data-integrations · seTTBoi · Apr 16, 2025
diff --git a/wrangler-api/src/main/java/io/cdap/wrangler/api/parser/ByteSize.java b/wrangler-api/src/main/java/io/cdap/wrangler/api/parser/ByteSize.java
@@ -0,0 +1,100 @@
+package io.cdap.wrangler.api.parser;
+
+import com.google.gson.JsonElement;
+import com.google.gson.JsonObject;
+
+import java.util.regex.Matcher;
+import java.util.regex.Pattern;
+
+/**
+ * Represents a byte size token parsed from a string (e.g., "10KB", "1.5MB").
+ * The value is stored in canonical units (bytes).
+ */
+public class ByteSize implements Token {
+  // The original token string.
+  private final String token;
+  // Canonical value stored in bytes.
+  private final long bytes;
+  // Token type.
+  private final TokenType tokenType = TokenType.BYTE_SIZE;
+
+  // Conversion constants (using 1024-based conversions)
+  private static final long KILOBYTE = 1024;
+  private static final long MEGABYTE = KILOBYTE * 1024;
+  private static final long GIGABYTE = MEGABYTE * 1024;
+  private static final long TERABYTE = GIGABYTE * 1024;
+
+  // Pattern to capture the numeric and unit portions (e.g., "10KB" or "1.5MB").
+  private static final Pattern PATTERN = Pattern.compile("([0-9]+(?:\\.[0-9]+)?)([a-zA-Z]+)");
+
+  /**
+   * Constructs a ByteSize token by parsing the input string.
+   *
+   * @param token the string representation (e.g., "10KB", "1.5MB")
+   * @throws IllegalArgumentException if the token is null, empty, or its format is invalid.
+   */
+  public ByteSize(String token) {
+    if (token == null || token.trim().isEmpty()) {
+      throw new IllegalArgumentException("Token cannot be null or empty");
+    }
+    this.token = token;
+    Matcher matcher = PATTERN.matcher(token);
+    if (!matcher.matches()) {
+      throw new IllegalArgumentException("Invalid byte size token format: " + token);
+    }
+
+    String numberStr = matcher.group(1);
+    String unitStr = matcher.group(2).toUpperCase(); // Normalize unit to upper-case
+
+    double value = Double.parseDouble(numberStr);
+    long multiplier;
+
+    // Determine multiplier based on the unit.
+    if ("B".equals(unitStr)) {
+      multiplier = 1;
+    } else if ("KB".equals(unitStr)) {
+      multiplier = KILOBYTE;
+    } else if ("MB".equals(unitStr)) {
+      multiplier = MEGABYTE;
+    } else if ("GB".equals(unitStr)) {
+      multiplier = GIGABYTE;
+    } else if ("TB".equals(unitStr)) {
+      multiplier = TERABYTE;
+    } else {
+      throw new IllegalArgumentException("Unsupported byte size unit: " + unitStr);
+    }
+
+    // Calculate canonical value in bytes.
+    this.bytes = (long) Math.round(value * multiplier);
+  }
+
+  /**
+   * Returns the byte size in canonical units (bytes).
+   *
+   * @return the size in bytes.
+   */
+  public long getBytes() {
+    return bytes;
+  }
+
+  @Override
+  public Object value() {
+    // Return the canonical value.
+    return bytes;
+  }
+
+  @Override
+  public TokenType type() {
+    return tokenType;
+  }
+
+  @Override
+  public JsonElement toJson() {
+    // Construct a JSON representation of this token.
+    JsonObject obj = new JsonObject();
+    obj.addProperty("token", token);
+    obj.addProperty("type", tokenType.name());
+    obj.addProperty("bytes", bytes);
+    return obj;
+  }
+}
diff --git a/wrangler-api/src/main/java/io/cdap/wrangler/api/parser/TimeDuration.java b/wrangler-api/src/main/java/io/cdap/wrangler/api/parser/TimeDuration.java
@@ -0,0 +1,92 @@
+package io.cdap.wrangler.api.parser;
+
+import com.google.gson.JsonElement;
+import com.google.gson.JsonObject;
+
+import java.util.regex.Matcher;
+import java.util.regex.Pattern;
+
+/**
+ * Represents a time duration token parsed from a string (e.g., "150ms", "2.1s").
+ * The value is stored in canonical units (milliseconds).
+ */
+public class TimeDuration implements Token {
+  // The original token string.
+  private final String token;
+  // Canonical value stored in milliseconds.
+  private final long milliseconds;
+  // Token type.
+  private final TokenType tokenType = TokenType.TIME_DURATION;
+
+  // Constant: 1 second = 1000 milliseconds.
+  private static final long SECOND_IN_MS = 1000;
+
+  // Pattern to capture the numeric and unit portions (e.g., "150ms" or "2.1s").
+  private static final Pattern PATTERN = Pattern.compile("([0-9]+(?:\\.[0-9]+)?)([a-zA-Z]+)");
+
+  /**
+   * Constructs a TimeDuration token by parsing the input string.
+   *
+   * @param token the string representation (e.g., "150ms", "2.1s")
+   * @throws IllegalArgumentException if the token is null, empty, or its format is invalid.
+   */
+  public TimeDuration(String token) {
+    if (token == null || token.trim().isEmpty()) {
+      throw new IllegalArgumentException("Token cannot be null or empty");
+    }
+    this.token = token;
+    Matcher matcher = PATTERN.matcher(token);
+    if (!matcher.matches()) {
+      throw new IllegalArgumentException("Invalid time duration token format: " + token);
+    }
+
+    String numberStr = matcher.group(1);
+    // Normalize the unit to lower-case for easier comparison.
