casesensitivity, stemming added as configurable query param

Author: gaborremenyi

src/main/java/eu/fusepool/p3/dictionarymatcher/DictionaryAnnotator.java

@@ -12,6 +12,7 @@
 import opennlp.tools.tokenize.TokenizerME;
 import opennlp.tools.tokenize.TokenizerModel;
 import opennlp.tools.util.Span;
+import org.apache.commons.lang.StringUtils;
 import org.arabidopsis.ahocorasick.AhoCorasick;
 import org.arabidopsis.ahocorasick.SearchResult;
 import org.tartarus.snowball.SnowballStemmer;
@@ -73,29 +74,29 @@ public DictionaryAnnotator(DictionaryStore _dictionary, String _stemmingLanguage
         }
 
         // if no stemming language configuration is provided set stemming language to None
-        if (stemmingLanguage == null || stemmingLanguage.isEmpty()) {
-            stemmingLanguage = "None";
+        if (StringUtils.isBlank(stemmingLanguage)) {
+            stemmingLanguage = "none";
         }
         // create a mapping between the language and the name of the class
         // responsible for the stemming of the current language
         languages = new HashMap<>();
-        languages.put("None", "");
-        languages.put("Danish", "danishStemmer");
-        languages.put("Dutch", "dutchStemmer");
-        languages.put("English", "englishStemmer");
-        languages.put("Finnish", "finnishStemmer");
-        languages.put("French", "frenchStemmer");
-        languages.put("German", "germanStemmer");
-        languages.put("Hungarian", "hungarianStemmer");
-        languages.put("Italian", "italianStemmer");
-        languages.put("Norwegian", "norwegianStemmer");
-        //languages.put("english2", "porterStemmer");
-        languages.put("Portuguese", "portugueseStemmer");
-        languages.put("Romanian", "romanianStemmer");
-        languages.put("Russian", "russianStemmer");
-        languages.put("Spanish", "spanishStemmer");
-        languages.put("Swedish", "swedishStemmer");
-        languages.put("Turkish", "turkishStemmer");
+        languages.put("none", "");
+        languages.put("danish", "danishStemmer");
+        languages.put("dutch", "dutchStemmer");
+        languages.put("english", "englishStemmer");
+        languages.put("finnish", "finnishStemmer");
+        languages.put("french", "frenchStemmer");
+        languages.put("german", "germanStemmer");
+        languages.put("hungarian", "hungarianStemmer");
+        languages.put("italian", "italianStemmer");
+        languages.put("norwegian", "norwegianStemmer");
+        languages.put("english2", "porterStemmer");
+        languages.put("portuguese", "portugueseStemmer");
+        languages.put("romanian", "romanianStemmer");
+        languages.put("russian", "russianStemmer");
+        languages.put("spanish", "spanishStemmer");
+        languages.put("swedish", "swedishStemmer");
+        languages.put("turkish", "turkishStemmer");
 
         originalDictionary = new DictionaryStore();
         processedDictionary = new DictionaryStore();

src/main/java/eu/fusepool/p3/transformer/dictionarymatcher/Arguments.java

@@ -10,7 +10,7 @@ public interface Arguments extends ArgumentsWithHelp {
             description = "The port on which the proxy shall listen")
     public int getPort();
 
-    @CommandLine(longName = "enableCors", shortName = {"C"},
+    @CommandLine(longName = "enableCors", shortName = {"C"}, required = false,
             description = "Enable a liberal CORS policy",
             isSwitch = true)
     public boolean enableCors();

src/main/java/eu/fusepool/p3/transformer/dictionarymatcher/DictionaryMatcherTransformer.java

@@ -72,12 +72,19 @@ public DictionaryMatcherTransformer(String queryString) {
             throw new TransformerException(HttpServletResponse.SC_BAD_REQUEST, "ERROR: Badly formatted query string: \"" + queryString + "\" \nUsage: http://<transformer>/?taxonomy=<taxonomy_URI>");
         }
 
+        // get taxonomy URI
         String taxonomy = queryParams.get("taxonomy");
 
         if (StringUtils.isEmpty(taxonomy)) {
             throw new TransformerException(HttpServletResponse.SC_BAD_REQUEST, "ERROR: Taxonomy URI was not provided! \nUsage: http://<transformer>/?taxonomy=<taxonomy_URI>");
         }
 
+        // get case sensitivity
+        boolean caseSensitivity = queryParams.get("casesensitive") != null;
+
+        // get stemming language
+        String stemmingLanguage = queryParams.get("stemming");
+
         // create new dictionaryAnnotator if it does not exists
         if (dictionaryAnnotator == null) {
             long start, end;
@@ -90,7 +97,7 @@ public DictionaryMatcherTransformer(String queryString) {
             System.out.print(" (" + dictionary.GetSize() + ") and creating transformer ...");
 
             // create the dictionary annotator instance
-            dictionaryAnnotator = new DictionaryAnnotator(dictionary, "English", false, 0, false);      // TODO get settings from query string
+            dictionaryAnnotator = new DictionaryAnnotator(dictionary, stemmingLanguage, caseSensitivity, 0, false);
 
             end = System.currentTimeMillis();
             System.out.println(" done [" + Double.toString((double) (end - start) / 1000) + " sec] .");