Support input charset and output charset. #5

Author: uw4

src/main/java/com/dashjoin/jsonata/cli/Main.java

@@ -9,6 +9,7 @@
 import java.io.PrintStream;
 import java.io.PrintWriter;
 import java.io.StringWriter;
+import java.nio.charset.Charset;
 import java.nio.file.Files;
 import java.util.Map;
 
@@ -44,16 +45,17 @@ void initCli() {
         options = new Options();
         options.addOption("v", "version", false, "Display version info");
         options.addOption("h", "help", false, "Display help and version info");
-        // TODO options.addOption("j", "jsonargs", true, "Specify arguments as JSON object");
         options.addOption(Option.builder("e").longOpt("expression")
                          .argName("file")
                          .hasArg()
                          .desc("JSONata expression file to evaluate")
                          //.required()
                          .build());
         options.addOption("i", "input", true, "JSON input file (- for stdin)");
+        options.addOption("ic", "icharset", true, "Input character set (default=UTF-8)");
         options.addOption("f", "format", true, "Input format (default=auto)");
         options.addOption("o", "output", true, "JSON output file (default=stdout)");
+        options.addOption("oc", "ocharset", true, "Output character set (default=UTF-8)");
         options.addOption("time", false, "Print performance timers to stderr");
         options.addOption("c", "compact", false, "Compact JSON output (don't prettify)");
         options.addOption(Option.builder("b").longOpt("bindings").
@@ -108,6 +110,9 @@ void run(String[] args) throws Throwable {
             bindingsStr = cmd.getOptionValue("b");
         Map<String, Object> bindingsObj = bindingsStr != null ? (Map<String, Object>)Json.parseJson(bindingsStr) : null;
 
+        String icharset = cmd.getOptionValue("ic", "UTF-8");
+        String ocharset = cmd.getOptionValue("oc", "UTF-8");
+
         InputStream in = null;
         Object input = null;
         {
@@ -132,15 +137,15 @@ void run(String[] args) throws Throwable {
             out = new FileOutputStream(arg);
         }
 
-        PrintStream pout = new PrintStream(out);
+        PrintStream pout = new PrintStream(out, false, ocharset);
 
         long t0 = System.currentTimeMillis();
 
         String formatStr = cmd.getOptionValue("f", TerminalUtil.InputFormat.auto.name());
         TerminalUtil.InputFormat format = TerminalUtil.InputFormat.valueOf(formatStr);
 
         if (in!=null)
-            input = TerminalUtil.readInput(in, format);
+            input = TerminalUtil.readInput(in, format, Charset.forName(icharset));
 
         long t1 = System.currentTimeMillis();
 

src/main/java/com/dashjoin/jsonata/cli/TerminalUtil.java

@@ -5,6 +5,8 @@
 import java.io.InputStream;
 import java.io.InputStreamReader;
 import java.io.UnsupportedEncodingException;
+import java.nio.charset.Charset;
+
 import com.dashjoin.jsonata.json.Json;
 
 /**
@@ -37,11 +39,12 @@ public static enum InputFormat {
      * 
      * @param in
      * @param format
+     * @param cs
      * @return Input object
      * @throws IOException
      * @throws UnsupportedEncodingException
      */
-    public static Object readInput(InputStream in, InputFormat format)
+    public static Object readInput(InputStream in, InputFormat format, Charset cs)
             throws IOException, UnsupportedEncodingException {
         Object input = null;
         switch (format) {
@@ -51,19 +54,19 @@ public static Object readInput(InputStream in, InputFormat format)
                 try {
                     // First try JSON, if there is an error,
                     // reset the stream and use string format
-                    return readInput(in, InputFormat.json);
+                    return readInput(in, InputFormat.json, cs);
                 } catch (Exception ex) {
                     if (in.markSupported()) {
                         in.reset();
-                        return readInput(in, InputFormat.string);
+                        return readInput(in, InputFormat.string, cs);
                     }
                     // We cannot reset the stream: throw the exception
                     // Need to specify the format manually in this case
                     throw ex;
                 }
             }
             case json: {
-                input = Json.parseJson(new InputStreamReader(in));
+                input = Json.parseJson(new InputStreamReader(in, cs));
                 break;
             }
             case string: {
@@ -72,7 +75,7 @@ public static Object readInput(InputStream in, InputFormat format)
                 for (int length; (length = in.read(buffer)) != -1;) {
                     result.write(buffer, 0, length);
                 }
-                input = result.toString("UTF-8");
+                input = result.toString(cs);
                 break;
             }
             default: