Skip to content

Commit

Permalink
update gcube
Browse files Browse the repository at this point in the history
  • Loading branch information
@wlangera
wlangera committed Jul 26, 2024
1 parent 337687b commit bd5d9cd
Showing 1 changed file with 153 additions and 31 deletions.
184 changes: 153 additions & 31 deletions source/gcube_integration_for_b3gbi.Rmd
View file
Original file line number Diff line number Diff line change
Expand Up @@ -74,7 +74,7 @@ The new `process_cube()` function is more flexible in this sense.

## Try gcube output as input

```
```r
## Create cube with gcube (4 time points, 1 species)
# Create a polygon to simulate occurrences
polygon <- st_polygon(list(cbind(c(5, 10, 8, 2, 3, 5), c(2, 1, 7, 9, 5, 2))))
Expand Down Expand Up @@ -219,7 +219,7 @@ This throws an error by `check_cell_size()`.
# b3gbi version 0.2.3

Review for [this pull request](https://github.com/b-cubed-eu/b3gbi/pull/25).
We install the [b3gbi](https://github.com/b-cubed-eu/b3gbi/) (version 0.2.3) and [gcube](https://github.com/b-cubed-eu/gcube) (version 0.1.0) packages.
We install the [b3gbi](https://github.com/b-cubed-eu/b3gbi/) (version 0.2.3) and [gcube](https://github.com/b-cubed-eu/gcube) (version 0.3.0) packages.

We create a datacube with **gcube** for 6 species over 6 time points.
First we define the spatial extend.
Expand Down Expand Up @@ -247,59 +247,181 @@ Let's simulate the cube.

```{r}
# Create dataframe with simulation function arguments
multi_species_args <- tibble(
plgn = rep(list(polygon), 6),
initial_average_abundance = rep(c(50, 100, 500), 2),
n_time_points = rep(6, 6),
temporal_function = c(simulate_random_walk, simulate_random_walk, rep(NA, 4)),
sd_step = c(1, 1, rep(NA, 4)),
spatial_autocorr = c(rep("random", 3), rep("clustered", 3)),
detection_probability = rep(c(0.8, 0.9, 1), 2),
coords_uncertainty_meters = rep(c(25, 30, 50), 2),
grid = rep(list(cube_grid), 6),
seed = 123
)
# Generate taxonomic hierarchy
multi_species_dataset <- generate_taxonomy(
num_species = multi_species_args,
num_genera = 4,
num_families = 2,
seed = 123)
multi_species_dataset <- tibble(
plgn = rep(list(polygon), 6),
n_time_points = rep(6, 6),
detection_probability = rep(c(0.8, 0.9, 1), 2),
coords_uncertainty_meters = rep(c(25, 30, 50), 2),
grid = rep(list(cube_grid), 6),
seed = 123
)
# Add taxonomic hierarchy and generate cube
map_occ_cube_df <- multi_species_dataset %>%
generate_taxonomy(num_genera = 4, num_families = 2, seed = 123) %>%
map_simulate_occurrences() %>%
map_sample_observations() %>%
map_filter_observations() %>%
map_add_coordinate_uncertainty() %>%
map_grid_designation(nested = FALSE)
map_grid_designation(nested = FALSE) %>%
select(-all_of(names(multi_species_dataset))) %>%
select(-occurrences, -observations_total, -observations)
glimpse(map_occ_cube_df)
```

This time we do not write out a csv, but we use the dataframe to process the cube.

```{r}
occ_cube_df_cleaned <- map_occ_cube_df %>%
select(-all_of(names(multi_species_args))) %>%
select(-occurrences, -observations_total, -observations) %>%
mutate(species_key = as.numeric(gsub("species", "", species)))
## Process cube with b3gbi
# Process cube with b3gbi
gcube_data <- process_cube(
cube_name = occ_cube_df_cleaned,
data_type = "df",
cube_name = map_occ_cube_df,
grid_type = "none",
cols_year = "time_point",
cols_cellCode = "id",
cols_occurrences = "n",
cols_scientificName = "species",
cols_minCoordinateUncertaintyInMeters = "min_coord_uncertainty",
cols_kingdom = "kingdom",
cols_family = "family",
cols_speciesKey = "species_key"
)
gcube_data
```

```{r}
total_occ_ts(gcube_data)
```

