ulan_match: Match names to the Getty ULAN
In mdlincoln/ulanr: Reconcile Names to the Getty Union List of Artist Names

Description Usage Arguments Value Functions Note Examples

Queries the Getty ULAN to find matching entries for a given string. You may filter the results by specifying an early or late date.

ulan_match(names, early_year = -9999, late_year = 2090,
  strictly_between = FALSE, method = c("sparql", "local"),
  max_results = 5, cutoff_score = NULL)

ulan_id(names, early_year = -9999, late_year = 2090,
  strictly_between = FALSE, method = c("sparql", "local"),
  max_results = 1)

ulan_data(names, early_year = -9999, late_year = 2090,
  strictly_between = FALSE, method = c("sparql", "local"),
  max_results = 1)

`names`	A character vector of names to match to a canonical ULAN id.
`early_year`	Match only artists who died after this year. Like `late_year`, this argument should be a numeric vector of length 1, or of the same length as `names`. If length 1, the same date restrictions will be used to match every value of `names`. Otherwise, each name match can be restricted to its own pair of early_year and late_year. If no `early_year` or `late_year` are specified, then artists from all time periods will be eligible for matching. Any NA values in `early_year` or `late_year` will be coerced to default maxima and minima.
`late_year`	Match only artists who were born before this year.
`strictly_between`	Method for filtering search results using the early_year/late_year parameters. TRUE will only include artists whose life dates fall within the range [late_year, early_year]. FALSE (the default) will include artists whose life dates intersect with [early_year, late_year]
`method`	This value determines which method will be used to match the name to a canonical ULAN id. `sparql` will query the Getty's live endpoint, relying on its Lucene index for finding close matches, while `local` instead uses string cosine similarity based on a local table of ULAN entries.
`max_results`	The maximum number of results to return. Defaults to 5. Depending on the query, the actual number of results returned may be lower. When `method = "sparql"` is used, values over 50 will be ignored.
`cutoff_score`	The minimum similarity score that must be returned by the chosen method for a candidate to be included in results. `NULL` will use default values for each method: 0.95 for the `local` method, and 3 for the `sparql` method.

A named list of data.frames, one per submitted name, with 7 columns and no more than max_results rows:

id: integer. ULAN id
pref_name: character. ULAN preferred name
birth_year: integer. Artist birth year, if assigned.
death_year: integer. Artist death year, if assigned
gender: character. Artist gender, if assigned.
nationality: character. Artist nationality, if assigned.
score: numeric. The score of the result. When method = "sparql", this is the Lucene index score. When method = "local", it will instead be a cosine similarity score. Results with a score below cutoff_score are dropped.

Unmatched names will return a data.frame with NAs for all values save name.

ulan_id: Return a vector of matching ULAN IDs. This function is a wrapper around ulan_match that simply returns a vector of the top match ID for each name given (as if max_results were 1). max_results will be ignored.
ulan_data: Return a dataframe of matching ULAN IDs with attributes, with an additional column names containing the original input vector. This function is a wrapper around ulan_match.

cutoff_score will be ingored for

method = "sparql" requires an internet connection.

## Not run: ulan_id("Rembrandt", early_year = 1600,
                 late_year = 1700, method = "sparql")
## End(Not run)
## Not run: ulan_id(c("Rembrandt", "Rothko"), early_year = c(1600, 1900),
                 late_year = c(1700, 2000), method = "sparql")
## End(Not run)