CouchDB Download Token Service

Documentation Status Code coverage Code coverage

Simple web service that allows downloading CouchDB document attachments with a per-document download token.

This project was created with the objetive of provide GET URLs for CouchDB document attachments without having specify the couch username and password in the URL itself or in the request headers.

The method I found to do this worrying about permissions is assigning each document a secret download_token field to each document that will have to match when doing requests to this service.

The service needs to have admin user credentials so it can query all document’s attachments and decide which clients give then. To do this you have to set a COUCHDB_URL environment variable of format http://admin_user:password@server_domain:5984. By default it uses localhost with no user credentials.

Example usage

$ pip3 install couchdb_download_token
$ pip3 install gunicorn
$ gunicorn couchdb_download_token:api
[2016-07-12 23:33:28 -0300] [7880] [INFO] Starting gunicorn 19.6.0
[2016-07-12 23:33:28 -0300] [7880] [INFO] Listening at: (7880)
[2016-07-12 23:33:28 -0300] [7880] [INFO] Using worker: sync
[2016-07-12 23:33:28 -0300] [7884] [INFO] Booting worker with pid: 7884
$ curl http://localhost:5984/my_database/my_document
    "_id": "my_document",
    "_rev": "4-763e041701ae3e55fd4af08dff93efc4",
    "info": "test document",
    "download_token": "123456",
    "_attachments": {
        "file.txt": {
            "content_type": "text/plain",
            "revpos": 2,
            "digest": "md5-U0f+Rrm7WPnsUGK3oD8t8g==",
            "length": 100,
            "stub": true
$ curl http://localhost:8000/my_database/my_document/file.txt?token=incorrect
< HTTP/1.1 403 Forbidden
< Server: gunicorn/19.6.0
$ curl http://localhost:8000/my_database/my_document/file.txt?token=123456
< HTTP/1.1 200 OK
< Server: gunicorn/19.6.0
< Date: Wed, 13 Jul 2016 03:02:07 GMT
< Connection: close
< Transfer-Encoding: chunked
< content-type: text/plain
File contents


This package was created with Cookiecutter and the audreyr/cookiecutter-pypackage project template.