FeforPetApi

PET API proposal

The idea is to replace the cheap executable (with 'pipe' over stderr...) by a library. The library in turn then could be wrapped in an executable (if someone really still need this), or better in a XML-RPC or other socket server. The library could also be called directly from Java (via JNI) or scripting languages like Python.

The main motivation is to have a more flexible configuration. Currently, the cheap binary has to be restarted (including time-consuming grammar loading) each time the number of readings to return is to be changed, for example. Using the new API, this option could be modified for each parse individually.

Generally, no new features will be added, just the interface should be made explicit. For example, it will not be possible to exchange the grammar during runtime, as this would require fundamental changes in the code.

Elements of the API

• basic configuration like grammar
• MRS globals
• access to type hierarchy (sub/supertype, name/code, but no feature structure unification)

Feature requests

• configurable start symbol for parsing (FrancisBond)

• pseudo TDL syntax for settings (common with LKB), as discussed in JerezTop. A pseudo TDL parser could simply call the setting methods of the API then

• change lexicon during runtime (not possible for the built-in lexicon, but probably for lexDB)

• change model during runtime