nl.basjes.parse.core

Class Disector

java.lang.Object

nl.basjes.parse.core.Disector

public abstract class Disector
extends Object

A Disector is a class capable of chopping a String into multiple values of specific types.

The lifecycle:

Parser setup

1. First instance is constructed
2. First instance is added to the Parser using Parser.addDisector(Disector)
3. The Parser calls getInputType() and getPossibleOutput() to know what this disector can deliver.

The parser now constructs a tree based on the available Disectors and what was requested.

Disectors setup

For each node in the tree a new instance of the required disector is created by calling getNewInstance() which calls initializeNewInstance(Disector). Note that only Disectors that are acually needed will be in the parse tree.

For each of those instances in the tree:

1. For each of the actually needed input+output combinations prepareForDisect(String, String) is called. This can be used to avoid needless CPU cycles during the actual run.
2. As a final step a call to prepareForRun() is done as an indication that all preparation input has been provided. A Disector can use this to finalize the runtime data structures so doing the actual disecting faster.

Disecting

• During a run the instance will be called with disect(nl.basjes.parse.core.Parsable<?>, java.lang.String) many times.
• In the disect(Parsable, String) the actual value to be worked on must be retrieved using Parsable.getParsableField(String, String)
• The result(s) of the disection must be put back using nl.basjes.parse.core.Parsable#addDisection(String, String, String, String, java.util.EnumSet)

Constructor Summary

Constructors

Constructor and Description
Disector()

Method Summary

Methods

Modifier and Type	Method and Description
abstract void	disect(Parsable<?> parsable, String inputname) This method must disect the provided field from the parsable into 'smaller' pieces.
abstract String	getInputType()
Disector	getNewInstance() Create an additional instance of this disector.
abstract List<String>	getPossibleOutput() What are all possible outputs that can be provided.
protected abstract void	initializeNewInstance(Disector newInstance)
abstract EnumSet<Casts>	prepareForDisect(String inputname, String outputname) This tells the disector that it should prepare that we will call it soon with 'inputname' and expect to get 'inputname.outputname' because inputname is of the type returned by getInputType and outputname was part of the answer from getPossibleOutput.
abstract void	prepareForRun() The framework will tell the disector that it should get ready to run.

Methods inherited from class java.lang.Object

clone, equals, finalize, getClass, hashCode, notify, notifyAll, toString, wait, wait, wait

Constructor Detail

Disector

public Disector()

Method Detail

disect

public abstract void disect(Parsable<?> parsable,
          String inputname)
                     throws DisectionFailure

This method must disect the provided field from the parsable into 'smaller' pieces.

Throws:

DisectionFailure

getInputType

public abstract String getInputType()

Returns:

The required typename of the input

getPossibleOutput

public abstract List<String> getPossibleOutput()

What are all possible outputs that can be provided.

Returns:

array of "type:name" values that indicates all the possible outputs

prepareForDisect

public abstract EnumSet<Casts> prepareForDisect(String inputname,
                              String outputname)

This tells the disector that it should prepare that we will call it soon with 'inputname' and expect to get 'inputname.outputname' because inputname is of the type returned by getInputType and outputname was part of the answer from getPossibleOutput. This can be used by the disector implementation to optimize the internal parsing algorithms and lookup tables and such. The disector must return the types to which this value can be mapped later on during the run.

prepareForRun

public abstract void prepareForRun()
                            throws InvalidDisectorException

The framework will tell the disector that it should get ready to run. I.e. finalize the bootstrapping.

Throws:

InvalidDisectorException

getNewInstance

public Disector getNewInstance()

Create an additional instance of this disector. This is needed because in the parse tree we may need the same disector multiple times. In order to optimize per node we need separate instances.

Returns:

New instance of this Disector

initializeNewInstance

protected abstract void initializeNewInstance(Disector newInstance)