README.md

In baasman/allegRo: This package provides an interface for interacting with an Allegro Graph server

allegRo

An API for the semantic graph database, AllegroGraph.

To use this package, you'll need to install your own AllegroGraph server. To install a free version, follow this link and follow the installation instructions.

Installation

After you've set up a working version of AllegroGraph, download the package as follows. As of right now it is not available on CRAN.

To install the latest version of this package

if (packageVersion("devtools") < 1.6) {
      install.packages("devtools")
    }
devtools::install_github("baasman/allegRo")

An introduction to the basic function supported by allegRo

We'll start off with creating a connection to the server, and specifying our credentials. For example, my test server is located on "localhost:10059", so that is where I'll be pointing it too. By specifying the testConnection = TRUE we can perform a simple get command to ensure we a receiving a response from the server.

url = "http://localhost:10059/"
user = "baasman"
password = "mypassword"

service = service(url,user,password,testConnection = TRUE)

If you receive the message saying you can safely move on, we can proceed with the actual work.

Lets take a look at the different catalogs and repositories on this server.

listCatalogs(service)

#create a new catalog object, I'll be working in my root catalog
cat = catalog(service,"root")

listRepositories(cat)

Lets create a new repository and link to it.

repositoryid = "testRepo"
expectedSize = 100

#to get a full list of arguments, look at the help documentation
createRepository(cat, repo = repositoryid, expectedSize = expectedSize)

#create repository object to work with
rep = repository(cat,"testRepo")

Suppose that want to add some triples to this store. There are multiple different ways to approach this, but for now, the easiest methods are either adding straight from a flat text file, or using a loop and dynamically changing the subject, predicate, or object.

subject = "<person>"
predicate = "<hasItem>"
objects = c('"sword"','"shield"')

for(i in 1:2){
  addStatement(rep,subj = subject,pred = predicate,
               obj = objects[i],context = "<testing>")
}

To add them from a text file, we can use the following function:

addStatementsFromFile(rep,
                      filepath = paste0(path.package("allegRo"),"/inst/extdata/testtrips.nq"),
                      context = "<testing>")

Finally, we can evaluate sparql queries on this test store. I already forgot what items our person has!

returnType = "data.table" # can also be dataframe, matrix, list or plain vector
query = "select ?items {<http://test.com/tmp#person> <http://test.com/tmp#hasItem> ?items }"
cleanUp = FALSE #will remove all XMLSchema, convert appropriate types to R types.

evalQuery(rep,returnType = returnType,
          query = query,limit = 10,cleanUp = cleanUp)

Another efficient way to return these triples is to match part of the triple pattern. Since all added triples are under the "" context, we can return all statements in which the context == ""

getStatements(rep,context = "<testing>", returnType = returnType, limit = 10, cleanUp = cleanUp)

#the following would work as well

getStatements(rep,pred = "<hasItem>", returnType = returnType, limit = 10, cleanUp = cleanUp)

#you can also supply multiple arguments when dealing with statements

getStatements(rep,subj = list("'sword'","'shield'"), returnType = returnType, limit = 10, cleanUp = cleanUp)

As of right now, you can return triples by SPARQL queries, statement matching, and text indices. Prolog is not yet supported.

baasman/allegRo documentation built on May 11, 2019, 5:22 p.m.