---
title: "GIFT tutorial"
author: "Pierre Denelle & Patrick Weigelt"
date: "`r Sys.Date()`"
output: 
  rmarkdown::html_vignette:
    toc: true
vignette: >
 %\VignetteIndexEntry{GIFT tutorial}
 %\VignetteEncoding{UTF-8} 
 %\VignetteEngine{knitr::rmarkdown}
editor_options: 
 chunk_output_type: console
---

<style>
body {
text-align: justify}
</style>

```{r setup, include=FALSE}

knitr::opts_chunk$set(echo = TRUE, message = FALSE, warning = FALSE,
                      fig.width = 4, fig.height = 4, fig.align = "center")
# Packages --------------------------------------------------------------------
suppressPackageStartupMessages({
  suppressWarnings({
    library("GIFT")
    library("dplyr")
    library("knitr")
    library("kableExtra")
    library("ggplot2")
    library("sf")
    library("rnaturalearth")
    library("rnaturalearthdata")
    library("tidyr")
    library("ape")
    library("patchwork")
  })
})

options(tinytex.verbose = TRUE)
```

```{r gottingen_logo, fig.show = 'hold', out.width = "20%", echo = FALSE}
knitr::include_graphics("../man/figures/biodiv_gottingen_logo.png")
knitr::include_graphics("../man/figures/GIFT.png")
```

<br>

The Global Inventory of Floras and Traits (GIFT) is a database of floras and
plant checklists, distributed worldwide. It also includes trait and
phylogenetic information. The GIFT R package grants an access to
the GIFT database.
<br>
This vignette illustrates the most common uses of the package with
detailed examples.

1. Retrieving plant checklists within a given area (example: Mediterranean)  
2. Getting the distribution of a plant species  
3. Retrieving trait information for a subset of plant species  
4. Retrieving environmental information for a list of polygons/regions  
5. Retrieving a plant phylogeny and plotting a trait coverage on it  

<br>

The following R packages are required to build this vignette:

```{r, echo=TRUE, eval=FALSE}
library("GIFT")
library("dplyr")
library("knitr")
library("kableExtra")
library("ggplot2")
library("sf")
library("rnaturalearth")
library("rnaturalearthdata")
library("tidyr")
library("ape")
library("patchwork")
```

<br><br>

**Side-note**  
Some of the following queries may take a long time to complete Increasing the
timeout makes it easier to complete larger downloads. This can be done as
follows:  
```{r, echo = TRUE, eval = FALSE}
options(timeout = max(1000, getOption("timeout")))
```

<br><br>

# 1. Checklists for a region
## 1.1. Shapefile
Let's assume we are interested in having a floristic knowledge of the western
part of the Mediterranean basin. For this purpose, we can simply use a
shape file of the region of interest and feed it to the `GIFT_checklists()`
function.  

We do provide a shape file of this region in the `GIFT` R package, which you
can access using the `data("western_mediterranean")` command.

```{r, fig.width = 3, fig.height = 3}
data("western_mediterranean")

world <- ne_coastline(scale = "medium", returnclass = "sf")
world_countries <- ne_countries(scale = "medium", returnclass = "sf")
# Fixing polygons crossing dateline
world <- st_wrap_dateline(world)
world_countries <- st_wrap_dateline(world_countries)

# Eckert IV projection
eckertIV <-
  "+proj=eck4 +lon_0=0 +x_0=0 +y_0=0 +ellps=WGS84 +datum=WGS84 +units=m +no_defs"

ggplot(world) +
  geom_sf(color = "gray50") +
  geom_sf(data = western_mediterranean, fill = "darkblue", color = "black",
          alpha = 0.5, size = 1) +
  labs(title = "Western Mediterranean basin") +
  lims(x = c(-20, 20), y = c(24, 48)) +
  theme_void()
```

<br>
*Please note that shapes used in GIFT are unprojected (Geographic Coordinate *
*System WGS84), and that all shapefiles provided should be in this CRS.*
*You can check the coordinate reference system of a sf object by using*
`sf::st_crs().`
<br>

## 1.2. Main arguments

Now that we have a shape for the region of interest, let's call
`GIFT_checklists()`. This wrapper function has many arguments, which we detail
in this subsection.
<br>
First, the taxonomic group of interest. We may be interested in a particular
group of plants, say only Angiosperms. In this case, we would set the
`taxon_name` argument like this `taxon_name = "Angiospermae"`. If we are
interested in a particular family of plants, let's say orchids, then 
`taxon_name = "Orchidaceae"`.
<br>
To see all the options for the `taxon_name` argument, you can run the
`GIFT_taxonomy()` function and look at the `taxon_name` column of its output.
<br>
Along with this first argument comes `complete_taxon`. This argument, set to
TRUE by default, determines whether only regions represented by checklists in
GIFT that completely cover **the taxon of interest** should be retrieved.
Figure 1 illustrates the principle.
<br><br>

```{r, fig.cap = "Figure 1. Principle of the complete_taxon argument", out.width = "50%",echo = FALSE}
knitr::include_graphics("../man/figures/GIFT_complete_taxon.svg")
```
<br>
In Figure 1, we want to retrieve checklists of Angiosperms. In the first
available region, region A, only one checklist is of interest. This checklist
is then always retrieved. In region B, there is only one checklist of orchids,
which is only a subset of Angiosperms. If `complete_taxon` is set to `TRUE`,
then this checklist won't be retrieved, otherwise yes. Finally, in region C,
there is a checklist for vascular plants and one for orchids. In both cases,
the checklist of vascular plants will be retrieved after filtering out the
non-angiosperm species. The checklist of Orchids is also retrieved in both
cases because it is not the only one available and because it can complete the
floristic knowledge for Angiosperms in this region.
<br>
<br>
The following arguments of `GIFT_checklists()` refer to the floristic status of
plant species. For example, we may be interested only in endemic or
naturalized species. The default value is to get all native species.
<br>
Similarly, two arguments are needed in the function. First, `floristic_group`
defines the group of interest. Second, `complete_floristic` indicates whether
or not to retrieve incomplete regions with respect to the selected floristic
group. The logic is detailed in Figure 2 and is similar to the `complete_taxon`
argument shown above  

