---
title: "BetaStability with vegan datasets"
author: "Yu Gao(gaoyu19920914@gmail.com)"
date: "`r Sys.Date()`"
output: rmarkdown::html_vignette
vignette: >
    %\VignetteIndexEntry{BetaStability with vegan datasets}
    %\VignetteEngine{knitr::rmarkdown}
    %\VignetteEncoding{UTF-8}
---

```{r setup, include = FALSE}
knitr::opts_chunk$set(
    collapse = TRUE,
    comment = "#>"
)
```

# BetaStability with vegan datasets

This vignette demonstrates the BetaStability package using `linearPred` with
multiple datasets (BCI, dune, and mite) from the vegan package. 

## Installation

First, install the package from GitHub:

```{r installation, eval = FALSE}
# Install from GitHub
# install.packages("devtools")
# devtools::install_github("gaoyu19920914/betaStability")

# OR install from BioConductor (in the future when it's available)
# if (!requireNamespace("BiocManager", quietly = TRUE))
#     install.packages("BiocManager")
# BiocManager::install("betaStability")
```

## Loading Required Packages

Load the BetaStability package and the vegan package for test data:

```{r load-packages}
library(betaStability)
library(vegan)
library(ggplot2)
data("BCI", "BCI.env", "mite", "mite.env", "dune", "dune.env")
```

## Demonstrating General Applicability with linearPred


```{r dataframe_preparing function}
df_prepare <- function(df) {
    # Remove columns where all elements are the same
    df <- df[, vapply(df, function(x) length(unique(x)) > 1, logical(1))]

    # Convert columns to integer if they contain different types of strings
    for (col in names(df)) {
        if (is.character(df[[col]]) && length(unique(df[[col]])) > 1) {
            df[[col]] <- as.integer(as.factor(df[[col]]))
        }
    }
    return(df)
}
```
### BCI Dataset with linearPred

Calculate stability for BCI dataset using `linearPred`:

```{r linear-bci}
# Calculate stability with linearPred
stability_BCI_linear <- betaStability(
    comtable = BCI,
    envmeta = BCI.env[, c("Precipitation", "Elevation", "EnvHet")],
    method = "linearPred"
)

# Inspect the result
head(stability_BCI_linear)
length(stability_BCI_linear)
```

### Dune Dataset with linearPred

Load the `dune` (community data) and `dune.env` (environmental metadata) 
datasets from the vegan package, then calculate stability:

```{r linear-dune}
data(dune)
data(dune.env)

# Inspect the data
head(dune)
head(dune.env)

# Dimensions of the datasets
cat("Dimensions of dune:", dim(dune), "\n")
cat("Dimensions of dune.env:", dim(dune.env), "\n")

# Process environmental data using df_prepare
dune.env_processed <- df_prepare(dune.env)

# Calculate stability with linearPred
stability_dune_linear <- betaStability(
    comtable = dune,
    envmeta = dune.env_processed,
    method = "linearPred"
)

# Inspect the result
head(stability_dune_linear)
length(stability_dune_linear)
```

### Mite Dataset with linearPred

Load the `mite` (community data) and `mite.env` (environmental metadata) 
datasets from the vegan package, then calculate stability:

```{r linear-mite}
data(mite)
data(mite.env)

# Inspect the data
head(mite)
head(mite.env)

# Dimensions of the datasets
cat("Dimensions of mite:", dim(mite), "\n")
cat("Dimensions of mite.env:", dim(mite.env), "\n")

# Process environmental data using df_prepare
mite.env_processed <- df_prepare(mite.env)

# Calculate stability with linearPred
stability_mite_linear <- betaStability(
    comtable = mite,
    envmeta = mite.env_processed,
    method = "linearPred"
)

# Inspect the result
head(stability_mite_linear)
length(stability_mite_linear)
```

### Visualizing Results

Visualize the stability results for all datasets using `linearPred`:

```{r plot-all-linear}
p_mite <- plotStability(stability_mite_linear)
p_mite

p_dune <- plotStability(stability_dune_linear)
p_dune

p_BCI <- plotStability(stability_BCI_linear)
p_BCI
```

## Summary

**General applicability across different datasets**: 
Using the simple and fast `linearPred` method, we successfully calculated 
stability for three different datasets (BCI, dune, and mite) representing 
different ecological systems. This demonstrates the versatility of the 
BetaStability package across various types of ecological data.

```{r}
print(sessionInfo())
```