Merge pull request #16551 from cavelo-anoop/tika_parser_config

Use setter method to derive the type of property
quarkusio · Apr 15, 2021 · 1321e3f · 1321e3f
2 parents 5e749f2 + 66635ca
commit 1321e3f
Show file tree

Hide file tree

Showing 2 changed files with 45 additions and 9 deletions.
diff --git a/extensions/tika/deployment/src/main/java/io/quarkus/tika/deployment/TikaProcessor.java b/extensions/tika/deployment/src/main/java/io/quarkus/tika/deployment/TikaProcessor.java
@@ -1,5 +1,6 @@
 package io.quarkus.tika.deployment;
 
+import java.lang.reflect.Method;
 import java.util.ArrayList;
 import java.util.Arrays;
 import java.util.Collections;
@@ -222,18 +223,24 @@ private static Class<?> loadParserClass(String parserName) {
     private static String getParserParamType(String parserName, String paramName) {
         try {
             Class<?> parserClass = loadParserClass(parserName);
-            String paramType = parserClass.getMethod("get" + capitalize(paramName), new Class[] {}).getReturnType()
-                    .getSimpleName().toLowerCase();
-            if (paramType.equals(boolean.class.getSimpleName())) {
-                // TikaConfig Param class does not recognize 'boolean', only 'bool'
-                // This whole reflection code is temporary anyway
-                paramType = "bool";
+            Method[] methods = parserClass.getMethods();
+            String setterMethodName = "set" + capitalize(paramName);
+            String paramType = null;
+            for (Method method : methods) {
+                if (method.getName().equals(setterMethodName) && method.getParameterCount() == 1) {
+                    paramType = method.getParameterTypes()[0].getSimpleName().toLowerCase();
+                    if (paramType.equals(boolean.class.getSimpleName())) {
+                        // TikaConfig Param class does not recognize 'boolean', only 'bool'
+                        // This whole reflection code is temporary anyway
+                        paramType = "bool";
+                    }
+                    return paramType;
+                }
             }
-            return paramType;
         } catch (Throwable t) {
-            final String errorMessage = "Parser " + parserName + " has no " + paramName + " property";
-            throw new TikaParseException(errorMessage);
+            throw new TikaParseException(String.format("Parser %s has no %s property", parserName, paramName));
         }
+        throw new TikaParseException(String.format("Parser %s has no %s property", parserName, paramName));
     }
 
     public static class TikaParserParameter {

diff --git a/extensions/tika/deployment/src/test/java/io/quarkus/tika/deployment/TikaProcessorTest.java b/extensions/tika/deployment/src/test/java/io/quarkus/tika/deployment/TikaProcessorTest.java
@@ -66,6 +66,35 @@ public void testPdfParserConfig() throws Exception {
         assertEquals("true", parserConfig.get(pdfParserFullName).get(0).getValue());
     }
 
+    @Test
+    public void testTesseractParserConfig() throws Exception {
+        String ocrParserFullName = "org.apache.tika.parser.ocr.TesseractOCRParser";
+        Map<String, List<TikaProcessor.TikaParserParameter>> parserConfig = getParserConfig(null, "ocr",
+                Collections.singletonMap("ocr",
+                        Collections.singletonMap("tesseract-path", "/opt/tesseract/")),
+                Collections.singletonMap("ocr", ocrParserFullName));
+        assertEquals(1, parserConfig.size());
+
+        assertEquals(1, parserConfig.get(ocrParserFullName).size());
+        assertEquals("tesseractPath", parserConfig.get(ocrParserFullName).get(0).getName());
+        assertEquals("/opt/tesseract/", parserConfig.get(ocrParserFullName).get(0).getValue());
+    }
+
+    @Test
+    public void testUnknownParserConfig() throws Exception {
+        String ocrParserFullName = "org.apache.tika.parser.ocr.TesseractOCRParser";
+        try {
+            Map<String, List<TikaProcessor.TikaParserParameter>> parserConfig = getParserConfig(null, "ocr",
+                    Collections.singletonMap("ocr",
+                            Collections.singletonMap("tesseract-unknown-opt", "/opt/tesseract/")),
+                    Collections.singletonMap("ocr", ocrParserFullName));
+        } catch (Exception e) {
+            // expected
+            assertEquals("Parser org.apache.tika.parser.ocr.TesseractOCRParser has no tesseractUnknownOpt property",
+                    e.getMessage());
+        }
+    }
+
     @Test
     public void testUnresolvableCustomAbbreviation() throws Exception {
         try {