Transition to new Rt back-calculation method

[seabbs]

This should really be looked at for next week and would make sense for someone to do who has not been involved in developing the code base as a validation.

[kathsherratt]

See PR #40 .

Remaining tasks before next Monday submission:

• [x] Finalise and merge PR
• [ ] Test run before next submission date both existing and new Rt methods
• [ ] Remove Epinow1 code
• [ ] Remove Epinow1 from update script, ensembling & scoring scripts

When possible:

• [ ] Run historical forecasts to get samples and formatted output

[seabbs]

PR looks good to me.

Has any work been done on checking out the QRA - is it at a point we could use it @nikosbosse?

Does everyone else have access to all our juicy compute? I have been spinning nodes up and using docker so there might be an easier way for people to get the access they need.

[kathsherratt]

Thanks for merging Sam.

QRA actually looks okay to me. When I was testing yesterday it produced different weights for the first time (0 to Rt-Epinow1, 25% TS-deaths, 75% deaths-on-cases) so looks like it is working.

Will update on QRA when it has some historical Rt-Epinow2 forecasts to work off.

Re compute, I don't have access, would be handy at some point though happy using met server for now.

[seabbs]

cool - for access ping seb.

Sounds good on the QRA front. The plan is to not use the EpiNow1 stuff at all right? I don't think it contains anything useful

[sbfnk]

You should have access - can you try logging in with your LSHTM details at https://portal.azure.com?

[kathsherratt]

Thanks Seb! I can log in with LSHTM details, but don't have any resource groups / VMs / anything else going on. Could you send me a link or add me in some way?

[seabbs]

Can you change group in the top right hand corner?

[kathsherratt]

Thanks both for the help on accessing the azure set up. Closing this issue as Epinow2 is up and running. See #44 for next steps!