Merge pull request #96 from b-cubed-eu/make-pkgdown-lighter

Make pkgdown lighter

Author: wlangera

NEWS.md

@@ -3,6 +3,7 @@
 * Add grant ID for Zenodo integration #94
 * Fix ROR url bug
 * Add publisher to metadata
+* Make pkgdown lighter #93
 
 # dubicube 0.11.0
 

codemeta.json

@@ -216,7 +216,7 @@
     },
     "SystemRequirements": null
   },
-  "fileSize": "389.183KB",
+  "fileSize": "1846.484KB",
   "citation": [
     {
       "@type": "SoftwareSourceCode",

inst/bootstrapping/bootstrap_results.rds

@@ -136,15 +136,14 @@ b3data_package <- read_package(
 
 # Load bird cube data
 bird_cube_belgium <- read_resource(b3data_package, "bird_cube_belgium_mgrs10")
-head(bird_cube_belgium)
 ```
 
 We process the cube with **b3gbi**.
 First, we select 2000 random rows to make the dataset smaller.
 This is to reduce the computation time for this tutorial.
 We select the data from 2011 - 2020.
 
-```{r}
+```r
 set.seed(123)
 
 # Make dataset smaller
@@ -161,6 +160,47 @@ processed_cube <- process_cube(
 processed_cube
 ```
 
+```{r, echo=FALSE, message=FALSE, results='hide'}
+if (
+  system.file("bootstrapping", "processed_cube.rds", package = "dubicube") == ""
+) {
+  # Read data package
+  b3data_package <- read_package(
+    "https://zenodo.org/records/15211029/files/datapackage.json"
+  )
+
+  # Load bird cube data
+  bird_cube_belgium <- read_resource(b3data_package, "bird_cube_belgium_mgrs10")
+
+  set.seed(123)
+
+  # Make dataset smaller
+  rows <- sample(nrow(bird_cube_belgium), 2000)
+  bird_cube_belgium <- bird_cube_belgium[rows, ]
+
+  # Process cube
+  processed_cube <- process_cube(
+    bird_cube_belgium,
+    first_year = 2011,
+    last_year = 2020,
+    cols_occurrences = "n"
+  )
+
+  saveRDS(
+    processed_cube,
+    file.path("..", "..", "inst", "bootstrapping", "processed_cube.rds")
+  )
+} else {
+  processed_cube <- readRDS(
+    system.file("bootstrapping", "processed_cube.rds", package = "dubicube")
+  )
+}
+```
+
+```{r, echo=FALSE}
+processed_cube
+```
+
 ### Analysis of the data
 
 Let's say we are interested in the mean number of observations per grid cell per year.
@@ -171,8 +211,8 @@ We create a function to calculate this.
 ```
 
 ```{r}
-# Function to calculate statistic of interest
-# Mean number of observations per grid cell per year
+# Function to calculate the statistic of interest
+# Mean observations per grid cell per year
 mean_obs <- function(data) {
   data %>%
     dplyr::mutate(x = mean(obs), .by = "cellCode") %>%
@@ -197,7 +237,7 @@ On their own, these values don’t reveal how much uncertainty surrounds them. T
 
 We use the `bootstrap_cube()` function to perform bootstrapping (see also the [bootstrap tutorial](https://b-cubed-eu.github.io/dubicube/articles/bootstrap-method-cubes.html)).
 
-```{r}
+```r
 bootstrap_results <- bootstrap_cube(
   data_cube = processed_cube,
   fun = mean_obs,
@@ -207,6 +247,35 @@ bootstrap_results <- bootstrap_cube(
 )
 ```
 
+```{r, echo=FALSE, message=FALSE, results='hide'}
+if (
+  system.file(
+    "bootstrapping", "bootstrap_results.rds", package = "dubicube"
+  ) == ""
+) {
+  bootstrap_results <- bootstrap_cube(
+    data_cube = processed_cube,
+    fun = mean_obs,
+    grouping_var = "year",
+    samples = 1000,
+    seed = 123
+  )
+
+  saveRDS(
+    bootstrap_results,
+    file.path("..", "..", "inst", "bootstrapping", "bootstrap_results.rds")
+  )
+} else {
+  bootstrap_results <- readRDS(
+    system.file("bootstrapping", "bootstrap_results.rds", package = "dubicube")
+  )
+}
+```
+
+```{r, echo=FALSE}
+print("Performing whole-cube bootstrap with `boot::boot()`.")
+```
+
 ### Interval calculation
 
 Now we can use the `calculate_bootstrap_ci()` function to calculate confidence limits. It relies on the following arguments:

inst/bootstrapping/processed_cube.rds

@@ -101,23 +101,22 @@ library(dubicube)     # Analysis of data quality & indicator uncertainty
 We load the bird cube data from the **b3data** data package using **frictionless** (see also [here](https://github.com/b-cubed-eu/b3data-scripts)).
 It is an occurrence cube for birds in Belgium between 2000 en 2024 using the MGRS grid at 10 km scale.
 
-```{r}
+```r
 # Read data package
 b3data_package <- read_package(
   "https://zenodo.org/records/15211029/files/datapackage.json"
 )
 
 # Load bird cube data
 bird_cube_belgium <- read_resource(b3data_package, "bird_cube_belgium_mgrs10")
-head(bird_cube_belgium)
 ```
 
 We process the cube with **b3gbi**.
 First, we select 2000 random rows to make the dataset smaller.
 This is to reduce the computation time for this tutorial.
 We select the data from 2011 - 2020.
 