<br>
```{r, fig.cap = "Figure 2. Principle of the complete_floristic argument", out.width = "50%", echo = FALSE}
knitr::include_graphics("../man/figures/GIFT_complete_floristic.svg")
```
<br>

The next set of arguments relate to the spatial match between the desired area
and the GIFT database. 
<br>

The main argument in this regard, when providing a shapefile or a set of
coordinates, is the `overlap` argument. This argument can take 4 options, each
of which produces different result, as shown in Figure 3.
<br>

```{r, fig.cap = "Figure 3. Principle of GIFT_spatial()", out.width = "50%", echo = FALSE}
knitr::include_graphics("../man/figures/GIFT_spatial.svg")
```
<br>

In Figure3, the GIFT polygons shown in orange either intersect, fall
inside or outside the provided shape file. The overlap argument below each GIFT
polygon illustrates in which situation a given GIFT polygon will or will not
be retrieved.

<br>

Another important spatial feature we provide is the possibility to remove
overlapping polygons. In fact, for many regions of the world, there are several
polygons in the GIFT database that cover them. If overlapping polygons are not
an issue for your case study, you can simply set `remove_overlap` to FALSE (top
right part of Figure 4). 
However, if you want to have only one polygon per region, you can set
`remove_overlap` to `TRUE`. In this case, the `GIFT_checklists()` will either
retrieve the smaller or the larger polygon. This depends on the values set for
the `area_threshold_mainland` argument as shown in Figure 4.
<br>

```{r, fig.cap = "Figure 4. Removing overlapping polygons with remove_overlap argument", out.width = "50%", echo = FALSE}
knitr::include_graphics("../man/figures/GIFT_no_overlap.svg")
```
<br>

`area_threshold_mainland` takes a value in $km^2$. If the area of the
smaller polygon is less than the threshold, then the larger overlapping polygon
is retrieved (lower left part in Figure 4). If the smaller polygon exceeds the
threshold, then it is retrieved (lower right part of Figure 4).
There is a similar argument for islands, `area_threshold_island`, which is set
to 0 $km^2$ by default. This way the smaller islands are always retrieved by
default.
<br><br>
Note also that polygons are considered to overlap if they exceed a
certain percentage of overlap. This percentage can be modified using the
`overlap_threshold` argument (Figure 5). This argument is set by default to
10%.
<br>

```{r, fig.cap = "Figure 5. Principle of the overlap_th argument", out.width = "50%", echo = FALSE}
knitr::include_graphics("../man/figures/GIFT_overlap_th.svg")
```
<br>

## 1.3. GIFT_checklists()

Now that we have covered the main arguments of `GIFT_checklists()`, we can
retrieve plant checklists for the Mediterranean region. `GIFT_checklists()`
returns a list with two elements. First the metadata of the checklists matching
the different criteria, named `$lists`. The second element is a `data.frame` of
all the checklists with the species composition per checklist (`$checklists`).
<br>
If you only want to retrieve the metadata, you can set the `list_set_only` 
argument to `TRUE`.
<br>

```{r, eval=FALSE}
ex_meta <- GIFT_checklists(taxon_name = "Angiospermae",
                           shp = western_mediterranean,
                           overlap = "centroid_inside",
                           list_set_only = TRUE)
```

<br>
And to retrieve the species composition:
<br>

```{r, eval = FALSE, echo = TRUE}
medit <- GIFT_checklists(taxon_name = "Angiospermae",
                         complete_taxon = TRUE,
                         floristic_group = "native",
                         complete_floristic = TRUE,
                         geo_type = "All",
                         shp = western_mediterranean,
                         overlap = "centroid_inside", 
                         remove_overlap = FALSE,
                         taxonomic_group = TRUE) # this argument adds two
# columns to the checklist: plant family and taxonomic group of each species
```

<br>

We can now have an estimation on the number of checklists with native
Angiosperm species in the western part of the Mediterranean basin, as well as
of the number of species.
<br>

```{r, eval=FALSE}
# Number of references covered
length(unique(medit[[2]]$ref_ID))
#   22 references

# Number of checklists covered (one reference can have several lists inside)
length(unique(medit[[2]]$list_ID))
#   115 checklists

# Number of species
length(unique(medit[[2]]$work_species))
#   12840 plant species
```

