Skip to main content
Version: 12.x (Current)

Usage

This service exposes several APIs, useful to upload, delete or download the files, described in details in the next paragraphs.

The routes can be prefixed by passing the --prefix option at startup.

Upload: POST /​

With a multipart request containing the file to upload. The file will be stored on the configured backend, and the file information will be stored on MongoDB.

Example of multipart request to upload a single file:

POST / HTTP/1.1
... other headers ...
Content-Type: multipart/form-data; boundary=----WebKitFormBoundaryePkpFF7tjBAqx29L

------WebKitFormBoundaryePkpFF7tjBAqx29L
Content-Disposition: form-data; filename="my-file-1.pdf"
Content-Type: application/pdf

... file content here ...
------WebKitFormBoundaryePkpFF7tjBAqx29L--
Content-Disposition: form-data; name="key-1"
Content-Type: text/plain

value-1
------WebKitFormBoundaryePkpFF7tjBAqx29L--

As it is possible to notice:

  • The first part is a *.pdf file named my-file-1.pdf

  • The second part is a plain text part with name="key-1".
    "key-1" will be the key of a field that will be saved inside the file document in files collection. "value-1" will be the value of the that field.

Continuing the example above, once the file is stored, the respective document in the files collection might be something like this:

{
  //Basic file data
  "_id" : "61c32e38e67eb6fc92ed50b9",
  "name" : "my-file-1.pdf",
  "file" : "078b04fa-6935-4686-ac5c-75f6b1ff5401.txt",
  "size" : 768,
  "location" : "/v2/files/download/078b04fa-6935-4686-ac5c-75f6b1ff5401.txt",
  "__STATE__" : "PUBLIC",
  "creatorId" : "public",
  "updaterId" : "public",
  //Specific metadata for the file
  "key-1": "value-1"
}
info

In the example described above, to be able to upload the additional metadata key-1, it is important to add a caster file and to specify its path in the environment variable ADDITIONAL_FUNCTION_CASTER_FILE_PATH. Without specifying a caster file, every additional metadata will be ignored.

caution

At the moment, the file part must be the last one of the multipart. Any other part that follows the file part won't be considered.

info

This API supports only one file per request. If the request contains more files, then only the first one will be uploaded. Consider using POST /bulk to upload multiple files at once.

Response​

In case of success, the response is a JSON with the following fields.

  • _id: unique MongoDB identifier.
  • name: original file name.
  • file: unique name of the file that should be used to retrieve it using this service.
  • size: size in bytes of the uploaded file.
  • location: the URL that can be used to download the file using the same service that performed the upload.

Upload many: POST /bulk​

With a multipart/form-data request containing all the files to upload and the respective metadata. All the files will be stored on the configured backend, and the files information will be stored on the configured CRUD. The order of the files inside the multipart is not important and files can have the same filename field.
The upload of multiple files is handled as an atomic operation, meaning that either all files are stored correctly or none of them are.

Consider the following simplified example of a HTTP multipart request:

POST /bulk HTTP/1.1
... other headers ...
Content-Type: multipart/form-data; boundary=----WebKitFormBoundaryePkpFF7tjBAqx29L

------WebKitFormBoundaryePkpFF7tjBAqx29L
Content-Disposition: form-data; filename="my-file-1.pdf"
Content-Type: application/pdf

... file content here ...
------WebKitFormBoundaryePkpFF7tjBAqx29L
Content-Disposition: form-data; filename="my-file-1.pdf"
Content-Type: application/pdf

... file content here ...
------WebKitFormBoundaryePkpFF7tjBAqx29L--

As it can be noticed, two files are being uploaded, both having the same filename field set tofilename="my-file-1.pdf".