-```{r}
+```r
 set.seed(123)
 
 # Make dataset smaller
@@ -134,24 +133,71 @@ processed_cube <- process_cube(
 processed_cube
 ```
 
+```{r, echo=FALSE, message=FALSE, results='hide'}
+if (
+  system.file("bootstrapping", "processed_cube.rds", package = "dubicube") == ""
+) {
+  # Read data package
+  b3data_package <- read_package(
+    "https://zenodo.org/records/15211029/files/datapackage.json"
+  )
+
+  # Load bird cube data
+  bird_cube_belgium <- read_resource(b3data_package, "bird_cube_belgium_mgrs10")
+
+  set.seed(123)
+
+  # Make dataset smaller
+  rows <- sample(nrow(bird_cube_belgium), 2000)
+  bird_cube_belgium <- bird_cube_belgium[rows, ]
+
+  # Process cube
+  processed_cube <- process_cube(
+    bird_cube_belgium,
+    first_year = 2011,
+    last_year = 2020,
+    cols_occurrences = "n"
+  )
+
+  saveRDS(
+    processed_cube,
+    file.path("..", "..", "inst", "inst", "bootstrapping", "processed_cube.rds")
+  )
+} else {
+  processed_cube <- readRDS(
+    system.file("bootstrapping", "processed_cube.rds", package = "dubicube")
+  )
+}
+```
+
+```{r, echo=FALSE}
+processed_cube
+```
+
 ### Analysis of the data
 
 Let's say we are interested in the mean number of observations per grid cell per year.
 We create a function to calculate this.
 
+```{r, echo=FALSE}
+# nolint start: object_usage_linter.
+```
+
 ```{r}
 # Function to calculate the statistic of interest
 # Mean observations per grid cell per year
 mean_obs <- function(data) {
-  obs <- x <- NULL
-
   data %>%
     dplyr::mutate(x = mean(obs), .by = "cellCode") %>%
     dplyr::summarise(diversity_val = mean(x), .by = "year") %>%
     as.data.frame()
 }
 ```
 
+```{r, echo=FALSE}
+# nolint end: object_usage_linter.
+```
+
 We get the following results:
 
 ```{r}
@@ -182,7 +228,7 @@ We use the `bootstrap_cube()` function to do this. It relies on the following ar
 - **`progress`**:  
   Logical flag to show a progress bar. Set to `TRUE` to enable progress reporting; default is `FALSE`.
 
-```{r}
+```r
 bootstrap_results <- bootstrap_cube(
   data_cube = processed_cube,
   fun = mean_obs,
@@ -192,6 +238,35 @@ bootstrap_results <- bootstrap_cube(
 )
 ```
 
+```{r, echo=FALSE, message=FALSE, results='hide'}
+if (
+  system.file(
+    "bootstrapping", "bootstrap_results.rds", package = "dubicube"
+  ) == ""
+) {
+  bootstrap_results <- bootstrap_cube(
+    data_cube = processed_cube,
+    fun = mean_obs,
+    grouping_var = "year",
+    samples = 1000,
+    seed = 123
+  )
+
+  saveRDS(
+    bootstrap_results,
+    file.path("..", "..", "inst", "bootstrapping", "bootstrap_results.rds")
+  )
+} else {
+  bootstrap_results <- readRDS(
+    system.file("bootstrapping", "bootstrap_results.rds", package = "dubicube")
+  )
+}
+```
+
+```{r, echo=FALSE}
+print("Performing whole-cube bootstrap with `boot::boot()`.")
+```
+
 This returned a list of `"boot"` objects. We can convert this to a dataframe 
 
 ```{r}

inst/bootstrapping/speed_comparison.rds

@@ -186,7 +186,7 @@ It relies on the following arguments:
     Logical flag to show a progress bar.
     Set to `TRUE` to enable progress reporting; default is `FALSE`.
 
-```{r}
+```r
 cv_results <- cross_validate_cube(
   data_cube = processed_cube,
   fun = mean_obs,
@@ -195,6 +195,30 @@ cv_results <- cross_validate_cube(
 )
 ```
 
+```{r, echo=FALSE, message=FALSE, results='hide'}
+if (
+  system.file(
+    "cross_validation", "cv_results.rds", package = "dubicube"
+  ) == ""
+) {
+  cv_results <- cross_validate_cube(
+    data_cube = processed_cube,
+    fun = mean_obs,
+    grouping_var = "year",
+    out_var = "taxonKey"
+  )
+
+  saveRDS(
+    cv_results,
+    file.path("..", "..", "inst", "cross_validation", "cv_results.rds")
+  )
+} else {
+  cv_results <- readRDS(
+    system.file("cross_validation", "cv_results.rds", package = "dubicube")
+  )
+}
+```
+
 ```{r}
 head(cv_results)
 ```
@@ -269,7 +293,7 @@ As stated in the documentation, it is also possible to cross-validate over a dat
 In this case, set the argument `processed_cube = FALSE`.
 This is implemented allow for flexible use of simple dataframes, while still encouraging the use of `b3gbi::process_cube()` as default functionality.
 
-``` r
+```r
 cv_results_df <- cross_validate_cube(
   data_cube = processed_cube$data,      # data.frame object
   fun = mean_obs,