Skip to content

Latest commit

 

History

History
141 lines (98 loc) · 3.12 KB

README.md

File metadata and controls

141 lines (98 loc) · 3.12 KB
Raw

Zenodo Uploader

Zenodo Uploader is a small application to efficiently upload large number of records and files to Zenodo via the Deposit API.

If you plan to make large uploads to Zenodo, please get in touch with us before hand at https://zenodo.org/support

Prerequisites

  • Docker and Docker-Compose
  • Python 3
  • virtulaenv-wrapper (for the mkvirtualenv command)

Install

$ git clone https://github.com/lnielsen/zenodo-uploader.git
$ cd zenodo-uploader
$ mkvirutalenv uploader
(uploader)$ pip install -r requirements.txt

Configure

You must edit config.py:

ACCESS_TOKEN = '...'
ENDPOINT = 'https://sandbox.zenodo.org/api/deposit/depositions'

Run

Start the RabbitMQ message broker (required by the worker):

$ docker-compose up -d

Note that the message broker have an admin interface you can see on http://localhost:15672 (login with guest/guest).

Initialize (or reset) the log database:

(uploader)$ python app.py init

This will create a SQLite3 database in the current working directory called log.db.

Finally start the Celery worker:

(uploader)$ python app.py worker

Note, if you make changes to the code, you'll have to restart the worker for the code changes to be picked up.

Usage

Dry-run

Process bulk upload (in dry-run mode which checks the local data):

(uploader)$ python app.py process -n testupload2/csvdir/

This will expect to find a directory with CSV files. Each CSV file is read, and each row is converted into an upload to Zenodo. See utils2.py for details on how these uploads are generated from the CSV (e.g. to fix metadata).

You'll need to configure the FILES_PATH_TEMPLATES in config.py. This option configures where to look for files for each upload.

Process

Process bulk upload (for real)

(uploader)$ python app.py process testupload2/csvdir/

You should now see the worker processing tasks.

Inspecting the logs

List all deposits:

(uploader)$ python app.py logs deposits
Deposit ID,Upload ID,Done
256643,testlin#BR0000010599976,True

List errors:

(uploader)$ python app.py logs errors
testlin#BR0000010599976;create;400;{"status": 400, "message": "Validation error.", "errors": [{"field": "metadata.communities", "message": "Invalid communities: belgiumherbarium"}]}

List request performance:

(uploader)$  python app.py logs timing
Type,Time,Bytes
Timestamp,Type,Time,Bytes
2019-01-18 16:14:03.193291,create,0.5013158321380615,0
2019-01-18 16:14:07.530581,upload,4.285969257354736,3692386
2019-01-18 16:14:08.946722,publish,1.360076904296875,0

List request performance (filter by type):

(uploader)$ python app.py logs timing -t upload
Type,Time,Bytes
upload,0.4276392459869385,3692386

You can reset the database at any point by running:

(uploader)$ python app.py init

Querying the log database

If you need more advanced querying of the database, you can simply open it in SQLite3:

$ sqlite3 log.db
SQLite version 3.24.0 2018-06-04 14:10:15
Enter ".help" for usage hints.
sqlite>