inital version of ICU chain XML config test file
authorMarc Cromme <marc@indexdata.dk>
Mon, 14 May 2007 10:07:48 +0000 (10:07 +0000)
committerMarc Cromme <marc@indexdata.dk>
Mon, 14 May 2007 10:07:48 +0000 (10:07 +0000)
src/test_icu_I18N.xml [new file with mode: 0644]

diff --git a/src/test_icu_I18N.xml b/src/test_icu_I18N.xml
new file mode 100644 (file)
index 0000000..17b593a
--- /dev/null
@@ -0,0 +1,29 @@
+<icu>
+  <icu_chain id="en:field" locale="en">
+    <display/>
+  </icu_chain>
+  <icu_chain id="en:sentence" locale="en">
+    <normalize rule="[:Control:] Any-Remove"/>
+    <tokenize rule="s"/>
+    <display/>
+    <normalize rule="[:Punctuation:] Remove"/>
+    <casemap rule="l"/>
+    <normal/>
+    <sort/>
+  </icu_chain>
+  <icu_chain id="en:word" locale="en">
+    <normalize rule="[:Control:] Any-Remove"/>
+    <tokenize rule="w"/>
+    <normalize rule="[[:WhiteSpace:][:Punctuation:]] Remove"/>
+    <display/>
+    <normalize rule="Lower"/>
+  </icu_chain>
+  <icu_chain id="en:number" locale="en">
+    <normalize rule="[:Letter:] Remove"/>
+    <tokenize rule="w"/>
+    <normalize rule="[:^Number:] Remove"/>
+    <display/>
+    <normal/>
+    <sort/>
+  </icu_chain>
+</icu>
\ No newline at end of file