---
title: "Start Ml job"
method: POST
path: "/job/ml"
tags: ["job"]
---

# Start Ml job

`POST /job/ml`

## Request body

- MLJobInput
  - `project_id` integer, required — The ID of the Project to analyse
  - `ml_job_type` 'QUOTATION_ATTRIBUTION' | 'TAG_RECOMMENDATION' | 'COREFERENCE_RESOLUTION' | 'DOCUMENT_EMBEDDING' | 'SENTENCE_EMBEDDING', required
  - `specific_ml_job_parameters` union, required — Specific parameters for the MLJob w.r.t it's type
    - QuotationAttributionParams
      - `ml_job_type` 'QUOTATION_ATTRIBUTION', required
      - `recompute` boolean — Whether to recompute already processed documents
    - DocTagRecommendationParams
      - `ml_job_type` 'TAG_RECOMMENDATION', required
      - `multi_class` boolean — Tags are mutually exclusive if `False`
      - `tag_ids` integer[] — Tags to consider. If empty, all tags applied to any document are considered.
      - `method` 'SIMPLE' | 'KNN' | 'EXCLUSIVE'
    - CoreferenceResolutionParams
      - `ml_job_type` 'COREFERENCE_RESOLUTION', required
      - `recompute` boolean — Whether to recompute already processed documents
    - DocumentEmbeddingParams
      - `ml_job_type` 'DOCUMENT_EMBEDDING', required
      - `recompute` boolean — Whether to recompute already processed documents
    - SentenceEmbeddingParams
      - `ml_job_type` 'SENTENCE_EMBEDDING', required
      - `recompute` boolean — Whether to recompute already processed documents

## Response `200`

Successful Response

- MlJobRead
  - `job_id` string, required — RQ job ID
  - `job_type` string, required — Type of the job
  - `project_id` integer, required — Project ID associated with the job
  - `status` 'queued' | 'finished' | 'failed' | 'started' | 'deferred' | 'scheduled' | 'stopped' | 'canceled', required
  - `status_message` string, nullable — Status message
  - `current_step` integer, required — Current step in the job process
  - `steps` string[], required — Total number of steps in the job process
  - `input` MLJobInput, required
    - `project_id` integer, required — The ID of the Project to analyse
    - `ml_job_type` 'QUOTATION_ATTRIBUTION' | 'TAG_RECOMMENDATION' | 'COREFERENCE_RESOLUTION' | 'DOCUMENT_EMBEDDING' | 'SENTENCE_EMBEDDING', required
    - `specific_ml_job_parameters` union, required — Specific parameters for the MLJob w.r.t it's type
      - QuotationAttributionParams
        - `ml_job_type` 'QUOTATION_ATTRIBUTION', required
        - `recompute` boolean — Whether to recompute already processed documents
      - DocTagRecommendationParams
        - `ml_job_type` 'TAG_RECOMMENDATION', required
        - `multi_class` boolean — Tags are mutually exclusive if `False`
        - `tag_ids` integer[] — Tags to consider. If empty, all tags applied to any document are considered.
        - `method` 'SIMPLE' | 'KNN' | 'EXCLUSIVE'
      - CoreferenceResolutionParams
        - `ml_job_type` 'COREFERENCE_RESOLUTION', required
        - `recompute` boolean — Whether to recompute already processed documents
      - DocumentEmbeddingParams
        - `ml_job_type` 'DOCUMENT_EMBEDDING', required
        - `recompute` boolean — Whether to recompute already processed documents
      - SentenceEmbeddingParams
        - `ml_job_type` 'SENTENCE_EMBEDDING', required
        - `recompute` boolean — Whether to recompute already processed documents
  - `output` unknown
  - `created` string, date-time, required — Created timestamp of the job
  - `finished` string, date-time, nullable — Finished timestamp of the job

## Other responses

- `422` — Validation Error

---

[API](https://skmtc.net/uhh-lt/apis/discourse-analysis-tool-suite-api.md) · [All operations](https://skmtc.net/uhh-lt/apis/discourse-analysis-tool-suite-api/llms.txt) · [OpenAPI document](https://skmtc-service-staging.skmtc.workers.dev/v1/apis/uhh-lt/discourse-analysis-tool-suite-api/revisions/6b22366ce2d1/schema)