<br>
You can now apply different values for the arguments detailed above. As you can
see, the number of checklists retrieved decreases as you become stricter on
some criteria.  For example, when removing overlapping regions:
<br>

```{r, eval = FALSE, echo = TRUE}
medit_no_overlap <- GIFT_checklists(shp = western_mediterranean,
                                    overlap = "centroid_inside",
                                    taxon_name = "Angiospermae",
                                    remove_overlap = TRUE)

# Number of references covered
length(unique(medit[[2]]$ref_ID)) # 23 references
length(unique(medit_no_overlap[[2]]$ref_ID)) # 22 references
```

Note that the function not only works with a shape file but can accept a
set of coordinates. The example below illustrates a case where you want to
retrieve GIFT checklists that intersect the coordinates of Göttingen.
<br>

```{r, echo = TRUE, eval = FALSE}
custom_point <- cbind(9.9, 51) # coordinates of Göttingen

got <- GIFT_checklists(coordinates = custom_point,
                       overlap = "extent_intersect",
                       taxon_name = "Angiospermae",
                       remove_overlap = TRUE,
                       list_set_only = TRUE)
```
<br><br>

To cite properly the references retrieved, you can run the function
`GIFT_references()` and look for the column `ref_long`. The column
`geo_entity_ref` associates each reference to a name.

<br>

## 1.4. Species richness map

Once we have downloaded a set of checklists, it is possible to map the species
richness of the taxonomic group of interest. To do this, we use a
combination of two functions: `GIFT_richness()` which returns either species
richness or trait coverage per polygon, and `GIFT_shapes()` which returns the
shapefile of a list of GIFT polygons.
<br>
The next two chunks illustrate this for the Angiosperms in the World and in the
Western part of the Mediterranean basin.
<br>

```{r, eval=FALSE, echo = TRUE}
gift_shapes <- GIFT_shapes() # retrieves all shapefiles by default
```

```{r, echo = TRUE, eval = FALSE, message=FALSE, fig.width = 8, fig.height = 4}
angio_rich <- GIFT_richness(taxon_name = "Angiospermae")

rich_map <- dplyr::left_join(gift_shapes, angio_rich, by = "entity_ID") %>%
  dplyr::filter(stats::complete.cases(total))

ggplot(world) +
  geom_sf(color = "gray50") +
  geom_sf(data = rich_map, aes(fill = total + 1)) +
  scale_fill_viridis_c("Species number\n(log-transformed)", trans = "log10",
                       labels = scales::number_format(accuracy = 1)) +
  labs(title = "Angiosperms", subtitle = "Projection EckertIV") +
  coord_sf(crs = eckertIV) +
  theme_void()
```

```{r, fig.cap = "", out.width = "90%",echo = FALSE}
knitr::include_graphics("../man/figures/angio_rich_map.png")
```

<br>
By customizing the code above, you can also produce a nicer map:

```{r, fig.cap = "", out.width = "90%",echo = FALSE}
knitr::include_graphics("../man/figures/angio_rich_map2.png")
```

<br>
Below is the `R` code to produce the above map if interested.
<br>

<details><summary>Fancier code</summary>
```{r, echo = TRUE, eval=FALSE}
# Background box
xmin <- st_bbox(world)[["xmin"]]; xmax <- st_bbox(world)[["xmax"]]
ymin <- st_bbox(world)[["ymin"]]; ymax <- st_bbox(world)[["ymax"]]
bb <- sf::st_union(sf::st_make_grid(st_bbox(c(xmin = xmin,
                                              xmax = xmax,
                                              ymax = ymax,
                                              ymin = ymin),
                                            crs = st_crs(4326)),
                                    n = 100))

# Equator line
equator <- st_linestring(matrix(c(-180, 0, 180, 0), ncol = 2, byrow = TRUE))
equator <- st_sfc(equator, crs = st_crs(world))

# Color code from Barthlott 2007
hexcode_barthlott2007 <- c("#fbf9ed", "#f3efcc", "#f6e39e", "#cbe784",
                           "#65c66a", "#0e8d4a", "#4a6fbf",
                           "#b877c2", "#f24dae", "#ed1c24")

ggplot(world) +
  geom_sf(data = bb, fill = "aliceblue") +
  geom_sf(data = equator, color = "gray50", linetype = "dashed",
          linewidth = 0.1) +
  geom_sf(data = world_countries, fill = "antiquewhite1", color = NA) +
  geom_sf(color = "gray50", linewidth = 0.1) +
  geom_sf(data = bb, fill = NA) +
  geom_sf(data = rich_map,
          aes(fill = ifelse(rich_map$entity_class %in%
                              c("Island/Mainland", "Mainland",
                                "Island Group", "Island Part"),
                            total + 1, NA)),
          size = 0.1) +
  geom_point(data = rich_map,
             aes(color = ifelse(rich_map$entity_class %in%
                                  c("Island"),
                                total + 1, NA),
                 geometry = geometry),
             stat = "sf_coordinates", size = 1, stroke = 0.5) +
  scale_color_gradientn(
    "Species number", trans = "log10", limits = c(1, 40000), 
    colours = hexcode_barthlott2007,
    breaks = c(1, 10, 100, 1000, 10000, 40000),
    labels = c(1, 10, 100, 1000, 10000, 40000),
    na.value = "transparent") +
  scale_fill_gradientn(
    "Species number", trans = "log10", limits = c(1, 40000), 
    colours = hexcode_barthlott2007,
    breaks = c(1, 10, 100, 1000, 10000, 40000),
    labels = c(1, 10, 100, 1000, 10000, 40000),
    na.value = "transparent") +
  labs(title = "Angiosperms", subtitle = "Projection EckertIV") +
  coord_sf(crs = eckertIV) +
  theme_void()
```
</details>
<br><br>

