openapi: '3.0.1'
info:
  description: ''
  version: '1.0.0'
  title: 'Parsr API'
  contact:
    email: 'opensource@axa.com'
    name: Parsr
    url: https://github.com/axa-group/Parsr

servers:
  - url: 'https://localhost:3001/api/v1'
    description: 'Localhost server'

paths:
  /document:
    post:
      operationId: postDocument
      tags:
        - 'Input'
      summary: Pipeline Input
      description: 'Entry point to add a file to the processing queue.'
      requestBody:
        content:
          multipart/form-data:
            schema:
              type: object
              properties:
                file:
                  type: string
                  format: binary
                config:
                  $ref: '#/components/schemas/config'
      responses:
        '202':
          description: Accepted
          content:
            text/plain:
              schema:
                type: string
                example: '{id}'
          headers:
            Location:
              description: Path to the queue for the given job.
              schema:
                type: string
                example: /api/v1.0/queue/{id}
        '415':
          description: Unsupported Media Type

  /queue/{id}:
    get:
      operationId: getQueueStatus
      tags:
        - 'Processing'
      summary: 'Get the status of the queue'
      description: 'Get the status of the queue'
      parameters:
        - $ref: '#/components/parameters/id'
      responses:
        '200':
          description: Ok. Returns the status of the queue
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/queueStatus'
        '201':
          description: Created. Returns the id of the document and links to generated resources.
          headers:
            Location:
              description: Path to the primary output data
              schema:
                type: string
                example: /api/v1.0/json/{id}
          content:
            application/json:
              schema:
                type: object
                example:
                  {
                    id: '{id}',
                    json: '/api/v1.0/json/{id}',
                    markdown: '/api/v1.0/markdown/{id}',
                    csv: '/api/v1.0/csv/{id}',
                  }
                properties:
                  id:
                    type: string
                  json:
                    type: string
                  xml:
                    type: string
                  markdown:
                    type: string
                  text:
                    type: string
                  confidences:
                    type: string
                  json-compact:
                    type: string
                  csv:
                    type: string
        '404':
          description: Not Found
        '500':
          description: Internal Server Error

  /json/{id}:
    get:
      operationId: getJson
      tags:
        - 'Output'
      summary: 'Get the JSON representation of the document'
      parameters:
        - $ref: '#/components/parameters/id'
      responses:
        '200':
          description: Ok
          content:
            application/json:
              schema:
                $ref: '#/components/schemas/json'
        '404':
          description: Not Found

  #  /xml/{id}:
  #    get:
  #      operationId: getXml
  #      tags:
  #      - "Output"
  #      summary: "Get the XML representation of the document"
  #      parameters:
  #      - $ref: '#/components/parameters/id'
  #      responses:
  #        "200":
  #          description: Ok
  #          content:
  #            application/xml:
  #              schema:
  #                type: array
  #                items:
  #                  $ref: '#/components/schemas/json'
  #                xml:
  #                  name: document
  #        "404":
  #          description: Not Found

  /text/{id}:
    get:
      operationId: getText
      tags:
        - 'Output'
      summary: 'Get the raw text representation of the document'
      parameters:
        - $ref: '#/components/parameters/id'
      responses:
        '200':
          description: Ok
          content:
            text/plain:
              schema:
                type: string
                example: Lorem ipsum dolor sit amet, consectetur adipiscing elit.
        '404':
          description: Not Found

  #  /confidences/{id}:
  #    get:
  #      operationId: getConfidences
  #      tags:
  #      - "Output"
  #      summary: "Get the confidences of every raw text words"
  #      parameters:
  #      - $ref: '#/components/parameters/id'
  #      responses:
  #        "200":
  #          description: Ok
  #          content:
  #            text/plain:
  #              schema:
  #                type: string
  #                example:
  #                  100 98.5 98 95 98.2 90 93 99.8
  #        "404":
  #          description: Not Found

  #  /json-compact/{id}:
  #    get:
  #      operationId: getJsonCompact
  #      tags:
  #      - "Output"
  #      summary: "Get the JSON representation of the document in a more compact format (to be implemented)"
  #      parameters:
  #      - $ref: '#/components/parameters/id'
  #      responses:
  #        "501":
  #          description: Not Implemented

