Bryce Frank

Create functions for querying models via API

1

The serverless API is more or less functional, and is accepting requests via HTTP. First, we should expand the `model` endpoint to consider various arguments. We need not be expansive...

A colleague points out that `bt` seems to refer to aboveground biomass, but the description is somewhat ambiguous to this case. I think AGB models are more common that total...

`UTF-8` warnings on `load_models()`

5

e.g. ``` There were 50 or more warnings (use warnings() to see the first 50) [ins] r$> warnings() Warning messages: 1: In readRDS(rds_path) : strings not representable in native encoding...

In v1, when we checked if model specifications were unique we relied on `dplyr::distinct` to determine this. However, `dplyr::distinct` does not work with S4 objects. It may be possible to...

also ensure refactoring of all models that use `en` (stem density) variable

I suspect `ModelSet` is poorly optimized. See especially this for-loop https://github.com/allometric/allometric/blob/c83744a3704d3f6081c6bb7752f68c1cb5b97047/R/FixedEffectsSet.R#L68-L80 We should use a profiler to see what takes the most time for a large model set (perhaps one...

It should be possible for users to maintain their own local models with the package. This would involve maintaining a separate publication and parameters directory, then running the publication files...