esteinig / critter Goto Github PK

I figured out a pattern for inheritance on the Distribution class in critter.blocks.distributions that load RealParameter instances on initiation. This way both Distribution and RealParameter identifiers can be set by users (for @ in the BEAST XMLs) while conforming to pydantic.BaseModel inheritance from the Distribution class.

Aim:

Production grade library for pathogen phylodynamics

Ideas:

• XMl wrappers for safe model building in BEAST
• bacterial phylodynamics module:
  • species specific configurations / workflows from raw data (TB, MRSA, Kleb...)
  • clonal expansion coalescent model (CaveDive) [collaboration]
  • birth-death skyline models (BDSky and MTBD) [based on our papers]
• SARS-CoV-2 models for variants and national / subnational configs [collaboration SSI]
• Nanopore preprocessing using Random Forests [based on preprint]
• integration with Nextflow to run BEAST + BEAGLE on GPU

Features:

• pydantic model oriented
• dynamic config of models
• test suites + docs
• user friendly releases
• command line interface

Deployment:

• FastAPI server + Vue
• continuous integration

Implement basic XML blocks based on BEAST 2.5:

• RealParameter --> <parameter/>
• Distribution base class --> <Gamma/>
• Priors --> <prior/>
• Operator --> <operator/>

Travis is nice for pytest deployment on merges into main but for releases ideally we should test a series of model configs and generated XMLs across BEST versions (using the validate option for example)

Probably need Docker container for it and maybe some GitHub actions or something? Haven't tried them yet.

Reimplemented YAML model configuration with pydantic - which really is more suitable for this than the base block [#2] implementations. It seems that the pydantic config classes and base blocks should be merged, so that the critter.config.CritterConfig class is not required to translate between them.

Model configs from YAML are great for sharing tempates of specific model configurations - e.g. a sliced sampling proportion birth death skyline model. Itm ight also provide a way forward to implement dynamic configuration of the parameters from the CLI.

Omg @Wytamma I cannot believe it there exists a project called BEASTling wrapping linguistic model XMLs - that is so funny, guess it's been around long enough to have a explored a significant portion of the BEAST namespace ^^

https://beastling.readthedocs.io/en/latest/overview.html#what-does-beastling-actually-do

Gotta rethink a name, that's also available on readthedocs.io

All those green unit test look great! I'll add some integration tests for critter (partly to learn how it works).

Include some of the useful blocks from feast

Unique identifier in distributions and parameters (using critter.utils.get_uuid) is instantiated at import, so that using the same import results in assignment of the same (replicated) identifiers.

Created a readthedocs hook, need to figure out a good structure. Do you have a template @Wytamma?