  /markdown/{id}:
    get:
      operationId: getMarkdown
      tags:
        - 'Output'
      summary: 'Get the Markdown representation of the document'
      parameters:
        - $ref: '#/components/parameters/id'
      responses:
        '200':
          description: Ok
          content:
            text/markdown:
              schema:
                type: string
                example: '## Lorem ipsum dolor sit amet, _consectetur adipiscing_ elit.'
        '404':
          description: Not Found

  /csv/{id}:
    get:
      operationId: getCsvList
      tags:
        - 'Output'
      summary: 'Get the list of every CSV file path'
      parameters:
        - $ref: '#/components/parameters/id'
      responses:
        '200':
          description: Ok
          content:
            application/json:
              schema:
                type: array
                items:
                  type: string
                  description: Paths to CSV files
                example: ['/api/v1.0/csv/{id}/1/1', '/api/v1.0/csv/{id}/{page}/{table}']
        '404':
          description: Not Found

  /csv/{id}/{page}/{table}:
    get:
      operationId: getCsv
      tags:
        - 'Output'
      summary: 'Get the CSV representation of a table'
      parameters:
        - $ref: '#/components/parameters/id'
        - name: page
          in: path
          description: Page number
          required: true
          schema:
            type: number
        - name: table
          in: path
          description: Table number
          required: true
          schema:
            type: number

      responses:
        '200':
          description: Ok
          content:
            text/csv:
              schema:
                type: string
                example: |
                  3x4 table;Empty column;Numbers
                  ;;
                  Item A;;3.14
                  "Item B
                  on two lines";;1,234.56
        '404':
          description: Not Found

components:
  parameters:
    id:
      name: id
      in: path
      description: ID of the document
      required: true
      schema:
        type: string

  schemas:
    queueStatus:
      type: object
      properties:
        progress-percentage:
          type: number
          example: 25
        status:
          type: string
          example: 'detecting reading order'
        start-date:
          type: string
          format: date-time
        estimated-remaining-time:
          type: number
          example: 60

    config:
      type: object
      properties:
        version:
          type: number
        extractor:
          type: object
          properties:
            pdf:
              type: string
              enum: ['pdfminer', 'tesseract', 'abbyy']
            img:
              type: string
              enum: ['tesseract', 'abbyy']
            language:
              type: string
        cleaner:
          oneOf:
            - type: string
            - type: object
        output:
          type: object
          properties:
            granularity:
              type: string
              enum: ['word', 'character', 'line', 'paragraph']
            formats:
              type: object
              properties:
                json:
                  type: boolean
                #json-compact:
                #  type: boolean
                text:
                  type: boolean
                markdown:
                  type: boolean
                #xml:
                #  type: boolean
                #confidences:
                #  type: boolean
                csv:
                  type: boolean
                #pdf:
                #  type: boolean
      example:
        verison: '1.0'

    element:
      type: object
      properties:
        id:
          type: number
          xml:
            attribute: true
        type:
          type: string
          enum:
            [
              character,
              drawing,
              heading,
              image,
              list,
              paragraph,
              table,
              table-cell,
              table-row,
              word,
              line,
              barcode,
            ]
          xml:
            attribute: true
        box:
          $ref: '#/components/schemas/box'
        content:
          type: array
          items:
            $ref: '#/components/schemas/element'
    box:
      type: object
      properties:
        l:
          type: number
        t:
          type: number
        w:
          type: number
        h:
          type: number
    json:
      type: object
      properties:
        metadata:
          type: array
          items:
            type: object
            properties:
              id:
                type: number
                xml:
                  attribute: true
              order:
                type: number
        pages:
          type: array
          items:
            type: object
            properties:
              pageNumber:
                type: number
                xml:
                  attribute: true
              box:
                $ref: '#/components/schemas/box'
              elements:
                type: array
                items:
                  $ref: '#/components/schemas/element'
        fonts:
          type: array
          items:
            type: object
            properties:
              id:
                type: number
                xml:
                  attribute: true
              name:
                type: string
              size:
                type: number