---
title: "Using Continuous Variables"
output: rmarkdown::html_vignette
vignette: >
  %\VignetteIndexEntry{Using Continuous Variables}
  %\VignetteEngine{knitr::rmarkdown}
  %\VignetteEncoding{UTF-8}
---

# Introduction

In many food webs, a consumer's ability to eat a resource is limited by their relative body sizes. The `infer_edgelist()` function allows you to layer a numerical rule (like a predator-prey mass ratio) on top of your categorical trait matching.

# Prepare Numerical Data

First, ensure your taxon data includes a numeric column for size (e.g., body mass in kg or length in cm). Here we will modify our traits dataframe

```{r, modify_traits}

library(pfwim)
library(dplyr)

# Adding body mass to our traits
traits_numeric <- traits %>%
  mutate(body_mass = case_when(
    species == "polar_bear" ~ 450,
    species == "seal"       ~ 100,
    species == "orca"       ~ 3000,
    species == "cod"        ~ 5,
    species == "plankton"   ~ 0.0001,
    TRUE                    ~ 10 # Default for others
  ))

head(traits_numeric)

```

# Define the Size Rule Function

The num_size_rule argument requires a function that takes two inputs (res_size, con_size) and returns 1 (feasible) or 0 (not feasible).

```{r, size_rule}

my_size_rule <- function(res_size, con_size) {
  ratio <- con_size / res_size
  ifelse(ratio >= 2 & ratio <= 100, 1, 0)
}

```

# Run Inference with Size Constraints

When `col_num_size` is provided, the function treats it as an additional trait type. If `certainty_req = "all"`, the interaction must satisfy all categorical matches AND the numerical size rule.

```{r, build_list}

edgelist_size <- infer_edgelist(
  data = traits_numeric,
  cat_combo_list = feeding_rules,
  col_taxon = "species",
  col_num_size = "body_mass",    # Point to the numeric column
  num_size_rule = my_size_rule,   # Apply our custom logic
  certainty_req = "all",
  hide_printout = TRUE
)

head(edgelist_size)
```