In this example, we query Arctos for all specimens of genus Microtus collected in Mongolia, currently held in the collection of mammals of the Museum of Southwestern Biology. After that we filter the downloaded data to find the specimens that were found to have nematods.
To begin, make sure to load the library:
First, we can view a list of all parameters we can ask Arctos to return by calling:
# Request a list of all result parameters. These are the names that can show up
# as columns in a dataframe returned by ArctosR.
result_params <- get_result_parameters()
# Explore all parameters.
View(result_params)
Each parameter has a category. If we are only interested in certain
categories of result parameter, we can filter the data.frame returned by
get_result_parameters()
like so:
# Check only core and record parameters.
result_params[result_params$category %in% c("core", "record"), 1:3]
Next, we find the number of specimens matching the query we wanted to perform by calling:
# Request just the number of records matching a query.
count <- get_record_count(country = "Mongolia", genus = "Microtus",
guid_prefix = "MSB:Mamm",
api_key=YOUR_API_KEY)
It is helpful to call this first to make sure that we aren’t asking
for too many items from Arctos. Next, to download data, we specify our
query, and then use the columns parameter to list all of the result
parameters we want from our query. Finally, we specify that we want to
download all records. This is necessary because Arctos paginates
results, returning only 100 at a time. Setting
all_records = TRUE
lets get_records
repeatedly
query Arctos until it receives all records from the search.
# Request to download all available data matching a query (specific columns).
microtus <- get_records(country = "Mongolia", genus = "Microtus",
guid_prefix = "MSB:Mamm",
columns = list("guid", "scientific_name", "dec_long",
"dec_lat", "verbatim_date", "parts",
"partdetail"),
all_records = TRUE,
api_key=YOUR_API_KEY)
In Arctos, some table entries, such as partdetail
are
themselves tables. We can expand these tables into data.frames
using:
# Expand a column that contains complex information in JSON format
expand_column(query = microtus, column_name = "partdetail")
The object returned from get_records
contains both data
and metadata about the request which is useful for making further
requests based on the data returned from the first request. To get the
data.frame of the response, use response_data
and pass in
the response:
Now, we write a custom function to check the partdetail
entries of each specimen for whether or not nematodes were present in
the specimen.
# Filter the data to keep only Microtus records in which nematodes were found
## Whole-word match for 'nematode' or 'nematodes'
pattern <- "\\bnematodes?\\b"
## A small function to check within data.frames in partdetail
has_nematode <- function(df) {
if (!is.data.frame(df) || is.null(df[["part_name"]])) {
return(FALSE)
} else {
return(any(grepl(pattern, df[["part_name"]], ignore.case = TRUE, perl = TRUE)))
}
}
## TRUE/FALSE mask for mic_df rows (whether they had nematodes or not)
mask <- vapply(microtus_df$partdetail, has_nematode, logical(1))
## Subset of microtus_df with matches
microtus_df_nematode <- microtus_df[mask, 1:5, drop = FALSE]
nrow(microtus_df) # Number of Microtus from Mongolia
nrow(microtus_df_nematode) # Number of Microtus from Mongolia that had nematodes