University of Chicago data API

openapi: 3.0.0
info:
  title: Fence OpenAPI Specification admin/user data API
  version: 0.1.0
  description: Access management for Gen3 data commons. Code is available on [GitHub](https://github.com/uc-cdis/fence).
  termsOfService: http://cdis.uchicago.edu/terms/
  contact:
    email: cdis@uchicago.edu
  license:
    name: Apache 2.0
    url: http://www.apache.org/licenses/LICENSE-2.0.html
servers:
- url: https://example.domain/
tags:
- name: data
  description: Generate signed URLs
paths:
  /data/{file_id}:
    delete:
      tags:
      - data
      summary: delete all locations of a stored data file and remove its record from indexd
      description: After a user uploads a data file and it is registered in indexd (before or after it is mapped into the graph via metadata submission), this endpoint will delete the file from its storage locations (saved in the record in indexd) and delete the record in indexd.
      security:
      - OAuth2:
        - data
      responses:
        204:
          description: successful operation
        500:
          description: internal server error; could not delete stored files, or not able to delete indexd record
          content-type: application/json
  /data/download/{file_id}:
    get:
      tags:
      - data
      summary: Create a signed URL for data download given a file_id
      description: Allow users to get a signed URL for a data file.
      security:
      - OAuth2:
        - user
      operationId: downloadSignedURL
      parameters:
      - name: file_id
        required: true
        in: path
        description: data UUID
        schema:
          type: string
      - name: protocol
        required: false
        in: query
        description: a protocol provided by storage provider, e.g. http, ftp, s3, gs
        schema:
          type: string
          enum:
          - http
          - s3
          - gs
      - name: expires_in
        required: false
        in: query
        description: the time (in seconds) in which return url is valid. Must be less than the configured maximum (default is 3600). If it's greater, the configured maximum will be used.
        schema:
          type: integer
      - name: redirect
        required: false
        in: query
        description: if set, then a redirect to the signed url will be sent. Otherwise, json data with the url is returned.
        schema:
          type: boolean
      - name: no_force_sign
        required: false
        in: query
        description: if `no_force_sign=True`, this will request to *not* sign the resulting URL (i.e. just provide the public url without using anonymous signing creds).
        schema:
          type: boolean
      - name: userProject
        required: false
        in: query
        description: a Google Project to bill for accessing data in requester pays buckets in Google Storage. Will override any configured default billing projects. If Fence is configured to automatically attempt to provide the necessary service account(s) billing permission, the provided Google Project will need to have given the Fence admin service account necessary permissions to create custom roles and set Project IAM policies. Please see README for more information.
        schema:
          type: string
      responses:
        '200':
          description: successful operation
          content:
            '*/*':
              schema:
                $ref: '#/components/schemas/SignedURL'
        '400':
          description: 'Invalid input: UUID not found or invalid location'
        '404':
          description: No location found for this file
  /ga4gh/drs/v1/objects/{object_id}/access/{access_id}:
    get:
      summary: GA4GH DRS Access API to get a URL for fetching bytes.
      description: Returns a URL that can be used to fetch the bytes of a DrsObject. This method only needs to be called when using an AccessMethod that contains an access_id (e.g., for servers that use signed URLs for fetching object bytes).
      operationId: GetAccessURL
      responses:
        '200':
          description: The access URL was found successfully.
          content:
            '*/*':
              schema:
                $ref: '#/components/schemas/AccessURL'
        '400':
          description: The request is malformed.
        '401':
          description: The request is unauthorized.
        '404':
          description: The requested access URL wasn't found
        '403':
          description: The requester is not authorized to perform this action.
        '500':
          description: An unexpected error occurred.
      parameters:
      - name: object_id
        schema:
          type: string
        in: path
        required: true
        description: An id of a DrsObject
      - name: access_id
        schema:
          type: string
        in: path
        required: true
        description: An access_id from the access_methods list of a DrsObject
      tags:
      - data
    post:
      summary: GA4GH DRS Access API to get a URL for fetching bytes.
      description: Returns a URL that can be used to fetch the bytes of a DrsObject. This method only needs to be called when using an AccessMethod that contains an access_id (e.g., for servers that use signed URLs for fetching object bytes).
      operationId: PostAccessURL
      requestBody:
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/RequestUrlPassport'
      responses:
        '200':
          description: The access URL was found successfully.
          content:
            '*/*':
              schema:
                $ref: '#/components/schemas/AccessURL'
        '400':
          description: The request is malformed.
        '401':
          description: The request is unauthorized.
        '404':
          description: The requested access URL wasn't found
        '403':
          description: The requester is not authorized to perform this action.
        '500':
          description: An unexpected error occurred.
      parameters:
      - name: object_id
        schema:
          type: string
        in: path
        required: true
        description: An id of a DrsObject
      - name: access_id
        schema:
          type: string
        in: path
        required: true
        description: An access_id from the access_methods list of a DrsObject
      tags:
      - data
  /data/upload:
    post:
      tags:
      - data
      summary: Create a new blank record in indexd to use for uploading a data file less than 5GB. For files larger than 5GB, please use multipart upload presigned url
      description: 'This is the first step on the API side for the data upload flow.  This

        endpoint causes fence to make a request to indexd to create a new, blank

        index record, and returns the GUID for this new record and a presigned

        URL which a client can then use to upload their data file directly to a

        storage bucket. No parameters are accepted because fence retrieves the

        username from the current token to send to indexd for the `uploader`

        field in the new record.