+    String unitStr = matcher.group(2).toLowerCase();
+
+    double value = Double.parseDouble(numberStr);
+    long multiplier;
+
+    // Determine multiplier based on the unit.
+    if ("ms".equals(unitStr)) {
+      multiplier = 1;
+    } else if ("s".equals(unitStr)) {
+      multiplier = SECOND_IN_MS;
+    } else {
+      throw new IllegalArgumentException("Unsupported time duration unit: " + unitStr);
+    }
+
+    // Calculate canonical value in milliseconds.
+    this.milliseconds = (long) Math.round(value * multiplier);
+  }
+
+  /**
+   * Returns the time duration in canonical units (milliseconds).
+   *
+   * @return the duration in milliseconds.
+   */
+  public long getMilliseconds() {
+    return milliseconds;
+  }
+
+  @Override
+  public Object value() {
+    // Return the canonical value.
+    return milliseconds;
+  }
+
+  @Override
+  public TokenType type() {
+    return tokenType;
+  }
+
+  @Override
+  public JsonElement toJson() {
+    // Construct a JSON representation of this token.
+    JsonObject obj = new JsonObject();
+    obj.addProperty("token", token);
+    obj.addProperty("type", tokenType.name());
+    obj.addProperty("milliseconds", milliseconds);
+    return obj;
+  }
+}
diff --git a/wrangler-api/src/main/java/io/cdap/wrangler/api/parser/TokenType.java b/wrangler-api/src/main/java/io/cdap/wrangler/api/parser/TokenType.java
@@ -152,5 +152,19 @@ public enum TokenType implements Serializable {
    * Represents the enumerated type for the object of type {@code String} with restrictions
    * on characters that can be present in a string.
    */
-  IDENTIFIER
+  IDENTIFIER,
+
+  /**
+   * Represents the enumerated type for a byte size token.
+   * This type is associated with tokens representing byte sizes, such as "10KB" or "1.5MB".
+   */
+  BYTE_SIZE,
+
+  /**
+   * Represents the enumerated type for a time duration token.
+   * This type is associated with tokens representing time durations, such as "150ms" or "2.1s".
+   */
+  TIME_DURATION
+
+
 }
diff --git a/wrangler-core/.vscode/settings.json b/wrangler-core/.vscode/settings.json
@@ -0,0 +1,3 @@
+{
+    "java.compile.nullAnalysis.mode": "automatic"
+}
diff --git a/wrangler-core/pom.xml b/wrangler-core/pom.xml
@@ -313,52 +313,58 @@
 
 
   <build>
-    <resources>
-      <resource>
-        <directory>src/main/resources</directory>
-        <filtering>true</filtering>
-      </resource>
-    </resources>
-    <plugins>
-      <plugin>
-        <groupId>org.antlr</groupId>
-        <artifactId>antlr4-maven-plugin</artifactId>
-        <version>${antlr4-maven-plugin.version}</version>
-        <configuration>
-          <visitor>true</visitor>
-        </configuration>
-        <executions>
-          <execution>
-            <goals>
-              <goal>antlr4</goal>
-            </goals>
-          </execution>
-        </executions>
-      </plugin>
-      <plugin>
-        <groupId>org.codehaus.mojo</groupId>
-        <artifactId>buildnumber-maven-plugin</artifactId>
-        <version>1.0</version>
-        <executions>
-          <execution>
-            <phase>generate-resources</phase>
-            <goals>
-              <goal>create</goal>
-            </goals>
-          </execution>
-        </executions>
-        <configuration>
-          <doCheck>true</doCheck>
-          <doUpdate>false</doUpdate>
-          <revisionOnScmFailure>false</revisionOnScmFailure>
-          <format>{0,date,yyyy-MM-dd-HH:mm:ss}_{1}</format>
-          <items>
-            <item>timestamp</item>
-            <item>${user.name}</item>
-          </items>
-        </configuration>
-      </plugin>
-    </plugins>
-  </build>
+  <sourceDirectory>src/main/java</sourceDirectory>
+  <testSourceDirectory>src/test/java</testSourceDirectory>
+
+  <resources>
+    <resource>
+      <directory>src/main/resources</directory>
+      <filtering>true</filtering>
+    </resource>
+  </resources>
+
+  <plugins>
+    <plugin>
+      <groupId>org.antlr</groupId>
+      <artifactId>antlr4-maven-plugin</artifactId>
+      <version>${antlr4-maven-plugin.version}</version>
+      <configuration>
+        <visitor>true</visitor>
+      </configuration>
+      <executions>
+        <execution>
+          <goals>
+            <goal>antlr4</goal>
+          </goals>
+        </execution>
+      </executions>
+    </plugin>
+
+    <plugin>
+      <groupId>org.codehaus.mojo</groupId>
+      <artifactId>buildnumber-maven-plugin</artifactId>
+      <version>1.0</version>
+      <executions>
+        <execution>
+          <phase>generate-resources</phase>
+          <goals>
+            <goal>create</goal>
+          </goals>
+        </execution>
+      </executions>
+      <configuration>
+        <doCheck>true</doCheck>
+        <doUpdate>false</doUpdate>
+        <revisionOnScmFailure>false</revisionOnScmFailure>
+        <format>{0,date,yyyy-MM-dd-HH:mm:ss}_{1}</format>
+        <items>
+          <item>timestamp</item>
+          <item>${user.name}</item>
+        </items>
+      </configuration>
+    </plugin>
+  </plugins>
+</build>
+
 
 </project>