Investigate if loading of Muon log data can be improved

[Anders Markvardsen]

For example, in LoadMuonNexus2 investigate if we can swap LoadMuonNexus2::loadLogs with equivalent code in LoadNexusLogs. Also, noted that there is a class LoadMuonLog, which appears from its description to possible do the same job as LoadNexusLogs.

This ticket is to investigate the code of LoadNexusLogs, LoadMuonLog and LoadMuonNexus/LoadMuonNexus2 to see if there is dublication of code which respect to the loading of log data.

[Karl Palmen]

I've investigated the possibility of using the LoadNexusLogs algorithm to do the work of the LoadMuonLog algorithm.

Firstly, LoadNexusLogs uses the new Nexus interface while LoadMuonLog uses the old nexus interface, so it is not advisable to use code from LoadMuonLog in LoadNexusLog without first converting it to use the new Nexus interface.

I've found that the logs have a different structure in the muon Nexus files to the neutron Nexus files. The neutron nexus logs are nested one level deeper then the muon logs and have components "time" and "value", while the muon logs has "name", "time" and "values". Furthermore, the neutron Nexus "time" has an attribute "start", which is the start time.

This attribute "start" has made it very hard to adapt LoadNexusLog to load muon logs (which don't have such a "start") and it turns out about 60 lines of code from LoadMuonLog would need to be converted to use the new Nexus interface and copied into LoadNexusLogs.

This suggests that this would not be worthwhile doing, until we change all the loadNexus code to use the new interface.

[Karl Palmen]

Investigation completed.

[Anders Markvardsen]

Yes investigation complet. It is worth keeping this investigation in mind when we have an code cleanup iteration in the future

[Stuart Campbell]

This ticket has been transferred to github issue ​4884