...
Aka the first batch of faults for the South Island. This version was archived on the RCC legacy storage in a very naive way. Every fault was saved in a single tar file. Thus every tar file has all the realizations, and for each realization we have a LF, HF and BB. This archived version needs to be cleaned to reduce the file size and have easier access to the waveforms.
...
This is the batch that completes all the inland faults on the South Island. Note that both results need to be aggregated to have everything. To archive this version, we decided that we would zip the three components of the waveform for a given station location into a single file (this means we will have as many .zip files as locations in the simulation). The way this was archived is very simple as it can be shown in the following block:
...
This means that every fault has the realizations. For each realization we have an Acc, Vel and xyts directories, where we have I copied the zipped accelerations, zipped velocities and e3d files respectively. Please note that only the Acc and Vel files are zipped. There is also a simple metadata.json file and the IM database for that particular version is also archived. The metadata file has the following contents:
...
This file was generated manually, but we could have some way to automate the generation after the archiving has finishedit should be automatically generated by the archiving process.
v18p5
This version completes the NZ faults by adding the North Island faults. This is the next archiving to be done, and it will follow the same structure as v18p4.
TODO
- Add symlinks symbolic links to v18p4 to give the impression that it is actually v17p9 + more faults. The same will be needed for v18p5.
Future
- We will use the binary workflow work-flow that Viktor has developed
- Metadata Meta-data should be easier to generate
- IM database may not exist anymore
- We need to provide a transparent way to clearly combine several events. For example, if v18p6 replaces some of the results from v17p9, this should be reflected somehow in the archive.
...