reprex
```r
library(dplyr)
#>
#> Attaching package: 'dplyr'
#> The following objects are masked from 'package:stats':
#>
#> filter, lag
#> The following objects are masked from 'package:base':
#>
#> intersect, setdiff, setequal, union
library(sf)
#> Linking to GEOS 3.12.1, GDAL 3.8.4, PROJ 9.3.1; sf_use_s2() is TRUE
library(gcube)
library(b3gbi)

# Create a polygon to simulate occurrences
polygon <- st_polygon(list(cbind(c(500, 1000, 1000, 600, 200, 100, 500),
c(200, 100, 700, 1000, 900, 500, 200))))

# Create grid for grid designation
cube_grid <- st_make_grid(
st_buffer(polygon, 50),
n = c(20, 20),
square = TRUE) %>%
st_sf()

# Create dataframe with simulation function arguments
multi_species_dataset <- tibble(
plgn = rep(list(polygon), 6),
n_time_points = rep(6, 6),
detection_probability = rep(c(0.8, 0.9, 1), 2),
coords_uncertainty_meters = rep(c(25, 30, 50), 2),
grid = rep(list(cube_grid), 6),
seed = 123
)

# Add taxonomic hierarchy and generate cube
map_occ_cube_df <- multi_species_dataset %>%
generate_taxonomy(num_genera = 4, num_families = 2, seed = 123) %>%
map_simulate_occurrences() %>%
map_sample_observations() %>%
map_filter_observations() %>%
map_add_coordinate_uncertainty() %>%
map_grid_designation(nested = FALSE) %>%
select(-all_of(names(multi_species_dataset))) %>%
select(-occurrences, -observations_total, -observations)
#> [1] [using unconditional Gaussian simulation]
#> [2] [using unconditional Gaussian simulation]
#> [3] [using unconditional Gaussian simulation]
#> [4] [using unconditional Gaussian simulation]
#> [5] [using unconditional Gaussian simulation]
#> [6] [using unconditional Gaussian simulation]

# Process cube with b3gbi
gcube_data <- process_cube(
cube_name = map_occ_cube_df,
grid_type = "none",
cols_year = "time_point",
cols_cellCode = "id",
cols_occurrences = "n",
cols_scientificName = "species",
cols_minCoordinateUncertaintyInMeters = "min_coord_uncertainty",
cols_kingdom = "kingdom",
cols_family = "family",
cols_speciesKey = "species_key"

)
gcube_data
#>
#> Simulated data cube for calculating biodiversity indicators
#>
#> Date Range: 1 - 5
#> Number of cells:
#> Grid reference system: none
#> Coordinate range:
#> NULL
#>
#> Total number of observations: 1382
#> Number of species represented: 6
#> Number of families represented:
#>
#> Kingdoms represented:
#>
#> First 10 rows of data (use n = to show more):
#>
#> # A tibble: 12,000 × 13
#> scientificName taxonKey genus family order class phylum kingdom year id
#> <chr> <dbl> <chr> <chr> <chr> <chr> <chr> <chr> <dbl> <chr>
#> 1 species1 1 genus3 family1 orde… clas… phylu… kingdo… 1 106
#> 2 species1 1 genus3 family1 orde… clas… phylu… kingdo… 1 109
#> 3 species1 1 genus3 family1 orde… clas… phylu… kingdo… 1 113
#> 4 species1 1 genus3 family1 orde… clas… phylu… kingdo… 1 117
#> 5 species1 1 genus3 family1 orde… clas… phylu… kingdo… 1 119
#> 6 species1 1 genus3 family1 orde… clas… phylu… kingdo… 1 124
#> 7 species1 1 genus3 family1 orde… clas… phylu… kingdo… 1 131
#> 8 species1 1 genus3 family1 orde… clas… phylu… kingdo… 1 134
#> 9 species1 1 genus3 family1 orde… clas… phylu… kingdo… 1 147
#> 10 species1 1 genus3 family1 orde… clas… phylu… kingdo… 1 154
#> # ℹ 11,990 more rows
#> # ℹ 3 more variables: obs <dbl>, minCoordinateUncertaintyInMeters <dbl>,
#> # geometry <POLYGON>

# Try calculate time series indicator
total_occ_ts(gcube_data)
#> Biodiversity indicator time series
#>
#> Name of indicator: Total Occurrences
#>
#> Date Range: 1 - 5
#>
#> Coordinate range represented:
#> xmin xmax ymin ymax
#> "NA" "NA" "NA" "NA"
#>
#> Number of species represented: 6
#> Kingdoms represented: NA
#>
#> First 10 rows of data (use n = to show more):
#>
#> # A tibble: 5 × 2
#> year diversity_val
#> <dbl> <dbl>
#> 1 1 244
#> 2 2 326
#> 3 3 198
#> 4 4 282
#> 5 5 332
```

<sup>Created on 2024-07-26 with [reprex v2.1.0](https://reprex.tidyverse.org)</sup>

0 comments on commit bd5d9cd

Please sign in to comment.