872-30dc-11ec-a351-acde48001122/result",
• "logsScorerUri": "http://localhost:8080/job/38ce1872-30dc-11ec-a351-acde48001122/logs/scorer"

}

Information about a finished job

Authorizations:

access_token

path Parameters

id required

string Job ID obtained from an async call

Responses

https://h2oai.github.io/job/{id}/result

Response samples

• 200

Content type

application/json

{

• "documentGuid": "file_8f00870c-6ed8-4cd7-5864-08d7dc0e99e2",
• "entities": [

  • {

    • "pageIndex": "0",
    • "dpiX": 200,
    • "entityId": "f6a142f6-e438-4ca5-b117-ceacc4695f28",
    • "sLabel": "DoBValue",
    • "sValue": "03/26/2016",
    • "labelConfidence": 0.6921263933181763,
    • "rzRect": {

      • "bx": 236.66666666666666,
      • "by": 990,
      • "ex": 430,
      • "ey": 962

      },
    • "rectImg": "(imgBase64)"

    },
  • "..."

  ],
• "pages": {

  • "0": {

    • "pageClass": "advertisement",
    • "classConfidence": 0.15744306375191164

    }

  },
• "entityConfidences": [

  • {

    • "entityClass": "DoBValue",
    • "topEntities": [

      • {

        • "entityValue": "03/26/2016",
        • "entityConfidence": 0.6921263933181763,
        • "entityId": "f6a142f6-e438-4ca5-b117-ceacc4695f28"

        },
      • {

        • "entityValue": "3/26/2016",
        • "entityConfidence": 0.7135215997695923,
        • "entityId": "933e4c41-a209-4600-8dbe-6b83f170f4f0"

        }

      ]

    },
  • "..."

  ]

}

Obtain logs from scoring

Authorizations:

access_token

path Parameters

id required

string Job ID obtained from an async call

Responses

https://h2oai.github.io/job/{id}/logs/scorer

Response samples

• 200

Content type

application/json

"string"

Pipeline

Return list of registered pipelines in given project

Authorizations:

access_token

query Parameters

projectId required	string
includeDeleted	boolean

Responses

https://h2oai.github.io/pipeline

Response samples

• 200

Content type

application/json

[

• {

  • "name": "dev",
  • "scoringUrl": "http://localhost:8080/async/model/dev/score",
  • "modelMetadata": {

    • "version": "26accf"

    },
  • "status": "Running",
  • "statusMessage": "status message",
  • "preSleepingState": "Running"

  },
• "..."

]

Publishes a pipeline

Authorizations:

access_token

Request Body schema: application/json

required

Pipeline configuration

any

Responses

https://h2oai.github.io/pipeline

Request samples

• Payload

Content type

application/json

null

Response samples

• 200

Content type

application/json

{

• "name": "pipeline-name",
• "scoringUrl": "http://localhost:8080/async/model/pipeline-name/score"

}

Get a config of a pipeline

Authorizations:

access_token

path Parameters

id required

string Pipeline ID

Responses

https://h2oai.github.io/pipeline/{id}

Response samples

• 200

Content type

*/*application/json*/*

No sample

Updates an existing pipeline

Authorizations:

access_token

path Parameters

id required

string Pipeline ID

Request Body schema: application/json

Pipeline definition

affinity	object
required	object (Autoscaler)
customPostProcessor required	string
customPipeline required	boolean
muteOutput required	boolean
name required	string
nodeSelector	object
pipelineConfig required	string
required	object (Resources)
	Array of objects (Tolerations)
workspaceId required	string
builtinPostProcessor	string Enum: "generic" "supply-chain"

Responses

https://h2oai.github.io/pipeline/{id}

Request samples

• Payload

Content type

application/json

{

• "affinity": { },
• "autoscaler": {

  • "maxReplicas": 4,
  • "minReplicas": 1

  },
• "customPostProcessor": "",
• "customPipeline": false,
• "muteOutput": false,
• "name": "pipe-99",
• "nodeSelector": { },
• "pipelineConfig": "spec:\n pipeline:\n steps:\n - tasks:\n - name: Intake\n type: PipelineTask\n class: argus.processors.ocr_processors.Intake\n parameters:\n root_docs_path: /input_files\n follow_symlinks: true\n - tasks:\n - name: PdfExtract\n type: PipelineTask\n class: argus.processors.ocr_processors.PdfTextExtract\n - tasks:\n - name: ImageNormalize\n class: argus.processors.ocr_processors.NormalizeImages\n parameters:\n resample_to_dpi: 300\n normalize_image_format: .jpg\n - tasks:\n - name: OCR\n class: argus.processors.ocr_processors.GenericOcr\n parameters:\n ocr_style: BestTextExtract\n - tasks:\n - name: Predict\n class: argus.processors.train_eval_processors.TrainEvalProcessor\n parameters:\n artifacts:\n base_model_name_or_path: {{ .labellingModelName }}\n - tasks:\n - name: Predict\n class: argus.processors.train_eval_processors.TrainEvalProcessor\n parameters:\n artifacts:\n base_model_name_or_path: {{ .classificationModelName }}\n - tasks:\n - name: PostProcess\n class: argus.processors.post_processors.generic_post_processor.PostProcessor\n parameters:\n output_format: json\n labeling_threshold: 0.8\n output_annoset_format:\n options: { }\n write_all_steps: false\n cache_processors: true\n artifacts:\n - version: \"\"\n source: http://mock-model-registry-service.default.svc.cluster.local:8080/{{ .labellingModelName }}:download\n name: {{ .labellingModelName }}\n - version: \"\"\n source: http://mock-model-registry-service.default.svc.cluster.local:8080/{{ .classificationModelName }}:download\n name: {{ .classificationModelName }}\n",
• "resources": {

  • "requests": {

    • "cpu": "500m",
    • "memory": "2G"

    },
  • "limits": {

    • "cpu": "4",
    • "memory": "6G"

    }

  },
• "tolerations": [

  • {

    • "effect": "NoExecute",
    • "key": "key1",
    • "operator": "Equal",
    • "value": "string",
    • "tolerationSeconds": 3600

    }

  ],
• "workspaceId": "string",
• "builtinPostProcessor": "generic"

}

Response samples

• 200

Content type

application/json

{

• "name": "pipeline-name",
• "scoringUrl": "http://localhost:8080/async/model/pipeline-name/score"

}

Deletes a published pipeline

Authorizations:

access_token

path Parameters

id required

string Pipeline ID

Responses

https://h2oai.github.io/pipeline/{id}

Get logs of a failing pipeline

Authorizations:

access_token

path Parameters

id required

string Pipeline ID

query Parameters

lines

number number of lines to return from log tail

Responses

https://h2oai.github.io/pipeline/{id}/logs

Telemetry

Total number of scored pages

Authorizations:

access_token

Responses

https://h2oai.github.io/telemetry

Response samples

• 200

Content type

application/json

{

• "total": 12,
• "lastScoringEventTime": "2023-04-12T23:20:50.52Z"

}

Number of scored pages by given pipeline

Authorizations:

access_token

path Parameters

id required

string Pipeline ID

Responses

https://h2oai.github.io/telemetry/{id}

Response samples

• 200

Content type

application/json

{

• "total": 12,
• "lastScoringEventTime": "2023-04-12T23:20:50.52Z"

}

---

Feedback

• Submit and view feedback for this page
• Send feedback about H2O Document AI to cloud-feedback@h2o.ai