Sam Abbott
@seabbs.bsky.social
Real-time infectious disease modelling. Developing methods for outbreak response, surveillance, and pandemic preparedness. samabbott.co.uk
Come join me on the epinowcast forum: https://community.epinowcast.org/latest
Come join me on the epinowcast forum: https://community.epinowcast.org/latest
Thanks yes makes sense. In r land the standard is quarto which gives you both a notebook and running which I like but I do like the idea of literate being very light weight.
February 4, 2026 at 7:32 PM
Thanks yes makes sense. In r land the standard is quarto which gives you both a notebook and running which I like but I do like the idea of literate being very light weight.
yes fair enough that makes sense.
February 4, 2026 at 6:27 PM
yes fair enough that makes sense.
(your home page is 👨🍳😗)
February 4, 2026 at 6:16 PM
(your home page is 👨🍳😗)
OOO I like the idea of having collapsable figures. I have just set up DocumenterVitepress and liking it a lot. Will take a look! Did you have a strong reason for literate.jl vs just writing markdown?
February 4, 2026 at 6:15 PM
OOO I like the idea of having collapsable figures. I have just set up DocumenterVitepress and liking it a lot. Will take a look! Did you have a strong reason for literate.jl vs just writing markdown?
The context here is I am trying to lock in some standards before spinning up a bunch of other packages that follow them.
February 4, 2026 at 5:57 PM
The context here is I am trying to lock in some standards before spinning up a bunch of other packages that follow them.
(I am aware of the meme that Julia packages don't have docs and so there is no answer to this question as there is no need to ask it)
February 4, 2026 at 5:34 PM
(I am aware of the meme that Julia packages don't have docs and so there is no answer to this question as there is no need to ask it)
Looks like a few tradeoffs between these options and really what I would like I think is quarto to Documenter (it looks like there is an option to make a quarto package site but I don't really want that).
February 4, 2026 at 5:33 PM
Looks like a few tradeoffs between these options and really what I would like I think is quarto to Documenter (it looks like there is an option to make a quarto package site but I don't really want that).
See the EpiAwareR package that is linked in the package it uses the Julia front end to replicate EpiNow2 amongst others in R
February 4, 2026 at 1:58 PM
See the EpiAwareR package that is linked in the package it uses the Julia front end to replicate EpiNow2 amongst others in R
If you have an idea for a small useful package that is a model “component” that could slot in nicely - again the forum is the place to gauge interest
February 4, 2026 at 1:40 PM
If you have an idea for a small useful package that is a model “component” that could slot in nicely - again the forum is the place to gauge interest
That’s because it hasn’t been written! Look at the case studies i.e the epinow2 one for that one. Well it’s all the same thing in this framework so it depends what you mean?
February 4, 2026 at 1:40 PM
That’s because it hasn’t been written! Look at the case studies i.e the epinow2 one for that one. Well it’s all the same thing in this framework so it depends what you mean?
How did this go? Claude and i been fighting to configure it to a point i can understand how to use it. Did you start with lazyvim?
February 4, 2026 at 1:38 PM
How did this go? Claude and i been fighting to configure it to a point i can understand how to use it. Did you start with lazyvim?
One of the main things to do for the DSL layer is think about alternative designs so any thoughts there very welcome indeed
February 4, 2026 at 10:51 AM
One of the main things to do for the DSL layer is think about alternative designs so any thoughts there very welcome indeed
I plan to start working again on the DSL layer in a the next few months. So yeah lots of places to get involved at different levels - the easiest way to track is probably the epinowcast forum.
February 4, 2026 at 10:44 AM
I plan to start working again on the DSL layer in a the next few months. So yeah lots of places to get involved at different levels - the easiest way to track is probably the epinowcast forum.
So the composability ideas are a proof of concept as the interface to Turing has changed and I think we might go a slightly different way.
The main thing we are doing at the moment is building out some nice robust model elements in Julia i.e github.com/EpiAware/Cen...
The main thing we are doing at the moment is building out some nice robust model elements in Julia i.e github.com/EpiAware/Cen...
GitHub - EpiAware/CensoredDistributions.jl: Additional censored event tools for Distributions.jl
Additional censored event tools for Distributions.jl - EpiAware/CensoredDistributions.jl
github.com
February 4, 2026 at 10:44 AM
So the composability ideas are a proof of concept as the interface to Turing has changed and I think we might go a slightly different way.
The main thing we are doing at the moment is building out some nice robust model elements in Julia i.e github.com/EpiAware/Cen...
The main thing we are doing at the moment is building out some nice robust model elements in Julia i.e github.com/EpiAware/Cen...
The approach
makes it easier for different experts to contribute pieces and could work well with AI coding assistants.
makes it easier for different experts to contribute pieces and could work well with AI coding assistants.
February 3, 2026 at 7:34 PM
The approach
makes it easier for different experts to contribute pieces and could work well with AI coding assistants.
makes it easier for different experts to contribute pieces and could work well with AI coding assistants.
This paper proposes a middle ground: modular building blocks that snap together and can be reused across projects. We built a proof of concept in Julia/R and showed it can replicate published COVID-19 and influenza analyses by mixing and matching components.
February 3, 2026 at 7:34 PM
This paper proposes a middle ground: modular building blocks that snap together and can be reused across projects. We built a proof of concept in Julia/R and showed it can replicate published COVID-19 and influenza analyses by mixing and matching components.
TLDR: Disease outbreak models need to be fast, accurate, and built collaboratively, but current approaches force a trade-off. You can either chain simple models together (flexible but loses information) or build one big custom model (rigorous but not reusable).
February 3, 2026 at 7:34 PM
TLDR: Disease outbreak models need to be fast, accurate, and built collaboratively, but current approaches force a trade-off. You can either chain simple models together (flexible but loses information) or build one big custom model (rigorous but not reusable).
The visual abstract is slightly mad at points but I haven't had the courage to return Nano Banana pro to do battle yet.
February 3, 2026 at 7:30 PM
The visual abstract is slightly mad at points but I haven't had the courage to return Nano Banana pro to do battle yet.
Paper site: epiaware.org/ComposablePr...
Paper: epiaware.org/ComposablePr...
Case studies: epiaware.org/ComposablePr...
Paper: epiaware.org/ComposablePr...
Case studies: epiaware.org/ComposablePr...
readme
epiaware.org
February 3, 2026 at 7:29 PM
Paper site: epiaware.org/ComposablePr...
Paper: epiaware.org/ComposablePr...
Case studies: epiaware.org/ComposablePr...
Paper: epiaware.org/ComposablePr...
Case studies: epiaware.org/ComposablePr...