Laboratoire ICAR » Plateforme TXM

Add corpus parameters to manage how the concordance lines displays the CQP tokens and structural units

Solution 1: tokens and structures¶

• use the corpus-language parameter/property to apply the right txt-renderer of that language when building the content of the contexts of the lines
  • a txt-renderer is a component producing raw text (txt) by using language specific typographic rules to display the graphical forms of tokens
  • the Pager already uses a txt-renderer component to build edition pages content
• add a new concordance-display-opening-structural-unit-separator corpus parameter
  • possible values: '<', '(', '|'
• add a new concordance-display-opening-structural-units parameter corpus parameter
  • possible values: div,sp,p
• add a new concordance-display-closing-structural-unit-separator corpus parameter
  • possible values: '>', ')', '|'
• add a new concordance-display-closing-structural-units parameter corpus parameter
  • possible values: div,sp,p

Solution 2: tokens versus structures¶

For some studies on infra-word units (like syllables), a double corpus strategy is often used:

• corpus A builds CQP tokens as text words and the infra-word units are encoded through concatenation inside the word form (eg 'gi-ra-fe')
• corpus B builds CQP tokens as infra-word units and text words are encoded as <word> structures (eg <word form="girafe" seg-form="gi-ra-fe">gi ra fe</word>

To display correctly word forms in text editions and concordance lines:

• add a new word-as-structure corpus parameter
  • possible values: 'word' or <empty>
  • if the parameter value is <empty> use corpus tokens to render text words
  • if the parameter value is the name of a structure use that structure to render text words accordingly

When building text edition and concordance line contexts, use that parameter value to render correctly word forms to ease reading.

Remark: there already exist some XSLT stylesheets to build correct text edition for those two different cases.