jkharv / ecologicalhypergraphs.jl Goto Github PK

Use hypergraphs to create foodweb models which include non-trophic effects

License: MIT License

Julia 100.00%

dynamics ecology ecology-modelling graph hypergraph network network-analysis ode ode-model julia-language

ecologicalhypergraphs.jl's Introduction

EcologicalHypergraphs.jl

EcologicalHypergraphs.jl allows you to do ecology using hypergraphs. Currently, it supports using hypergraphs do create large dynamical foodweb models which can incorporate non-trophic effects. In the future, it will support statistical study of these hypergraphs as well. It may also support embedding these hypergraphs in spatial networks in the future to facilitate metacommunity models.

Getting started

This package isn't yet in the Julia general registry. You can still install it directly from this git repository.

] add github.com/jkharv/EcologicalHypergraphs.jl

Examples

There are examples in the example directory of this repository.

Integration with other packages

SimpleHypergraphs.jl
- Conversion functions are planned for converting to and from SimpleHypergraphs.jl types.

EcologicalNetworks.jl

EcologicalHypergraphs.jl currently has constructors for creating hypergraphs from EcologicalNetworks.jl types, allowing access to all the structural models that it provides. We will eventually include functions that can project hypergraphs into pairwise graphs using EcologicalNetworks.jl types, allowing the use of existing analysis tools in the package.

Symbolics.jl

Dynamical models done in EcologicalHypergraphs.jl are specified using Symbolics.jl. You can do anything that Symbolics.jl allows you to do to the resultant model.

ModelingToolkit.jl

Symbolics.jl is used via ModelingToolkit.jl meaning that parameters are tagged differently than variables and creating numerical systems ModelingToolkit.jl does all of its usual symbolic manipulations on the system.

ecologicalhypergraphs.jl's People

Contributors

Stargazers

Watchers

ecologicalhypergraphs.jl's Issues

Fix explicit multivariable function templates

I broke those while rewriting the function template system. I should fix that.

Create macro (or function) for setting species biomass initial conditions.

This can't be done in the @functional_form macro like the parameters are because u0 is a property of a species, not a node or interaction.

Add tests for dynamics functionality

The infrastructure is set up but there are no tests actually written or it.

Rewrite `Node` type to use `Dict{Num, Float64}` for vars and params.

Right now the Node type keeps four vectors of Nums and DistributionOptions, two for parameters, and two for variables. These should each be replaced with a Dict{Num, Float64} for each of parameters and variables.

Doing this obviates the need for a bunch of the code in build_system.jl.

Get rid of `DistributionOption`

This doesn't provide any actual benefit over using a simpe Union{Real, Distribution}

Implement `AbstractGraph` interface from `Graphs.jl`?

Update documentation for `@functional_form`

The current documentation predates both multiple variables and specifying parameter values.

Simplify predicates testing for producers and consumers

Right now they need two arguments, an Edge, and a trophic level dict. This produces a hard requirement for EcologicalNetworks.jl. But for these functions all we need is to check in and out degree.

Conversions for `SimpleHypergraphs.jl` types

Create functions to allow conversion of our types to those used in SimpleHypergraphs.jl

Create a new `CommunityMatrix` type.

Community matrices in this package should be represented by a special purpose type.

CM should be parametrically types to allow Num and Real based matrices.
CM is indexable by species name and cartesian index (string or int).
CM includes parameter and variable dicts.
CM implements the AbstractArray interface.
Existing code is rewritten to use this new type rather than Matrix{Num}

https://docs.julialang.org/en/v1/devdocs/offset-arrays/#man-custom-indices
https://docs.julialang.org/en/v1/manual/interfaces/#man-interface-array

Store variable values in `EcologicalHypergraph`

The variables appear in many nodes, so it doesn't make sense to store their values locally to a node. Setting initial conditions then required iterating through all of the nodes which contain that variable.