We can also produce maps of richness at intermediate scales. Here is the code
and the map of Angiosperms in the Western Mediterranean basin.  
<br>

```{r, echo = TRUE, eval = FALSE, message=FALSE, fig.width = 8, fig.height = 4}
med_shape <- gift_shapes[which(gift_shapes$entity_ID %in% 
                                 unique(medit[[2]]$entity_ID)), ]

med_rich <- angio_rich[which(angio_rich$entity_ID %in% 
                               unique(medit[[2]]$entity_ID)), ]

med_map <- dplyr::left_join(med_shape, med_rich, by = "entity_ID") %>%
  dplyr::filter(stats::complete.cases(total))

ggplot(world) +
  geom_sf(color = "gray50") +
  geom_sf(data = western_mediterranean,
          fill = "darkblue", color = "black", alpha = 0.1, size = 1) +
  geom_sf(data = med_map, aes(fill = total)) +
  scale_fill_viridis_c("Species number") +
  labs(title = "Angiosperms in the Western Mediterranean basin") +
  lims(x = c(-20, 20), y = c(24, 48)) +
  theme_void()
```

```{r, fig.cap = "", out.width = "90%",echo = FALSE}
knitr::include_graphics("../man/figures/angio_medit.png")
```

<br><br>

# 2. Distribution of species

The `GIFT` `R` package also allows for retrieving the spatial distribution of a
focal plant species.
<br>

## 2.1. Available species
To know what plant species are available, you can first run the function
`GIFT_species()`.

```{r, echo = TRUE, eval = FALSE}
all_sp <- GIFT_species()
```

364571 species are currently available in the database. This number may
increase with new releases of the database.
See the
[dedicated section](https://biogeomacro.github.io/GIFT/articles/GIFT_advanced_users.html#species)
in the advanced vignettes for more details.

## 2.2. Species names and taxonomic harmonization

Since GIFT is a collection of checklists in which authors use their own
taxonomic knowledge to describe species, there is a step of taxonomic
harmonization when including checklists in the database. The most commonly used
backbone is the
[World Checklists of Vascular Plants (WCVP)](https://powo.science.kew.org/).
<br>
Both original and harmonized names are stored in the database and you can use
the `GIFT_species_lookup()` function to look up the differences for particular
species. For example, the wood anemone *Anemonoides nemorosa*.
<br>

```{r anemone_lookup}
anemone_lookup <- GIFT_species_lookup(genus = "Anemonoides",
                                      epithet = "nemorosa")

kable(anemone_lookup, "html") %>%
  kable_styling(full_width = FALSE)
```

Looking at the output table, you can see the original species names and their
identification numbers (`name_ID`) **before** taxonomic harmonization. The
species names and IDs **after** taxonomic harmonization are the last columns on
the right starting with the prefix `work_`.

<br><br>

## 2.3. Species distribution
<br>
Now that we have a focal species and its harmonized name, we can retrieve its
distribution using `GIFT_species_distribution()`.
<br>
Note that here we set the `aggregation` argument to `TRUE` in order to
have only one floristic status per polygon. See the function's help page for
more details.
<br>

```{r}
anemone_distr <- GIFT_species_distribution(
  genus = "Anemonoides", epithet = "nemorosa", aggregation = TRUE)

anemone_statuses <- anemone_distr %>%
  mutate(native = ifelse(native == 1, "native", "non-native"),
         naturalized = ifelse(naturalized == 1, "naturalized",
                              "non-naturalized"),
         endemic_list = ifelse(endemic_list == 1, "endemic_list",
                               "non-endemic_list")) %>%
  dplyr::select(entity_ID, native, naturalized, endemic_list)

table(anemone_statuses$endemic_list)
```

This species is not listed as endemic in any of the GIFT polygons. Let's check
the places where it is listed as native or naturalized.
<br>

```{r}
table(paste(anemone_statuses$native, anemone_statuses$naturalized,
            sep = "_"))
```

Looking at the different combinations of statuses, we can distinguish several
situations: in 13 polygons, there is no status available. The species is listed
as native and non-naturalized (or naturalized status is missing) in 99+12=111
polygons. It is naturalized and non native in 5 polygons.
<br>
More surprising are the cases where the species is non-native and
non-naturalized, which happens in 2 polygons. This particular combination
can occur in unstable cases where the species is in the process of becoming
naturalized.
<br>

