The SMAF XML interface to PET is enabled via the -tok=smaf option to PET.

SMAF XML can be sent directly to the PET input stream (unfortunately, each complete input must correspond to a single line -- no newlines!). Eg.

<smaf><lattice init='v0' final='v2'><edge type='pos' id='p1' deps='t1'><slot name="tag">NP1</slot></edge><edge type='token' id='t2' cfrom='2' cto='4' source='v1' target='v2'>won</edge><edge type='token' id='t1' cfrom='0' cto='1' source='v0' target='v1'>FooCorp</edge></lattice></smaf>

Alternatively the SMAF XML may be provided as the contents of a file (here the complete input may be spread of multiple lines). Syntax: '@' + FILENAME. Eg.

@/tmp/sample.smaf

You must provide a configuration file. This provides an interpretation for the edge types used provided in the SMAF input.

Add the following to your PET config file:

;; SMAF conf
smaf-conf := "saf.conf".

A sample SMAF config file is given below:

define gMap.carg (synsem lkeys keyrel carg) STRING

token.[] -> edgeType='tok' tokenStr=content
wordForm.[] -> edgeType='morph' stem=content.stem partialTree=content.partial-tree
ersatz.[] -> edgeType='tok+morph' stem=content.name tokenStr=content.name gMap.carg=content.surface inject='t' analyseMorph='t'
pos.[] -> edgeType='morph' fallback='' pos=content.tag gMap.carg=deps.content

The sample SMAF given above makes use of the edge type token (where the token string is "simple content", that is the text content of the XML element) ; and of the edge type pos (linked to a token element via its deps attribute, with part-of-speech tag stored in the slot named tag, and with semantics obtained from the content of the associated token).

For interpretation of part-of-speech tags, see the PET config setting posmapping (enabled via command line option -default-les).