Added PandasUtil#formatPandasCategorical(List<List<?>>) utility method

jpmml · Dec 14, 2023 · d1a244d · d1a244d
1 parent 1b96efe
commit d1a244d
Show file tree

Hide file tree

Showing 2 changed files with 40 additions and 8 deletions.
diff --git a/pmml-lightgbm/src/main/java/org/jpmml/lightgbm/PandasUtil.java b/pmml-lightgbm/src/main/java/org/jpmml/lightgbm/PandasUtil.java
@@ -43,13 +43,27 @@ public List<List<?>> parsePandasCategorical(String string){
 
 		JsonElement element = JsonParser.parseString(string);
 
-		Gson gson = new GsonBuilder()
-		    .setObjectToNumberStrategy(ToNumberPolicy.LONG_OR_DOUBLE)
-		    .create();
+		Gson gson = createGson();
 
 		return gson.fromJson(element, ListOfLists.class);
 	}
 
+	static
+	public String formatPandasCategorical(List<List<?>> objects){
+		Gson gson = createGson();
+
+		return PandasUtil.PREFIX_PANDAS_CATEGORICAL + gson.toJson(objects, ListOfLists.class);
+	}
+
+	static
+	private Gson createGson(){
+		Gson result = new GsonBuilder()
+			.setObjectToNumberStrategy(ToNumberPolicy.LONG_OR_DOUBLE)
+			.create();
+
+		return result;
+	}
+
 	static
 	private class ListOfLists extends ArrayList<List<?>> {
 	}

diff --git a/pmml-lightgbm/src/test/java/org/jpmml/lightgbm/PandasUtilTest.java b/pmml-lightgbm/src/test/java/org/jpmml/lightgbm/PandasUtilTest.java
@@ -29,18 +29,36 @@
 public class PandasUtilTest {
 
 	@Test
-	public void parse() throws Exception {
-		List<List<?>> pandasCategories = parsePandasCategorical("null");
+	public void parse(){
+		assertNull(parsePandasCategorical("null"));
 
-		assertNull(pandasCategories);
+		assertEquals(PandasUtilTest.CATEGORIES_LIST_OF_LISTS, parsePandasCategorical(PandasUtilTest.CATEGORIES_STRING));
+	}
 
-		pandasCategories = parsePandasCategorical("[[\"null\", \"A\", \"B, B\", \"C, [C], C\"], [-2, -1, 0, 1, 2], [-2.0, -1.0, 0.0, 1.0, 2.0], [false, true]]");
+	@Test
+	public void format(){
+		assertEquals("null", formatPandasCategorical(null));
 
-		assertEquals(Arrays.asList(Arrays.asList("null", "A", "B, B", "C, [C], C"), Arrays.asList(-2L, -1L, 0L, 1L, 2L), Arrays.asList(-2d, -1d, 0d, 1d, 2d), Arrays.asList(Boolean.FALSE, Boolean.TRUE)), pandasCategories);
+		assertEquals(removeWhitespace(PandasUtilTest.CATEGORIES_STRING), removeWhitespace(formatPandasCategorical(PandasUtilTest.CATEGORIES_LIST_OF_LISTS)));
 	}
 
 	static
 	private List<List<?>> parsePandasCategorical(String value){
 		return PandasUtil.parsePandasCategorical(PandasUtil.PREFIX_PANDAS_CATEGORICAL + value);
 	}
+
+	static
+	private String formatPandasCategorical(List<List<?>> objects){
+		String result = PandasUtil.formatPandasCategorical(objects);
+
+		return result.substring(PandasUtil.PREFIX_PANDAS_CATEGORICAL.length());
+	}
+
+	static
+	private String removeWhitespace(String string){
+		return string.replaceAll("\\s","");
+	}
+
+	private static final String CATEGORIES_STRING = "[[\"null\", \"A\", \"B, B\", \"C, [C], C\"], [-2, -1, 0, 1, 2], [-2.0, -1.0, 0.0, 1.0, 2.0], [false, true]]";
+	private static final List<List<?>> CATEGORIES_LIST_OF_LISTS = Arrays.asList(Arrays.asList("null", "A", "B, B", "C, [C], C"), Arrays.asList(-2L, -1L, 0L, 1L, 2L), Arrays.asList(-2d, -1d, 0d, 1d, 2d), Arrays.asList(Boolean.FALSE, Boolean.TRUE));
 }