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

## Re-exporting nflreadr functions

You can re-export `nflreadr` functions in your package by using the following roxygen template:

```{r eval = FALSE}
#' @inherit nflreadr::load_nextgen_stats
#' @export
# Need to add own examples if the function name is different
#' @examples
#' load_ng_stats(2020)
load_ng_stats <- nflreadr::load_nextgen_stats
```

## Using progressr with from_url functions

The `from_url` family of functions can be adapted with the `progressively()` function decorator to signal progress. Here's an example, adapted from the `load_rosters()` source code:

```{r eval = FALSE}
load_rosters <- function(seasons = 1999:2020){

  # Create a progressor function inside your function that knows how many "steps" there will be
  p <- progressr::progressor(steps = length(seasons)) 

  # Form the URLs to pass into rds_from_url
  urls <- paste0(
    "https://github.com/nflverse/nflverse-rosters/",
    "raw/master/data/seasons/roster_",
    seasons, 
    ".rds")

  # Use `progressively()` to wrap the existing function rds_from_url and 
  # call the `p()` progressor function as a signal immediately after each iteration
  
  out <- purrr::map_dfr(urls, progressively(rds_from_url,p))
  
  return(out)
}
```

This will cause p() to execute once for each rds_from_url call, effectively "signalling" progress.

In order to receive progress updates, the user must wrap the function (load_rosters, in this case) with `progressr::with_progress()` as shown here:

```{r eval = FALSE}
progressr::with_progress(load_rosters(2010:2020))
```

For more information, please see the [progressr](https://progressr.futureverse.org) documentation.