Now that we know in which polygons the species occurs and with what status, we
can map its distribution using the GIFT shapes we retrieved earlier with
`GIFT_shapes()`.

```{r, echo = TRUE, eval = FALSE, fig.width = 8, fig.height = 4}
# We rename the statuses based on the distinct combinations
anemone_statuses <- anemone_statuses %>%
  mutate(Status = case_when(
    native == "native" & naturalized == "non-naturalized" ~ "native",
    native == "native" & is.na(naturalized) ~ "native",
    native == "non-native" & is.na(naturalized) ~ "non-native",
    native == "non-native" & naturalized == "naturalized" ~ "naturalized",
    native == "non-native" & naturalized == "non-naturalized" ~ "non-native",
    is.na(native) & is.na(naturalized) ~ "unknown"
  ))

# Merge with the shapes
anemone_shape <- gift_shapes[which(gift_shapes$entity_ID %in% 
                                     unique(anemone_distr$entity_ID)), ]
anemone_map <- dplyr::left_join(anemone_shape, anemone_statuses,
                                by = "entity_ID")

# Area of distribution with floristic status
ggplot(world) +
  geom_sf(color = "gray70") +
  geom_sf(data = anemone_map, color = "black", aes(fill = as.factor(Status))) +
  scale_fill_brewer("Status", palette = "Set2") +
  labs(title = expression(paste("Distribution map of ",
                                italic("Anemonoides nemorosa"))),
       subtitle = "Unprojected (GCS: WGS84)") +
  lims(x = c(-65, 170), y = c(-45, 70)) +
  theme_void()
```

```{r, fig.cap = "", out.width = "90%",echo = FALSE}
knitr::include_graphics("../man/figures/anemone_map.png")
```

<br>
By customizing the code above, you can also produce a nicer map:

```{r, fig.cap = "", out.width = "90%",echo = FALSE}
knitr::include_graphics("../man/figures/anemone_map2.png")
```

<br>
Below is the `R` code to produce the above map if interested.
<br>

<details><summary>Fancier code</summary>
```{r, eval=FALSE}
anemone_map_plot_bg_parts <-
  ggplot(world) +
  geom_sf(data = bb, fill = "aliceblue", color = NA) +
  geom_sf(data = equator, color = "gray50", linetype = "dashed",
          linewidth = 0.1) +
  geom_sf(data = world_countries, fill = "antiquewhite1", color = NA) +
  geom_sf(color = "gray50", linewidth = 0.1) +
  geom_sf(data = bb, fill = NA) +
  geom_sf(data = anemone_map, color = "black", aes(fill = as.factor(Status))) +
  scale_fill_manual("Status",
                    values = c("native" = "#2c7bb6",
                               "naturalized" = "#d7191c",
                               "non-native" = "#fdae61",
                               "unknown" = "#abd9e9")) +
  labs(title = expression(paste("b) Distribution map of ",
                                italic("Anemonoides nemorosa")))) +
  theme_void() +
  theme(axis.title = element_blank(),
        axis.text = element_blank(),
        axis.ticks = element_blank())

(anemone_map_plot_bg_parts +
    lims(x = c(-69, 61), y = c(37, 70)) + # Europe & Newfoundland
    theme(panel.border = element_rect(fill = NA, linewidth = 1)) +
    theme(legend.position = "bottom")
  |
    anemone_map_plot_bg_parts +
    lims(x = c(165, 178), y = c(-47, -35)) + # New Zealand
    labs(title = "") +
    guides(fill = "none") +
    theme(panel.border = element_rect(fill = NA, linewidth = 1)))
```
</details>
<br>

<br><br>

# 3. Trait data

Trait information at the species or at a higher taxonomic level is also
provided in the `GIFT` R package.
<br>

## 3.1. Metadata
<br>
There are many functional traits available in GIFT. Each of these traits has
an identification number called `trait_ID`. Since the two functions for
retrieving trait values, `GIFT_traits()` and `GIFT_traits_raw()`, rely on these
IDs, the first step is to call the function `GIFT_traits_meta()` to know what
the ID of the desired trait is.
<br>
For example, let's say we want to retrieve the maximum vegetative heights of
plant species.
<br>
```{r}
trait_meta <- GIFT_traits_meta()
trait_meta[which(trait_meta$Trait2 == "Plant_height_max"), ]
```
<br>
We can see that the ID of this trait is `1.6.2`. Now that we have the ID, we
can use `GIFT_traits()` to retrieve the growth form values for different plant
species.

<br>

## 3.2. Trait values
### 3.2.1. Species level
There are two functions to access trait values. First, `GIFT_traits_raw()`
returns all trait values for a given species and a given trait. These trait
values can then vary.
<br>
Second, `GIFT_traits()` returns an aggregated trait value at the species level.
The aggregation simply takes the mean for continuous traits or the most
frequent entry for categorical traits. However, for some specific cases, the
aggregation takes either the minimum or the maximum, as for the trait we chose.
<br>
Let's retrieve the raw and aggregated values for the maximum vegetative height
of plants (trait_ID 1.6.2).
<br>

```{r, echo = TRUE, eval = FALSE}
height <- GIFT_traits(trait_IDs = c("1.6.2"), agreement = 0.66,
                      bias_ref = FALSE, bias_deriv = FALSE)

height_raw <- GIFT_traits_raw(trait_IDs = c("1.6.2"))

# Raw values
as.numeric(height_raw[which(height_raw$work_species == "Fagus sylvatica"),
                      "trait_value"])

# Aggregated value
as.numeric(height[which(height$work_species == "Fagus sylvatica"),
                  "trait_value_1.6.2"])
```

There were three maximum heights for *Fagus sylvatica*, 30, 35, and 50 meters,
which led to an aggregated value of 50 meters.
<br>
And if you want to look up the references that led to the aggregated trait
value, you can run this chunk:  

```{r, echo = TRUE, eval=FALSE}
references <- GIFT_references(GIFT_version = "beta")

unique(unlist(strsplit(height$references_1.6.2, ",")))

references <- references[
  which(references$ref_ID %in% 
          unique(unlist(strsplit(height$references_1.6.2, ",")))), ]
references[1:2, ]
```

<br>

### 3.2.2. Taxonomic level

Traits can also be retrieved at a higher taxonomic level, using
`GIFT_traits_tax()`.
<br>
As an example, we here attempt to retrieve three traits at the family level.
The three traits asked are woodiness (`trait_ID = "1.1.1"`),
growth form (`trait_ID = "1.2.1"`) and whether the plant is a climber
(`trait_ID = "1.4.1"`).

```{r, echo=TRUE, eval=FALSE}
trait_tax <- GIFT_traits_tax(trait_IDs = c("1.1.1", "1.2.1", "1.4.1"),
                             bias_ref = FALSE, bias_deriv = FALSE)

trait_tax[1:3, ]
```

Of the three requested traits, the growth form was not available at the family
level. Therefore, the output table contains trait values for the other two
traits at the family level.