Apart from the multiple files, the multipart request accepts other parts useful to provide additional data to be saved the in the file document stored in the files collection; please notice that these parts must have application/json content-type.
In the following we describe with an example how to associate some structured metadata to a file
Consider the following example of multipart request:

POST /bulk HTTP/1.1
... other headers ...
Content-Type: multipart/form-data; boundary=----WebKitFormBoundaryePkpFF7tjBAqx29L

------WebKitFormBoundaryePkpFF7tjBAqx29L
Content-Disposition: form-data; filename="my-file-1.pdf"
Content-Type: application/pdf

... file content here ...
------WebKitFormBoundaryePkpFF7tjBAqx29L--
Content-Disposition: form-data; name="my-file1.pdf"
Content-Type: application/json

{
  "anAdditionalProperty": "a_string_value",
  "anotherAdditionalProperty": 34.12
}
------WebKitFormBoundaryePkpFF7tjBAqx29L--

As it is possible to notice:

  1. The first part is a file named my-file-1.pdf.

  2. The second part is a text part of type application/json with name whose value is the name of the file which will receive the metadata contained in the part.

Continuing the example above, once the file is stored, the respective document in the files collection might be something like this:

{
  //Basic file data
  "_id" : "61c32e38e67eb6fc92ed50b9",
  "name" : "my-file-1.pdf",
  "file" : "078b04fa-6935-4686-ac5c-75f6b1ff5401.pdf",
  "size" : 768,
  "location" : "/v2/files/download/078b04fa-6935-4686-ac5c-75f6b1ff5401.pdf",
  "__STATE__" : "PUBLIC",
  "creatorId" : "public",
  "updaterId" : "public",
  //Specific metadata for the file
  "anAdditionalProperty": "a_string_value",
  "anotherAdditionalProperty": 34.12
}
caution

Keep in mind the fact that File Service uses a CRUD service to store file documents in the files collection, hence you'll need to register all the possible keys that will be stored at the first level of the file document in the files CRUD collection definition.

info

File info still pass through the function that the user can specify by means of the ADDITIONAL_FUNCTION_CASTER_FILE_PATH environment variable before being passed to the CRUD service.

Response​

At the moment, as stated earlier, the upload of multiple files at once is handled as an atomic operation, meaning that either all of the files are stored properly or none of them are.

In the case of success, a list with stored files' data is returned.
In the event of an error while storing one of the files, the service ensures that none file is stored, and an error is returned to the caller.

Continuing the last example of upload, in the case of success, the service answers with 200 and the following application/json response:

[
  {
    "_id": "61c338e6e67eb6fc92ed50ba",
    "name": "my-file-1.pdf",
    "file": "02aa2106-661d-4fcc-9299-c72884ad0f28.pdf",
    "size": 768,
    "location": "/v2/files/download/02aa2106-661d-4fcc-9299-c72884ad0f28.pdf",
    "anAdditionalProperty": "a_string_value",
    "anotherAdditionalProperty": 34.12
  }
]

Consider consulting the Open API definition of the service to get more information about the service responses.

Download: GET /download/:file​

To download a single file that was previously uploaded, specifying its file field.

Query parameters​

  • download=1: to download the file as an attachment.
  • downloadWithOriginalName=1: to download the file with its original name as attachment.
  • useOriginalName=1: to get the file with its original name.
info

The GET /download/:file API, when used as link in the browser has different behaviors based on these query parameters; in particular with one of the download* query parameters the file is downloaded as attachment. Instead, the use* query parameter opens the requested file in the browser window.

This rule applies for all files except for HTML which are always downloaded preventing the browser from rendering them.

This API also supports a multi-bucket configuration. In this case, the scope parameter should be added to the path, to select the specific bucket instance on which to operate (GET /scopeValue/download/:file, where scopeValue is the name of the scope).

Download many: GET /download​

To download a ZIP archive containing all the files specified in the query parameters.

Query parameters​

  • files required: array of strings (or single string) specifying the file field of the desired files to download
  • downloadWithOriginalName=1: to download the files with their original names.

In case of a multiple-bucket files-service, the scope parameter should be added to the path to select the specific bucket instance on which to operate (GET /scopeValue/download, where scopeValue is the name of the scope).

