Use the following commands as test scripts
cd test
cl upload data
cl ancestors data
cd test
cl upload data
cl upload code
cl run :data :code 'python code/sort.py < data/lines.txt'
cl ancestors run-python
cl ancestors 0xdbb9daec17bb4964bc86a559c0927464
cl ancestors run_python
The goal of CodaLab Worksheets is to faciliate transparent, reproducible, and collaborative research in computation- and data-intensive areas such as machine learning.
The CodaLab Bundle Service allows users to create bundles, which are immutable directories containing code or data. Bundles are either uploaded or created from other bundles by executing arbitrary commands. When the latter happens, all the provenance information is preserved. In addition, users can create worksheets, which interleave bundles with free-form textual descriptions, allowing one to easily describe an experimental workflow.
The CodaLab frontend holds the React front-end web interface for CodaLab worksheets.
To get started visit the official CodaLab Worksheets instance For more information about the platform, visit our Wiki
If you're interested in contributing or setting up your own CodaLab Worksheets instance, get in touch with us
- Official CodaLab Worksheets Instance: live instance of CodaLab Worksheets
- CodaLab Worksheets Wiki: all documentation
We provide a convenience script that uses docker-compose to bring up a full fledged CodaLab server in this repo.
To get started, make sure you have a recent version of Docker and docker-compose installed.
Simply run ./codalab_service.py start -i to bring up a fresh instance of CodaLab with the default configuration.
If you've made local changes to the codebase and would like to rebuild docker images from the current state of the codebase,
use ./codalab_service.py start -b.
If you've previously started an instance and thus have the database and root account initialization done, omit the -i flag to just bring the service up like so: ./codalab_service.py start
For more information, check out the Wiki page