Correcting Search of Polish records

From Koha Test Wiki MediaWiki MySQL Binary

This page explains how to configure Zebra search engine to transliterate Polish special characters into a latin form.

This only works for a Koha installed with ICU option.

How to

Goto Zebra configuration directory : etc/zebradb.

Edit etc/words-icu.xml (and etc/phrases-icu.xml if exists).

After root node icu_chain, add :

  <transliterate rule="{ ą > a "/>
  <transliterate rule="{ Ą > a "/>
  <transliterate rule="{ ć > c "/>
  <transliterate rule="{ Ć > c "/>
  <transliterate rule="{ ę > e "/>
  <transliterate rule="{ Ę > e "/>
  <transliterate rule="{ ł > l "/>
  <transliterate rule="{ Ł > l "/>
  <transliterate rule="{ ń > n "/>
  <transliterate rule="{ Ń > n "/>
  <transliterate rule="{ ó > o "/>
  <transliterate rule="{ Ó > o "/>
  <transliterate rule="{ ś > s "/>
  <transliterate rule="{ Ś > s "/>
  <transliterate rule="{ ź > z "/>
  <transliterate rule="{ Ź > z "/>
  <transliterate rule="{ ż > z "/>
  <transliterate rule="{ Ż > z "/>

Reindex all database.

Result

You will be able to find records written in latin or polish, using latin or polish search terms.

Example with author "Miłosz Oskar Wladysław de Lubicz". Performing a search with "Milosz" or "Miłosz" will return records containing "Milosz" or "Miłosz".

See also Correcting_Search_of_Arabic_records and ICU_chains_configuration