swe-phrasestructure-sparv

Converts swedish phrase structures from Mamba-Dep dependency analysis using a rule-based heuristic. Below is the complete list of possible phrase structure labels:

NP: noun phrase
NP-wh: noun phrase with a relativizer, e.g. "whose mother"
PrP: prepositional phrase
PrP-wh: prepositional phrase with a relativizer, t.e. "in which"
SBAR: subordinate clause introduced by a subordinator
S: clause
S-wh: clause introduced by a relativizer
S-imp: clause in the imperative
VP-sup: verb phrase using the supine
VP-att: verb phrase with the infinitive, including the infinitive marker "att"
VP-inf: verb phrase with the infinitive, without an infinitive marker
VP-fin: finite verb phrase
ADJP: adjective phrase
ADVP: adverb phrase
ADVP-wh: adverb phrase with a relativizer
QP: numeral phrase

This analysis is used with Sparv. Check out Sparv's quick start guide to get started!

To use this analysis, add the following lines under export.annotations in the Sparv corpus configuration file:

- phrase_structure.phrase  # Phrase segments
- phrase_structure.phrase:phrase_structure.name  # Phrase names
- phrase_structure.phrase:phrase_structure.func  # Phrase functions

For more info on how to use Sparv, check out the Sparv documentation.

Example output:

<phrase func="ROOT" name="S">
  <phrase func="SS" name="NP">
    <token>Alfred</token>
    <token>Bernhard</token>
    <token>Nobel</token>
  </phrase>
  <token>var</token>
  <phrase func="SP" name="NP">
    <phrase func="DT" name="NP">
      <token>en</token>
      <phrase func="AT" name="ADJP">
        <token>svensk</token>
        <phrase func="CJ" name="NP">
          <token>kemist</token>
        </phrase>
      </phrase>
      <token>och</token>
    </phrase>
    <token>stiftare</token>
    <phrase func="ET" name="PrP">
      <token>av</token>
      <phrase func="PA" name="NP">
        <token>Nobelpriset</token>
        <token>.</token>
      </phrase>
    </phrase>
  </phrase>
</phrase>