---
title: "7. GIMME"
output:
  rmarkdown::html_vignette:
    toc: true
vignette: >
  %\VignetteIndexEntry{7. GIMME}
  %\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)
```

`build_gimme()` searches individual models and **promotes paths shared by enough
people** to the group level. As with uSEM, pass the shipped `day` column as
`time`. We fit a handful of students so the example runs quickly.

```{r gimme}
students <- subset(srl, name %in% c("Grace", "Eve", "Aisha", "Alice",
                                    "Bob", "Diana", "Frank", "Heidi"))

gimme_fit <- build_gimme(students, vars = vars, id = "name", time = "day",
                         ar = TRUE, groupcutoff = 0.75, seed = 1)

gimme_fit
```

The printout lists the group-level paths and then shows, for the temporal and
contemporaneous networks, the **proportion of subjects** carrying each path —
the quantity GIMME displays.

# Tidy tables

`edges()` returns one tidy row per path with a `level` column marking
group-level versus individual-level paths; `coefs()` gives the per-subject
estimates.

```{r gimme-tables}
head(edges(gimme_fit))

head(coefs(gimme_fit))
```

```{r gimme-matrices}
matrices(gimme_fit)
```

# Plot

`plot()` draws the faithful gimme-style mixed network: **dashed** edges are
lag-1 (temporal), **solid** edges are contemporaneous, edge width is the
proportion of subjects with the path, and **black** edges are group-level.

```{r plot-gimme, eval=has_cograph}
plot(gimme_fit)
```