Apart from Geonames [http://www.geonames.org/], the online gazetteers referenced by the geoparser are all hosted by Edina [http://edina.ac.uk/]‘s Unlock service [http://edina.ac.uk/unlock/]. There are two web services, Unlock Places and Unlock Text.
Unlock Places [http://edina.ac.uk/unlock/places/] provides an API to allow the user to match a placename string against one or more gazetteers. The geoparser uses this to find candidate locations for placename strings extracted by the geotagging step. The API provides a simple and flexible way to search different gazetteers, with results available in a range of formats.
For example:
returns an xml file of candidate locations matching “Edinburgh” (108 candidates at time of writing) from all the available gazetteers, including whatever attributes the gazetteer provides - such as latitude, longitude, feature codes, population and alternative names.
To restrict the search to a particular gazetteer one simply specifies it in the request:
This returns matches for “Halicarnassus” in the copy of Pleiades+ hosted by Edina. (See Pleiades+ discussion in the Gazetteers chapter for more on versions of Pleiades.)
See the Unlock Places [http://edina.ac.uk/unlock/places/] website for documentation and examples of use of the service.
The Unlock Text [http://edina.ac.uk/unlock/texts/] API is a more sophisticated tool and requires the user to create an account. It allows you to submit complete texts for geoparsing, either individually or in bundles. It is, in effect, an online version of the geoparser pipeline and is probably the simplest way of using the geoparser, if no customisation is required. There are subtle and unavoidable differences between the Unlock version and the downloadable geoparser package, meaning that results will not necessarily be identical. If your needs are complex it may be better to install a local copy of the geoparser.
See Getting Started with Unlock Text [http://edina.ac.uk/unlock/texts/getstarted.html] for how to use the online service. At the time of writing, a visualisation component is in development, so that users can submit full texts and see map-based results immediately.
At the time of writing the Unlock Text documentation doesn’t explicitly show how to use the Pleiades+ gazetteer hosted by Edina, but this can be specified much as for the Unlock Text example above, by including a gazetteer:gazname pair in the JSON data item for “src”, like this:
{"src":"http://synapse.inf.ed.ac.uk/~kate/gap/unlockTests/book9plain.txt",
"gazetteer":"plplus"}
For more on using Unlock Text with classical works, see blog posts on the Google Ancient Places (GAP) project at Unlocking Text [http://googleancientplaces.wordpress.com/2012/12/21/unlocking-text/] and Even More Unlocked [http://googleancientplaces.wordpress.com/2013/01/21/even-more-unlocked/]. The GapVis interface (version 1 [http://nrabinowitz.github.io/gapvis/] and version 2 [http://enridaga.github.io/gapvis/gap2/]) uses Unlock Text as part of its back-end engine, to produce an interface to Herodotus’ Histories and other texts, for students of classics.