Resumable File Upload

Notes for resumable file upload.

Resumable File Upload

This document describes how to upload a large file to Seafile in chunks.

Examples below assume the Seafile service is available at http://{server}.

Workflow

1. Call the file-uploaded-bytes API to check whether resumable upload is supported and to get the current uploaded offset.
2. Get a normal upload link.
3. Upload the file chunk by chunk to the upload link with the required headers.

Check Support and Uploaded Bytes

Use the following API before starting or resuming a chunked upload.

GET http://{server}/api/v2.1/repos/{repo_id}/file-uploaded-bytes/?parent_dir={parent_dir}&file_name={file_name}

Request parameters

• repo_id
• parent_dir
• file_name

Sample request

Bash

curl -v \
  -H 'Authorization: Token e71c00e93af863ba9bcddb61a46bb4de11d713fc' \
  -H 'Accept: application/json; charset=utf-8; indent=4' \
  "http://{server}/api/v2.1/repos/09b7d3c0-5f0d-49be-9318-7ca136f386cd/file-uploaded-bytes/?parent_dir=/path-in-seafile-repo/&file_name=test.md"

If the response contains the header Accept-Ranges: bytes, the Seafile server supports resumable upload.

The response body contains the number of bytes already stored on the server for this file.

Sample response

If the file has not been uploaded before:

JSON

{
  "uploadedBytes": 0
}

If 149946368 bytes have already been uploaded:

JSON

{
  "uploadedBytes": 149946368
}

Use uploadedBytes as the starting offset for the next chunk.

Errors

• 400 parent_dir or file_name is invalid.
• 404 Library or folder not found.
• 500 Internal server error.

Get Upload Link

Get the upload link in the same way as a normal file upload. Then use that link for all chunk upload requests.

Upload File in Chunks

A large file is uploaded as multiple chunks. The chunk size is chosen by the client. Chunks must be uploaded in order. After the last chunk is uploaded, the server indexes the full file and saves it.

After getting the upload link and uploadedBytes, POST each chunk to the upload link.

POST http://{server}/seafhttp/upload-api/73c5d117-3bcf-48a0-aa2a-3f48d5274ae3

Request parameters

The multipart form fields are the same as for a normal file upload, but chunked upload requires two extra headers:

• Content-Range: the byte range of the current chunk. Refer to https://developer.mozilla.org/en-US/docs/Web/HTTP/Headers/Content-Range for details.
• Content-Disposition: attachment; filename="filename.txt": in chunked upload, the server reads the file name from this header instead of from the file form field.

Unlike a normal upload, chunked upload supports only one file per request.

Sample request

Upload a chunk to /path-in-seafile-repo/. If test.md already exists there, replace it.

Bash

curl \
  -H "Content-Range: bytes 149946368-150994943/1587609600" \
  -H "Content-Disposition: attachment; filename=\"test.md\"" \
  -F [email protected] \
  -F parent_dir=/path-in-seafile-repo/ \
  -F replace=1 \
  http://{server}/seafhttp/upload-api/73c5d117-3bcf-48a0-aa2a-3f48d5274ae3

In this example:

• 149946368-150994943 is the byte range of the current chunk.
• 1587609600 is the total file size in bytes.

Sample response

If the uploaded chunk is not the last chunk, the server returns:

JSON

{
  "success": true
}

If the uploaded chunk is the last chunk, the server returns the same response as a normal file upload.

Errors

• 400 Bad request.
• 403 Permission denied.
• 440 Invalid filename.
• 442 File size is too large.
• 443 Out of quota.
• 500 Internal server error.