fix dups task

brown-uk · Oct 5, 2023 · c342c54 · c342c54
1 parent 0ed7a29
commit c342c54
Show file tree

Hide file tree

Showing 2 changed files with 44 additions and 39 deletions.
diff --git a/build.gradle b/build.gradle
@@ -135,45 +135,11 @@ task sortDict(type: JavaExec) {
 }
 
 
-task dups {
-  doLast {
-	def dictDir = file("data/dict")
-
-	def files = dictDir.listFiles().findAll { it.name.endsWith('.lst') \
-	    && ! (it.name in ['dot-abbr.lst', 'geo-ukr-hydro.lst', 'geo-ukr-koatuu.lst']) }
-
-	def lines = files.collect { File f ->
-            boolean geo = f.name.contains('geo')
-            f.readLines()
-            .findAll { ! it.startsWith('#') && ! it.startsWith(' +cs=') } 
-            .collect { it
-//            .replaceFirst(/^ \+cs=/, '')
-                .trim()
-                .replaceAll(/\s*#.*/, '')
-                .replaceAll(/( \/[a-z0-9]+)(\s+|$)/, '$1... ')
-                .replaceAll(/\.[^<.: ]+/, '...')
-                .replaceAll(/\s*:coll/, '')
-                .replaceAll(/:&adjp:(actv|pasv)(:(imperf|perf))+/, '') 
-            }
-            .collect { line ->
-                geo ? "$line :geo" : line 
-            }
-        }
-        .flatten()
-	    .grep { it }
-
-//    lines = lines.grep { ! it.startsWith('#') }
-
-    def dups = lines.countBy{it}.grep{it.value > 1 }.collect{it.key}
-
-    java.text.Collator coll = java.text.Collator.getInstance(new Locale("uk", "UA"));
-    coll.setStrength(java.text.Collator.IDENTICAL)
-    coll.setDecomposition(java.text.Collator.NO_DECOMPOSITION)
-
+task checkDups(type: JavaExec) {
+    classpath = sourceSets.test.runtimeClasspath
+    mainClass = "org.dict_uk.check.CheckDups"
 
-    println dups.toSorted(coll).join("\n")
-    println "Found ${dups.size} dups"
-  }
+    workingDir = projectDir
 }
 
 task checkAnim {
@@ -344,7 +310,7 @@ task testExpand(type: JavaExec, dependsOn: [reverseAffix, autogen, classes]) {
 // Main targets
 //
 
-task expandForCorp(type: JavaExec, dependsOn: [reverseAffix, autogen/*, testExpand*/, dups]) {
+task expandForCorp(type: JavaExec, dependsOn: [reverseAffix, autogen/*, testExpand*/, checkDups]) {
     classpath = sourceSets.main.runtimeClasspath
     mainClass = "org.dict_uk.expand.ExpandAll"
 

diff --git a/src/test/groovy/org/dict_uk/check/CheckDups.groovy b/src/test/groovy/org/dict_uk/check/CheckDups.groovy
@@ -0,0 +1,39 @@
+#!/bin/env groovy
+
+package org.dict_uk.check
+
+def dictDir = new File("data/dict")
+
+def files = dictDir.listFiles().findAll { it.name.endsWith('.lst') \
+    && ! (it.name in ['dot-abbr.lst', 'geo-ukr-hydro.lst', 'geo-ukr-koatuu.lst']) }
+
+def lines = files.collect { File f ->
+            boolean geo = f.name.contains('geo')
+            f.readLines()
+            .findAll { ! it.startsWith('#') && ! it.startsWith(' +cs=') } 
+            .collect { it
+//            .replaceFirst(/^ \+cs=/, '')
+                .trim()
+                .replaceAll(/\s*#.*/, '')
+                .replaceAll(/( \/[a-z0-9]+)(\s+|$)/, '$1... ')
+                .replaceAll(/\.[^<.: ]+/, '...')
+                .replaceAll(/\s*:coll/, '')
+                .replaceAll(/:&adjp:(actv|pasv)(:(imperf|perf))+/, '') 
+            }
+            .collect { line ->
+                geo ? "$line :geo" : line 
+            }
+        }
+        .flatten()
+        .grep { it }
+
+//    lines = lines.grep { ! it.startsWith('#') }
+
+def dups = lines.countBy{it}.grep{it.value > 1 }.collect{it.key.toString()}
+
+java.text.Collator coll = java.text.Collator.getInstance(new Locale("uk", "UA"));
+coll.setStrength(java.text.Collator.IDENTICAL)
+coll.setDecomposition(java.text.Collator.NO_DECOMPOSITION)
+
+println dups.toSorted(coll).join("\n")
+println "Found ${dups.size()} dups"