<br>

## 3.3. Trait coverage

We can also retrieve trait coverage information for polygons, using the same
function as for species richness `GIFT_coverage()`.
<br>
In combination with the previously loaded shapes, we can also map the trait
coverage.
<br>

```{r, echo = TRUE, eval = FALSE, fig.width = 8, fig.height = 4}
angio_height <- GIFT_coverage(what = "trait_coverage",
                              taxon_name = "Angiospermae",
                              trait_ID = "1.6.2")

angio_height_shape <- gift_shapes[which(gift_shapes$entity_ID %in% 
                                          unique(angio_height$entity_ID)), ]

angio_height_map <- dplyr::left_join(
  angio_height_shape, angio_height, by = "entity_ID")

angio_height_map <- angio_height_map[complete.cases(angio_height_map$native), ]

ggplot(world) +
  geom_sf(color = "gray50") +
  geom_sf(data = angio_height_map[complete.cases(angio_height_map$native), ],
          aes(fill = native)) +
  scale_fill_viridis_c("Coverage (%)") +
  labs(title = "Coverage for maximal vegetative height of Angiosperms",
       subtitle = "Projection EckertIV") +
  coord_sf(crs = eckertIV) +
  theme_void()
```

```{r, fig.cap = "", out.width = "90%",echo = FALSE}
knitr::include_graphics("../man/figures/angio_height_plot.png")
```

<br>
By customizing the code above, you can also produce a nicer map:

```{r, fig.cap = "", out.width = "90%",echo = FALSE}
knitr::include_graphics("../man/figures/angio_height_plot2.png")
```

<br>
Below is the `R` code to produce the above map if interested.
<br>

<details><summary>Fancier code</summary>
```{r, eval=FALSE}
ggplot(world) +
  geom_sf(data = bb, fill = "aliceblue") +
  geom_sf(data = equator, color = "gray50", linetype = "dashed",
          linewidth = 0.1) +
  geom_sf(data = world_countries, fill = "antiquewhite1", color = NA) +
  geom_sf(color = "gray50", linewidth = 0.1) +
  geom_sf(data = bb, fill = NA) +
  geom_sf(data = angio_height_map,
          aes(fill = ifelse(angio_height_map$entity_class %in%
                              c("Island/Mainland", "Mainland",
                                "Island Group", "Island Part"),
                            100*native, NA)), size = 0.1) +
  geom_point(data = angio_height_map,
             aes(color = ifelse(angio_height_map$entity_class %in%
                                  c("Island"),
                                100*native, NA),
                 geometry = geometry),
             stat = "sf_coordinates", size = 1, stroke = 0.5) +
  scale_color_gradientn(
    "Coverage (%)", 
    colours = rev(RColorBrewer::brewer.pal(9, name = "PuBuGn")),
    limits = c(0, 100),
    na.value = "transparent") +
  scale_fill_gradientn(
    "Coverage (%)", 
    colours = rev(RColorBrewer::brewer.pal(9, name = "PuBuGn")),
    limits = c(0, 100),
    na.value = "transparent") +
  labs(title = "Coverage for maximal vegetative height of Angiosperms",
       subtitle = "Projection EckertIV") +
  coord_sf(crs = eckertIV) +
  theme_void()
```
</details>
<br>

<br><br>

# 4. Environmental variables

Finally, a set of summary statistics for many environmental variables can be
retrieved for each GIFT polygon.

## 4.1. Metadata
We here illustrate how to retrieve environmental variables, summarized at the
polygon level, for a subset of polygons.
<br>
We here retrieve environmental variables for the polygons falling into the
western Mediterranean basin, retrieved in the section 1.
<br>
To know what variables are available, you can run these two metadata functions: `GIFT_env_meta_misc()` and `GIFT_env_meta_raster()`. They respectively give
access to the list of miscellaneous variables and raster layers available in
the GIFT database.
<br>
The references to cite when using environmental variables are also available
through these functions (column `ref_long` of the outputs).  
<br>
```{r env_meta}
misc_env <- GIFT_env_meta_misc()
raster_env <- GIFT_env_meta_raster()
```
<br>

## 4.2. Environmental values
Let's say we want to retrieve the perimeter and biome of each polygon as
well as elevation and mean temperature. For these two raster layers, we need to
define summary statistics, since the polygons are usually larger than the
raster resolution.
<br>
There are many summary statistics available, you can check the help page of
`GIFT_env()` to see them all. Let's get the mean and median of the
elevation and the maximal value of the average temperature.
<br>
```{r, echo = TRUE, eval = FALSE}
med_env <- GIFT_env(entity_ID = unique(medit[[2]]$entity_ID),
                    miscellaneous = c("perimeter", "biome"),
                    rasterlayer = c("mn30_grd", "wc2.0_bio_30s_01"),
                    sumstat = list(c("mean", "med"), "max"))

med_env[1, ]
```

We see here that the region of El Hierro has an average altitude of 579 meters
above sea level and an average annual temperature of 21.6 degrees Celsius. 

