dwc_mapping.Rmd

---
title: "Darwin Core mapping"
subtitle: "For: my_dataset_title"
author:
- author_1
- author_2
date: "`r Sys.Date()`"
output:
  html_document:
    df_print: paged
    number_sections: yes
    toc: yes
    toc_depth: 3
    toc_float: yes
#  pdf_document:
#    df_print: kable
#    number_sections: yes
#    toc: yes
#    toc_depth: 3
---

# Setup 

```{r setup, include = FALSE}
knitr::opts_chunk$set(echo = TRUE, warning = FALSE, message = TRUE)
```

Load libraries:

```{r message = FALSE}
library(tidyverse)      # To do data science
library(tidylog)        # To provide feedback on dplyr functions
library(magrittr)       # To use %<>% pipes
library(here)           # To find files
library(janitor)        # To clean input data
library(readxl)         # To read Excel files
library(digest)         # To generate hashes
library(rgbif)          # To use GBIF services
```

# Read source data

Create a data frame `input_data` from the source data:

```{r}
input_data <- read_excel(path = here("data", "raw", "checklist.xlsx")) 
```

Preview data:

```{r}
input_data %>% head(n = 5)
```

# Process source data

## Tidy data

Clean data somewhat:

```{r}
input_data %<>% remove_empty("rows")
```

## Scientific names

Use the [GBIF nameparser](https://www.gbif.org/tools/name-parser) to retrieve nomenclatural information for the scientific names in the checklist:

```{r}
parsed_names <- input_data %>%
  distinct(scientific_name) %>%
  pull() %>% # Create vector from dataframe
  parsenames() # An rgbif function
```

Show scientific names with nomenclatural issues, i.e. not of `type = SCIENTIFIC` or that could not be fully parsed. Note: these are not necessarily incorrect.

```{r}
parsed_names %>%
  select(scientificname, type, parsed, parsedpartially, rankmarker) %>%
  filter(!(type == "SCIENTIFIC" & parsed == "TRUE" & parsedpartially == "FALSE"))
```

Correct names and reparse:

```{r correct and reparse}
input_data %<>% mutate(scientific_name = recode(scientific_name,
  "AseroÙ rubra" = "Asero rubra"
))

# Redo parsing
parsed_names <- input_data %>%
  distinct(scientific_name) %>%
  pull() %>%
  parsenames()

# Show names with nomenclatural issues again
parsed_names %>%
  select(scientificname, type, parsed, parsedpartially, rankmarker) %>%
  filter(!(type == "SCIENTIFIC" & parsed == "TRUE" & parsedpartially == "FALSE"))
```

## Taxon ranks

The nameparser function also provides information about the rank of the taxon (in `rankmarker`). Here we join this information with our checklist. Cleaning these ranks will done in the Taxon Core mapping:

```{r}
input_data %<>% left_join(
  parsed_names %>%
  select(scientificname, rankmarker),
  by = c("scientific_name" = "scientificname"))
```

## Taxon IDs

To link taxa with information in the extension(s), each taxon needs a unique and relatively stable `taxonID`. Here we create one in the form of `dataset_shortname:taxon:hash`, where `hash` is unique code based on scientific name and kingdom (that will remain the same as long as scientific name and kingdom remain the same):

```{r}
vdigest <- Vectorize(digest) # Vectorize digest function to work with vectors
input_data %<>% mutate(taxon_id = paste(
  "my_dataset_shortname", # e.g. "alien-fishes-checklist"
  "taxon",
  vdigest(paste(scientific_name, kingdom), algo = "md5"),
  sep = ":"
))
```

## Preview data

Show the number of taxa and distributions per kingdom and rank:

```{r}
input_data %>%
  group_by(kingdom, rankmarker) %>%
  summarize(
    `# taxa` = n_distinct(taxon_id),
    `# distributions` = n()
  ) %>%
  adorn_totals("row")
```

Preview data:

```{r}
input_data %>% head()
```

# Taxon core

## Pre-processing

Create a dataframe with unique taxa only (ignoring multiple distribution rows):

```{r}
taxon <- input_data %>% distinct(taxon_id, .keep_all = TRUE)
```

## Term mapping

Map the data to [Darwin Core Taxon](http://rs.gbif.org/core/dwc_taxon_2015-04-24.xml).

Start with record-level terms which contain metadata about the dataset (which is generally the same for all records).