        This API also supports utilizing the newer authorization within the indexing

        service by providing an "authz" field.


        If an "authz" field is provided, the authorization checks are slightly different:

        your user must have *both* "create" and "write-storage" permission on the resources

        you are including in the "authz" list.


        In addition to a different authorization check, when "authz" is provided the

        "uploader" field in indexd with *not* be populated.


        Previous authorization check requires a more general, global upload permission:

        "file_upload" on "/data_file" resource. When "authz" is *not* provided, this

        endpoint will check for that permission for your user.


        Accepts a "guid" field in the request body. If "guid" is provided, it checks indexd for an existing record. If not found, it raises a 404.

        '
      security:
      - OAuth2:
        - user
        - data
      operationId: uploadBlank
      requestBody:
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/RequestUploadBlank'
      responses:
        201:
          description: successful operation; created new record in indexd
          content:
            application/json:
              schema:
                type: object
                properties:
                  guid:
                    type: string
                    description: the GUID from the newly created record in indexd
                  url:
                    type: string
                    description: the presigned URL usable for data upload
        404:
          description: Record with <guid> not found.
  /data/upload/{file_id}:
    get:
      tags:
      - data
      summary: Create a signed URL for data upload specified by file_id
      description: Allow users to get a signed URL to upload data given the location of preference.
      security:
      - OAuth2:
        - user
      operationId: uploadSignedURL
      parameters:
      - name: file_id
        required: true
        in: path
        description: data UUID
        schema:
          type: string
      - name: protocol
        required: false
        in: query
        description: a protocol provided by storage provider, e.g. http, ftp, s3, gs
        schema:
          type: string
          enum:
          - http
          - s3
      - name: expires_in
        required: false
        in: query
        description: the time (in seconds) in which return url is valid. Must be less than the configured maximum (default is 3600). If it's greater, the configured maximum will be used.
        schema:
          type: integer
      - name: file_name
        required: false
        in: query
        description: the requested file name in the cloud bucket you will upload to. If not provided, will use the GUID/file_id
        schema:
          type: string
      - name: bucket
        required: false
        in: query
        description: the requested bucket to upload to. If not provided, defaults to the configured DATA_UPLOAD_BUCKET.
        schema:
          type: string
      responses:
        '200':
          description: successful operation
          content:
            '*/*':
              schema:
                $ref: '#/components/schemas/SignedURL'
        '400':
          description: 'Invalid input: UUID not found or invalid location'
  /multipart/init:
    post:
      tags:
      - data
      summary: Initilize a multipart upload for AWS bucket. It returns an uploadId which can be used for uploading object parts (see /multipart/upload) and completing the upload (see /multipart/complete)
      description: 'For uploading the big file with the size is larger than 5GB in data upload flow, Fence needs to provide a list of endpoints for supporting multipart upload presigned url This is the first step on the API side for the multipart upload presigned url. This endpoint causes fence to make a request to indexd to create a new, blank index record, and returns the GUID for this new record and an uploadId for multipart upload presigned url.


