Skip to content

Latest commit

 

History

History
43 lines (35 loc) · 1.67 KB

cqp_dialect.md

File metadata and controls

43 lines (35 loc) · 1.67 KB
Raw

CQP Dialect used for CQP2SPARQL

The CQP dialect for this project is mostly based on the SketchEngine CQP dialect, which is described here: https://www.sketchengine.eu/documentation/corpus-querying/

Key differences from various standards

  1. Logical operators can be both words and symbols:
  • Conjunction: and, &
  • Disjunction: or, |
  • Negation is currently only !
  1. Tokens can have symbolic labels, not only numerical
  2. Global constraints can be separated by any AND variant, or ::
  3. Comments are allowed after the global constraints after the symbol #
  4. Quantifying regular expressions can be applied to bracketed groups of tokens: ([] [])*

Names and conventions

  1. Attributes are represented as :<attr_name>, where namespaces are predefined in config files of a specific tool: conll, rdfs, powla, etc.
  2. Segments names are represented as :<type_name>, where namespaces are predefined in config files, and type_name corresponds to a specific type of which should be the enclosing segments.

Examples

  1. Sentences that contain exactly two verbs: one with a lemma է (to be), and another one which is either a perfective converb or an imperfective converb:
(    v1:[conll:LEM='է' and rdfs:type='olia:Verb'] 
     v2:[a='eanc:ImperfectiveConverb' or rdfs:type='eanc:PerfectiveConverb']) 
     within 
         (<powla:sentence/> !containing v3:[rdfs:type='olia:Verb'])
) & v1.ID != v3.ID and v2.ID != v3.ID
  1. Sentences that contain more than one recursive clauses with that.
(
	[conll:WORD='that']
	[]+
){2,}
within <powla:sentence/>