A Routing Policy Specification Language parsing library forked from RIPE-NNC's whois project. The library contains the subset of classes required to parse and represent RPSL objects, along with added functionality to facilitate inspection of complex attributes.

$ mvn package

for deployment instructions, refer to the rpsl4j-generator

String rpslDocumentString = "...";
Set<RpslObject> objectSet = new HashSet<>():
RpslObjectStringReader rpslReader = new RpslObjectStringReader(rpslDocumentString);

for(String objStr : rpslReader)
    objectSet.add(RpslObject.parse(objStr));

You can also use RpslObjectFileReader or the more general RpslObjectStreamReader if they are a better fit for your project.

RpslObject routeExample = RpslObject.parse("route: 1.1.1.0/24\n
origin: AS1\n mnt-by: MNT-EXAMPLE # Need to change owner\n");

// The value of the class attribute is retrieved as follows
RpslAttribute routePrefixAttr = routeExample.getTypeAttribute();

// The value of a specific attribute can be retrieved using the AttributeType enumerator
RpslAttribute originASAttr = routeExample.findAttribute(AttributeType.ORIGIN);

// Many attribute types have corresponding classes that provide useful encapsulation and functionality.
// Refer to the net.ripe.db.whois.common.rpsl.attrs package
//
AutNum originAS = AutNum.parse(originAsAttr.getCleanValue());
AddressPrefixRange routePrefix = AddressPrefixRange.parse(routePrefixAttr.getCleanValue());

// If you want the full text of an attribute (including comments etc), use RpslAttribute#getValue() instead of RpslAttribute#getCleanValue()
System.out.println(routeExample.findAttribute(AttributeType.MNT_BY).getValue());

When rpsl4j-parser is compiled, parsers and lexers are generated from bison and jflex sources for many RPSL attributes. These can be found in the net.ripe.db.whois.common.generated package and are located in the target/generated-sources/jflex directory. These are primarily used for validity checks but, in combination with the AttributeLexerWrapper class, can be used to inspect supported attributes more deeply.

The simplest interface to this functionality is the RpslAttribute#getTokenList method. This method will either return a list of syntax tokens (describd below) or an empty list, so it's important to check which attributes lexer's exist for before calling it.

The method returns the Java type List<Pair<String, List<String>>>, which represents the type [(TOKEN_TYPE, [TOKEN_VALUES])]. An example of this process follows:

RpslAttribute importAttribute = RpslAttrbute.parse(AttributeType.IMPORT, "import: from AS2 accept AS1, 1.2.3.4/5");

// Generate the token list for the attribute
List<Pair<String, List<String>>> importTokenList = importAttribtue.getTokenList();

System.out.println(importTokenList); // prints "[(from, [AS2]), (accept, [AS1, 1.2.3.4/5])]""

// Iterate through the tokens
for(Pair<String, List<String>> tokenPair : exportAttribute.getTokenList()) {
    String tokenType = tokenPair.getLeft();
    List<String> values = tokenPair.getRight();

    if(tokenType.equals("from"))
        setOrigins(values);
    else if(tokenType.equals("accept"))
        setIncoming(values);
}

AttributeLexerWrapper works by generating a table of lexer states and their corresponding token names via reflection, then running the lexer over an attribute; recording tokens and runs of values. Unfortunately there is no easy way of predetermining the structure of a token list without thoroughly studying the attribute's corresponding lexer. The best approach at present is by experiment and testing.

refer to the rpsl4j-generator library for more complex uses of AttributeLexerWrapper.

Excluding code already licensed by RIPE under the BSD license, the source code is licensed under the GNU Affero General Public License.