<br>

## 4.3. Map

Using the previously loaded shapes, we can also map a specific environmental
variable for all GIFT polygons.

```{r, echo = TRUE, eval = FALSE, fig.width = 8, fig.height = 4}
world_temp <- GIFT_env(entity_ID = unique(angio_rich$entity_ID),
                       rasterlayer = c("wc2.0_bio_30s_01"),
                       sumstat = c("mean"))

temp_shape <- gift_shapes[which(gift_shapes$entity_ID %in% 
                                  unique(angio_rich$entity_ID)), ]

temp_map <- dplyr::left_join(temp_shape, world_temp, by = "entity_ID")

ggplot(world) +
  geom_sf(color = "gray50") +
  geom_sf(data = temp_map, aes(fill = mean_wc2.0_bio_30s_01)) +
  scale_fill_viridis_c("Celsius degrees") +
  labs(title = "Average temperature",
       subtitle = "Projection EckertIV") +
  coord_sf(crs = eckertIV) +
  theme_void()
```

```{r, fig.cap = "", out.width = "90%",echo = FALSE}
knitr::include_graphics("../man/figures/temp_plot.png")
```

<br>
By customizing the code above, you can also produce a nicer map:

```{r, fig.cap = "", out.width = "90%",echo = FALSE}
knitr::include_graphics("../man/figures/temp_plot2.png")
```

<br>
Below is the `R` code to produce the above map if interested.
<br>

<details><summary>Fancier code</summary>
```{r, eval=FALSE}
ggplot(world) +
  geom_sf(data = bb, fill = "aliceblue") +
  geom_sf(data = equator, color = "gray50", linetype = "dashed",
          linewidth = 0.1) +
  geom_sf(data = world_countries, fill = "antiquewhite1", color = NA) +
  geom_sf(color = "gray50", linewidth = 0.1) +
  geom_sf(data = bb, fill = NA) +
  geom_sf(data = temp_map,
          aes(fill = ifelse(temp_map$entity_class %in%
                              c("Island/Mainland", "Mainland",
                                "Island Group", "Island Part"),
                            mean_wc2.0_bio_30s_01, NA)), size = 0.1) +
  geom_point(data = temp_map,
             aes(color = ifelse(temp_map$entity_class %in%
                                  c("Island"),
                                mean_wc2.0_bio_30s_01, NA),
                 geometry = geometry),
             stat = "sf_coordinates", size = 1, stroke = 0.5) +
  scale_color_gradientn(
    "°C", 
    colours = RColorBrewer::brewer.pal(9, name = "Reds"),
    limits = c(-20, 30),
    na.value = "transparent") +
  scale_fill_gradientn(
    "°C", 
    colours = RColorBrewer::brewer.pal(9, name = "Reds"),
    limits = c(-20, 30),
    na.value = "transparent") +
  labs(title = "Average temperature",
       subtitle = "Projection EckertIV") +
  coord_sf(crs = eckertIV) +
  theme_void()
```
</details>
<br>

<br><br>

# 5. Phylogeny

To build a phylogeny for all species in the GIFT database, we have used the
U.PhyloMaker R package 
[Jin & Qian (2022)](https://doi.org/10.1016/j.pld.2022.12.007). For seed
plants, we
selected the megatree GBOTB.extended.WP.tre by
[Smith and Brown (2018)](https://doi.org/10.1002/ajb2.1019) and the phylogeny
for pteridophytes in
[Zanne et al. (2014)](https://www.nature.com/articles/nature12872).

```{r, eval=FALSE, echo=TRUE}
# Retrieving phylogeny, taxonomy and species from GIFT
phy <- GIFT_phylogeny(clade = "Tracheophyta", GIFT_version = "beta")
tax <- GIFT_taxonomy(GIFT_version = "beta")
gift_sp <- GIFT_species(GIFT_version = "beta")

gf <- GIFT_traits(trait_IDs = "1.2.1", agreement = 0.66, bias_ref = FALSE,
                  bias_deriv = FALSE, GIFT_version = "beta")
```

```{r, eval=FALSE, echo=TRUE}
# Replacing space with _ for the species names
gf$work_species <- gsub(" ", "_", gf$work_species, fixed = TRUE)
```


```{r, eval=FALSE, echo=TRUE}
# Retrieving family of each species
sp_fam <- GIFT_taxgroup(work_ID = unique(gift_sp$work_ID),
                        taxon_lvl = "family", GIFT_version = "beta")
sp_genus_fam <- data.frame(
  work_ID = unique(gift_sp$work_ID),
  work_species = unique(gift_sp$work_species),
  family = sp_fam)
sp_genus_fam <- left_join(sp_genus_fam,
                          gift_sp[, c("work_ID", "work_genus")],
                          by = "work_ID")
colnames(sp_genus_fam)[colnames(sp_genus_fam) == "work_genus"] <- "genus"

# Problem with hybrid species on the tip labels of the phylo tree
phy$tip.label[substring(phy$tip.label, 1, 2) == "x_"] <-
  substring(phy$tip.label[substring(phy$tip.label, 1, 2) == "x_"],
            3,
            nchar(phy$tip.label[substring(phy$tip.label, 1, 2) == "×_"]))

phy$tip.label[substring(phy$tip.label, 1, 2) == "×_"] <-
  substring(phy$tip.label[substring(phy$tip.label, 1, 2) == "×_"],
            3,
            nchar(phy$tip.label[substring(phy$tip.label, 1, 2) == "×_"]))
```