### language

```{r}
taxon %<>% mutate(dwc_language = "en") # e.g. "en"
```

### license

```{r}
taxon %<>% mutate(dwc_license = "my_license") # e.g. "http://creativecommons.org/publicdomain/zero/1.0/"
```

### rightsHolder

```{r}
taxon %<>% mutate(dwc_rightsHolder = "my_rights_holder") # e.g. "INBO"
```

### datasetID

```{r}
taxon %<>% mutate(dwc_datasetID = "my_dataset_doi") # e.g. "https://doi.org/10.15468/xvuzfh"
```

### institutionCode

```{r}
taxon %<>% mutate(dwc_institutionCode = "my_institution_code") # e.g. "INBO"
```

### datasetName

```{r}
taxon %<>% mutate(dwc_datasetName = "my_dataset_title") # e.g. "Checklist of non-native freshwater fishes in Flanders, Belgium"
```

The following terms contain information about the taxon:

### taxonID

```{r}
taxon %<>% mutate(dwc_taxonID = taxon_id)
```

### scientificName

```{r}
taxon %<>% mutate(dwc_scientificName = scientific_name)
```

### kingdom

Inspect values:

```{r}
taxon %>%
  group_by(kingdom) %>%
  count()
```

Map values:

```{r}
taxon %<>% mutate(dwc_kingdom = kingdom)
```

### taxonRank

Inspect values:

```{r}
taxon %>%
  group_by(rankmarker) %>%
  count()
```

Map values by recoding to the [GBIF rank vocabulary](http://rs.gbif.org/vocabulary/gbif/rank_2015-04-24.xml):

```{r}
taxon %<>% mutate(dwc_taxonRank = recode(rankmarker,
  "agg."      = "speciesAggregate",
  "infrasp."  = "infraspecificname",
  "sp."       = "species",
  "var."      = "variety",
  .default    = "",
  .missing    = ""
))
```

Inspect mapped values: 

```{r}
taxon %>%
  group_by(rankmarker, dwc_taxonRank) %>%
  count()
```

### nomenclaturalCode

```{r}
taxon %<>% mutate(dwc_nomenclaturalCode = "my_nomenclaturalCode") # e.g. "ICZN"
```

## Post-processing

Only keep the Darwin Core columns:

```{r}
taxon %<>% select(starts_with("dwc_"))
```

Drop the `dwc_` prefix:

```{r}
colnames(taxon) <- str_remove(colnames(taxon), "dwc_")
```

Preview data:

```{r}
taxon %>% head()
```

Save to CSV:

```{r}
write_csv(taxon, here("data", "processed", "taxon.csv"), na = "")
```

# Distribution extension

## Pre-processing

Create a dataframe with all data:

```{r}
distribution <- input_data
```

## Term mapping

Map the data to [Species Distribution](http://rs.gbif.org/extension/gbif/1.0/distribution.xml).

### taxonID

```{r}
distribution %<>% mutate(dwc_taxonID = taxon_id)
```

### locality

Inspect values:

```{r}
distribution %>%
  group_by(country_code, locality) %>%
  count()
```

Map values to `input_locality` if provided, otherwise use the country name:

```{r}
distribution %<>% mutate(dwc_locality = case_when(
  !is.na(locality) ~ locality,
  country_code == "BE" ~ "Belgium",
  country_code == "GB" ~ "United Kingdom",
  country_code == "MK" ~ "Macedonia",
  country_code == "NL" ~ "The Netherlands",
  TRUE ~ "" # In other cases leave empty
))
```

Inspect mapped values: 

```{r}
distribution %>%
  group_by(country_code, locality, dwc_locality) %>%
  count()
```

### countryCode

Inspect values:

```{r}
distribution %>%
  group_by(country_code) %>%
  count()
```

Map values:

```{r}
distribution %<>% mutate(dwc_countryCode = country_code) 
```

### occurrenceStatus

Inspect values:

```{r}
distribution %>%
  group_by(occurrence_status) %>%
  count()
```

Map values:

```{r}
distribution %<>% mutate(dwc_occurrenceStatus = occurrence_status) 
```

### threatStatus

Inspect values:

```{r}
distribution %>%
  group_by(threat_status) %>%
  count()
```

Map values by recoding to the [IUCN threat status vocabulary](http://rs.gbif.org/vocabulary/iucn/threat_status.xml):

```{r}
distribution %<>% mutate(dwc_threatStatus = recode(threat_status,
  "endangered" = "EN",
  "vulnerable" = "VU"
))
```

Inspect mapped values: 

```{r}
distribution %>%
  group_by(threat_status, dwc_threatStatus) %>%
  count()
```

### source

Inspect values:

```{r}
distribution %>%
  group_by(source) %>%
  count() %>%
  head() # Remove to show all values
```

Map values:

```{r}
distribution %<>% mutate(dwc_source = source) 
```

### occurrenceRemarks

Inspect values:

```{r}
distribution %>%
  group_by(remarks) %>%
  count() %>%
  head() # Remove to show all values
```

Map values:

```{r}
distribution %<>% mutate(dwc_occurrenceRemarks = remarks) 
```

## Post-processing

Only keep the Darwin Core columns:

```{r}
distribution %<>% select(starts_with("dwc_"))
```

Drop the `dwc_` prefix:

```{r}
colnames(distribution) <- str_remove(colnames(distribution), "dwc_")
```

Preview data:

```{r}
distribution %>% head()
```

Save to CSV:

```{r}
write_csv(distribution, here("data", "processed", "distribution.csv"), na = "")
```