Levels of Testing/Verifications
Unit tests : Function-level testing, best to be developed by Viktor after testing-friendly refactorisation of code
Regression tests : Develop an automated e-to-e workflow covering hazard curve, deaggregation, but no
Verification : Produce executive summary of statistics to help researchers to verify (after gaining some confidence, we will complete regression tests)
Feature Matrix / Userclass
| Internal | EAP | Public | |
|---|---|---|---|
| Event search | |||
| Dataset | All Cybershake v18p6 | ||
| single site | O | ||
| multiple site | O | ||
| custom name | O | ||
| Intensity measures (all?) | O | ||
| Download IMs for all sites | O | ||
| Download Acc. time series for all sites | O | ||
Site search | |||
| Data set | All Cybershake v18p6 | ||
| Intensity measures (all?) | O | ||
| NZ code spectra | O | ||
| Disaggregation Exceedance | 50%, 10% 2% | ||
| Disaggregation result | O | ||
| by Type | O | ||
| by Epsilon | O | ||
| by Top contributing faults | O | ||
| Ground Motion Selection |
Deployment & Version control
4 separate servers to be deployed.
- Bleeding Edgy http://hypocentre:5000 (Bleeding Edgy) : Latest code, may be broken.
- Internal http://dev.(awsURL) : Latest master (that passed regression/unit tests), scientific correctness is not guaranteed.
- EAP http://eap.(awsURL) : When ready for sharing with the EA evaluation. At this stage, code change should be kept strictly minimal.
- awsURL (Public): When EA evaluation has reached satisfactory level and is ready for general public release.
