All notable changes to this project will be documented in this file. This project adheres to Semantic Versioning.
- Improved network robustness
- Better error handling when requesting non-existing job status
- Zenodo integration and DOI
- Don't litter slurm-xyz.out files all over the place
- Update Docker to Ubuntu 18.04
- Improved network error handling
- Faster API reinstall
- Security Overly-broad permissions on auto-created remote directories
- Improved network error handling
- Better HPC support * Extra scheduler options (e.g. to select a GPU node) * Allow specifying the number of cores, per step * Allow specifying a time limit, per step and/or per job
- Improved specialisations * Allow multiple projects per specialisation * Versioning of specialisations * Automatic remote installation of updates * Rewritten documentation
- Usability * Improved job logs * Improved server log * Incremental remote job log updates * Log level configurable via environment variable
- Robustness * Automatically reconnect if the network connection drops * Better error-handling and retrying throughout * Better error messages * Database recovery in the event of a crash
- Technical improvements * Faster data transfer * Cleaner code, using Cerulean instead of Xenon 1 * Better code documentation * Improved and cleaned up unit tests * Rebuilt integration test * Type annotations and checking using mypy
Initial release