        Accepts a "guid" field in the request body. If "guid" is provided, it checks indexd for an existing record. If not found, it raises a 404.'
      security:
      - OAuth2:
        - user
        - data
      operationId: multipartUploadInit
      requestBody:
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/InitMultipartUpload'
      responses:
        201:
          description: successful operation; created new blank record in indexd and uploadId for multipart upload
          content:
            application/json:
              schema:
                type: object
                properties:
                  guid:
                    type: string
                    description: the GUID from the newly created record in indexd
                  uploadId:
                    type: string
                    description: the uploadId for multipart upload presigned URL usable for data upload
        404:
          description: Record with <guid> not found.
  /multipart/upload:
    post:
      tags:
      - data
      summary: Upload an object part given uploadId
      description: This is the second step on the API side for the multipart upload presigned url. The object is chopped into multiple parts with part number start from 1. Each part is submitted to the cloud resource with this endpoint by providing the part number a long with uploadId. It is worth to mention that the minimal supported part size is 5 MB as describe in https://docs.aws.amazon.com/AmazonS3/latest/dev/qfacts.html
      security:
      - OAuth2:
        - user
        - data
      operationId: multipartUpload
      requestBody:
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/RequestMultipartUpload'
      responses:
        200:
          description: successful operation
          content:
            application/json:
              schema:
                type: object
                properties:
                  presigned_url:
                    type: string
                    description: the presigned URL for uploading part
  /multipart/complete:
    post:
      tags:
      - data
      summary: Complete aws multipart upload
      description: This is the last step for the multipart upload presigned url. All the parts which were submitted need to be combined together. This enpoint takes a list of the part info (partNumber, Etag) and uploadId in order to finish the upload
      security:
      - OAuth2:
        - user
        - data
      operationId: multipartUploadComplete
      requestBody:
        content:
          application/json:
            schema:
              $ref: '#/components/schemas/CompleteMultipartUpload'
      responses:
        200:
          description: successful operation
  /data/buckets:
    get:
      tags:
      - data
      summary: Get list of buckets and relevant bucket info
      description: Returns a list of buckets and relevant bucket information provided in the fence-config under AWS_BUCKETS and GS_BUCKETS. It filters out sensitive information about the bucket and only shows information like bucket region.
      operationId: getBuckets
      responses:
        '200':
          description: successful operation
          content:
            '*/*':
              schema:
                $ref: '#/components/schemas/BucketInfo'
components:
  schemas:
    RequestUploadBlank:
      type: object
      required: null
      properties:
        file_name:
          type: string
          required: true
          description: the file name to use for this upload
        bucket:
          type: string
          required: true
          description: the requested bucket to upload to. If not provided, defaults to the configured DATA_UPLOAD_BUCKET.
        guid:
          type: string
          required: false
          description: GUID to be assigned to the object
        expires_in:
          type: integer
          description: optional integer specifying the presigned URL lifetime
        authz:
          type: array
          items:
            type: string
          description: requested authorization resources to be set on the resulting indexed record. You must have proper authorization to set this
      example:
        guid: 123456abcd
        file_name: my_file.bam
        bucket: bucket1
        expires_in: 1200
        authz:
        - /programs/A
    BucketInfo:
      type: object
      properties:
        S3_BUCKETS:
          type: object
          description: list and info on S3 buckets configured in the fence config
          properties:
            bucket1:
              type: object
              properties:
                region:
                  type: string
            bucket2:
              type: object
              properties:
                region:
                  type: string
        GS_BUCKETS:
          type: object
          description: list and info on GS buckets configured in the fence config
          properties:
            bucket1:
              type: object
              properties:
                region:
                  type: string
            bucket2:
              type: object
              properties:
                region:
                  type: string
    AccessURL:
      type: object
      required:
      - url
      properties:
        url:
          type: string
          description: A fully resolvable URL that can be used to fetch the actual object bytes.
    RequestUrlPassport:
      type: object
      required: true
      properties:
        passports:
          type: array
          items:
            type: string
          required: true
          description: The list of GA4GH passports containing the authorizations for a given user to access the requested file. NOTE - the key "passports" is configurable in fence via the GA4GH_DRS_POSTED_PASSPORT_FIELD field. The default of "passports" has been provided.
    SignedURL:
      type: object
      properties:
        url:
          type: string
          description: the signed url issued to
    CompleteMultipartUpload:
      type: object
      required: null
      properties:
        key:
          type: string
          description: the key of the object in the format of GUID/filename
        uploadId:
          type: string
          description: the uploadId for multipart presigned URL upload usable for data upload
        parts:
          type: array
          items:
            type: object
            properties:
              partNumber:
                type: integer
                description: The part number
              ETag:
                type: string
                description: the etag that cooresponds to the the object part with the partnumber
          description: List of part numbers info needed for multipart completion
        expires_in:
          type: integer
          description: optional integer specifying the presigned URL lifetime
        bucket:
          type: string
          required: false
          description: bucket to upload to
    RequestMultipartUpload:
      type: object
      required: null
      properties:
        key:
          type: string
          description: the key of the object in the format of GUID/filename
        uploadId:
          type: string
          description: the uploadId for multipart presigned URL upload usable for data upload
        partNumber:
          type: integer
          description: the part number of the part (start from 1)
        expires_in:
          type: integer
          description: optional integer specifying the presigned URL lifetime
        bucket:
          type: string
          required: false
          description: bucket to upload\ to
    InitMultipartUpload:
      type: object
      required: null
      properties:
        file_name:
          type: string
          required: true
          description: the file name to use for this upload
        guid:
          type: string
          required: false
          description: GUID to be assigned to the object
        expires_in:
          type: integer
          description: optional integer specifying the presigned URL lifetime
        authz:
          type: array
          items:
            type: string
          description: requested authorization resources to be set on the resulting indexed record. You must have proper authorization to set this
        bucket:
          type: string
          required: false
          description: bucket to upload to
      example:
        file_name: my_file.bam
        expires_in: 1200
        authz:
        - /programs/A
        bucket: bucket-1
  securitySchemes:
    OAuth2:
      type: oauth2
      flows:
        authorizationCode:
          authorizationUrl: /oauth/authorize
          tokenUrl: /oauth/token
          scopes:
            user: generic user access