---
title: "4. Subject-by-subject networks"
output:
  rmarkdown::html_vignette:
    toc: true
vignette: >
  %\VignetteIndexEntry{4. Subject-by-subject networks}
  %\VignetteEngine{knitr::rmarkdown}
  %\VignetteEncoding{UTF-8}
---

```{r setup, include=FALSE}
knitr::opts_chunk$set(
  collapse = TRUE, comment = "#>", message = FALSE, warning = FALSE,
  fig.width = 7, fig.height = 5.5
)
```

```{r libraries}
library(idiographic)
data(srl)
vars <- c("efficacy", "value", "planning", "monitoring", "effort")
has_cograph <- requireNamespace("cograph", quietly = TRUE)
```

When the analysis target is the **idiographic map for every individual**, fit
one model per person with `build_var_each()` (OLS) or `graphical_var_each()`
(regularized). Each returns a named collection of per-subject results.

# One OLS VAR per person

```{r var-each}
var_each <- build_var_each(srl, vars = vars, id = "name", scale = TRUE)

var_each
```

The collection prints a one-line-per-cohort summary (subjects, variables, edge
counts). Reach into one person's result by name and use the usual verbs on it:

```{r var-each-one}
grace <- var_each[["Grace"]]

head(edges(grace))

matrices(grace)
```

`plot()` draws a chosen subject directly — pass `subject` instead of indexing:

```{r plot-var-each, eval=has_cograph}
plot(var_each, subject = "Grace")
```

# One graphical VAR per person

`graphical_var_each()` does the same with sparse estimation. Here we fit a
handful of students so the example runs quickly; drop the `subset()` to fit
everyone.

```{r gvar-each}
students <- subset(srl, name %in% c("Grace", "Eve", "Aisha", "Bob"))

gvar_each <- graphical_var_each(students, vars = vars, id = "name",
                                n_lambda = 8, gamma = 0)

gvar_each

summary(gvar_each[["Eve"]])
```

```{r plot-gvar-each, eval=has_cograph}
plot(gvar_each, subject = "Eve")
```