In the next chunk, we calculate the trait coverage (for growth form) at the
genus and family level.

```{r, eval=FALSE, echo=TRUE}
sp_genus_fam <- left_join(sp_genus_fam,
                          gf[, c("work_ID", "trait_value_1.2.1")],
                          by = "work_ID")

genus_gf <- sp_genus_fam %>%
  group_by(genus) %>%
  mutate(prop_gf = round(100*sum(is.na(trait_value_1.2.1))/n(), 2)) %>%
  ungroup() %>%
  dplyr::select(-work_ID, -work_species, -family, -trait_value_1.2.1) %>%
  distinct(.keep_all = TRUE)

fam_gf <- sp_genus_fam %>%
  group_by(family) %>%
  mutate(prop_gf = round(100*sum(is.na(trait_value_1.2.1))/n(), 2)) %>%
  ungroup() %>%
  dplyr::select(-work_ID, -work_species, -genus, -trait_value_1.2.1) %>%
  distinct(.keep_all = TRUE)

sp_genus_fam$species <- gsub("([[:punct:]])|\\s+", "_",
                             sp_genus_fam$work_species)

# Keeping one species per genus only
one_sp_per_gen <- data.frame()
for(i in 1:n_distinct(sp_genus_fam$genus)){ # loop over genera
  # Focal genus
  focal_gen <- unique(sp_genus_fam$genus)[i]
  # All species in that genus
  gen_sp_i <- sp_genus_fam[which(sp_genus_fam$genus == focal_gen),
                           "species"]
  # Species from the genus available in the phylogeny
  gen_sp_i <- gen_sp_i[gen_sp_i %in% phy$tip.label]
  # Taking the first one (if at least one is available)
  gen_sp_i <- gen_sp_i[1]
  
  one_sp_per_gen <- rbind(one_sp_per_gen,
                          data.frame(species = gen_sp_i,
                                     genus = focal_gen))
}

# Adding the trait coverage per genus
one_sp_per_gen <- left_join(one_sp_per_gen, genus_gf, by = "genus")

# Adding the trait coverage per family
one_sp_per_gen <- left_join(one_sp_per_gen,
                            sp_genus_fam[!duplicated(sp_genus_fam$genus),
                                         c("genus", "family")],
                            by = "genus")
colnames(one_sp_per_gen)[colnames(one_sp_per_gen) == "prop_gf"] <-
  "prop_gf_gen"
one_sp_per_gen <- left_join(one_sp_per_gen, fam_gf, by = "family")
colnames(one_sp_per_gen)[colnames(one_sp_per_gen) == "prop_gf"] <-
  "prop_gf_fam"
```

We now prune the tree at the genus level.

```{r, eval=FALSE, echo=TRUE}
phy_gen <- ape::keep.tip(
  phy = phy,
  tip = one_sp_per_gen[complete.cases(one_sp_per_gen$species), "species"])
```

In the following plot, there is only one tip per genus. The two outer rings
illustrate the coverage of growth form per genus and per family (outer ring).
For the family ring, the width of each family is set by the number of genera
it contains.
<br>
To get the following plot, you need to install/load the following packages:

```{r, echo=TRUE, eval = FALSE}
library("BiocManager")
install("ggtree")
library("ggtree")
library("tidytree")
install("ggtreeExtra")
library("ggtreeExtra")
```

<br>

```{r, echo = TRUE, eval = FALSE, fig.width = 6, fig.height = 6}
ggtree(phy_gen, color = "grey70", layout = "circular") %<+% one_sp_per_gen +
  geom_fruit(geom = geom_tile,
             mapping = aes(fill = prop_gf_gen),
             width = 50,
             offset = 0.1) +
  geom_fruit(geom = geom_tile,
             mapping = aes(color = prop_gf_fam, fill = prop_gf_fam),
             width = 50,
             offset = 0.1,
             show.legend = FALSE) +
  scale_color_viridis_c() +
  scale_fill_viridis_c("Growth form availability per genus (%)") +
  theme(legend.position = "bottom")
```

```{r, fig.cap = "", out.width = "100%",echo = FALSE}
knitr::include_graphics("../man/figures/phy_tree_plot.png")
```

<br><br>

# 6. References

When using the GIFT database and the GIFT R package, here are the two
resources to cite:  

Denelle, P., Weigelt, P., & Kreft, H. (2023). GIFT—An R package to access the
Global Inventory of Floras and Traits. *Methods in Ecology and Evolution*,
00, 1–11.
[https://doi.org/10.1111/2041-210X.14213](https://doi.org/10.1111/2041-210X.14213). 
  
Weigelt, P., König, C. & Kreft, H. (2020) GIFT – A Global Inventory of Floras
and Traits for macroecology and biogeography. *Journal of Biogeography*, 
[https://doi.org/10.1111/jbi.13623](https://doi.org/10.1111/jbi.13623).