Response​

At the moment the download of multiple files is handled as an atomic operation, meaning that the API returns either all requested files in a (streamed) ZIP archive or an error.

Get files list: GET /files/​

Get a list of files metadata managed by the Files Service in the bucket; the API supports pagination via query parameters.

Please note that this API is available from v2.7.0

Query parameters​

  • limit: integer value specifying the number of documents to be retrieved per request. Default to 25.
  • page: integer value representing the index of the requested page (starting from 1).
  • dateFrom: date value used to filter the files. Files older than this value will not be shown.

In the multi-bucket configuration, the scope parameter should be added to the path to select the specific bucket instance on which to operate (GET /scopeValue/files/, where scopeValue is the name of the scope).

Response​

List of file metadata objects as retrieved from the Crud Service that match query parameters (if present).

Update file metadata: PATCH /:filename/metadata​

This API allow to modify the metadata of a specific file. Only custom metadata fields, set with a caster file, can be modified.

Please note that this API is available from v2.10.0

The body must be a JSON containing the new values you want to set to the metadata.

caution

The endpoint only modifies custom metadata fields defined in the caster file additionalPropertiesValidator. The other fields won't be modified, and silently ignored if added to the body.

danger

Make sure to always specify all the metadata fields if you want to modify only some of them. Any field not specified may be removed.

In the multi-bucket configuration, the scope parameter should be added to the path to select the specific bucket instance on which to operate (PATCH /scopeValue/:filename/metadata, where scopeValue is the name of the scope).

Example​

Consider a configuration with the following caster file:

'use strict'

module.exports = function caster(doc) {
  return {
      metadata: {
        tags: (doc.tags || '').split(','),
        authorId: doc.authorId || undefined,
        ownerId: doc.ownerId || undefined,
      }
  }
}

module.exports.additionalPropertiesValidator = {
  tags: { type: 'string' },
  authorId: { type: 'string' },
  ownerId: { type: 'string' },
}

The collection contains this document:

{
  "_id" : "61c32e38e67eb6fc92ed50b9",
  "name" : "my-file-1.pdf",
  "file" : "078b04fa-6935-4686-ac5c-75f6b1ff5401.txt",
  "size" : 768,
  "location" : "/v2/files/download/078b04fa-6935-4686-ac5c-75f6b1ff5401.txt",
  "__STATE__" : "PUBLIC",
  "creatorId" : "public",
  "updaterId" : "public",
  "metadata": {
    "authorId": "someOldVal",
    "ownerId": "someOtherOldVal",
    "tags": "some,tags"
  }
}

You can update the authorId and the ownerId with the following http call:

curl http://my-files-service/filesMetadata \
    -X PATCH \
    -raw-data '{
       "authorId": "someNewValue",
       "ownerId": "someOtherNewValue"
     }'

In this case, the object returned by the caster function will replace the specified metadata CRUD field. For this reason, any field not specified in the body will be undefined in the resulting document. The endpoint will return 204 No Content in case of successful operation.

This is the resulting document after the PATCH:

{
  "_id" : "61c32e38e67eb6fc92ed50b9",
  "name" : "my-file-1.pdf",
  "file" : "078b04fa-6935-4686-ac5c-75f6b1ff5401.txt",
  "size" : 768,
  "location" : "/v2/files/download/078b04fa-6935-4686-ac5c-75f6b1ff5401.txt",
  "__STATE__" : "PUBLIC",
  "creatorId" : "public",
  "updaterId" : "public",
  "metadata": {
    "authorId": "someNewValue",
    "ownerId": "someOtherNewValue"
  }
}

Response​

List of file metadata objects as retrieved from the Crud Service that match query parameters (if present).

Delete: DELETE /:file​

To delete a file that was previously uploaded.

In the multi-bucket configuration, the scope parameter should be added to the path to select the specific bucket instance on which to operate (DELETE /scopeValue/:file, where scopeValue is the name of the scope).