WarcDigest: Add raw() method to access the raw uncanonicalized string…

… value

By putting this on WarcDigest itself it means we don't have to double up all the digest accessors.

#74

src/org/netpreserve/jwarc/WarcDigest.java

@@ -13,21 +13,20 @@
 import java.util.Objects;
 
 public class WarcDigest {
-
-    private final String algorithm;
-    private final String value;
+    private final String raw;
+    private String algorithm;
+    private String value;
 
     public WarcDigest(String digest) {
-        int i = digest.indexOf(':');
-        if (i == -1) {
-            throw new IllegalArgumentException("Invalid WARC-Digest");
-        }
-        this.algorithm = digest.substring(0, i);
-        this.value = base32Encode(digest.substring(i + 1), this.algorithm);
+        Objects.requireNonNull(digest);
+        raw = digest;
     }
 
     public WarcDigest(String algorithm, String value) {
-        this.algorithm = algorithm;
+        Objects.requireNonNull(algorithm);
+        Objects.requireNonNull(value);
+        raw = algorithm + ":" + value;
+        this.algorithm = canonicalizeAlgorithm(algorithm);
         this.value = base32Encode(value, algorithm);
     }
 
@@ -36,14 +35,35 @@ public WarcDigest(String algorithm, byte[] value) {
     }
 
     public WarcDigest(MessageDigest messageDigest) {
-        algorithm = messageDigest.getAlgorithm().replace("-", "").toLowerCase(Locale.US);
+        algorithm = canonicalizeAlgorithm(messageDigest.getAlgorithm());
         value = base32Encode(messageDigest.digest());
+        raw = algorithm + ":" + value;
+    }
+
+    private static String canonicalizeAlgorithm(String algorithm) {
+        return algorithm.replace("-", "").toLowerCase(Locale.US);
+    }
+
+    private void parse() {
+        if (value != null) return;
+        int i = raw.indexOf(':');
+        if (i == -1) {
+            throw new IllegalArgumentException("Invalid WARC-Digest");
+        }
+        this.algorithm = canonicalizeAlgorithm(raw.substring(0, i));
+        this.value = base32Encode(raw.substring(i + 1), this.algorithm);
     }
 
     public String algorithm() {
+        parse();
         return algorithm;
     }
 
+    private String value() {
+        parse();
+        return value;
+    }
+
     public String hex() {
         return hexEncode(bytes());
     }
@@ -53,22 +73,29 @@ public String base16() {
     }
 
     public String base32() {
-        return value;
+        return value();
     }
 
     public String base64() {
         return base64Encode(bytes());
     }
 
-    public byte[] bytes() { return base32Decode(value); }
+    public byte[] bytes() { return base32Decode(value()); }
+
+    /**
+     * Returns the original digest string without any canonicalization.
+     */
+    public String raw() {
+        return raw;
+    }
 
     @Override
     public String toString() {
         return prefixedBase32();
     }
 
     public String prefixedBase32() {
-        return algorithm + ":" + value;
+        return algorithm() + ":" + value();
     }
 
     static String hexEncode(byte[] data) {
@@ -197,13 +224,13 @@ public boolean equals(Object o) {
         if (this == o) return true;
         if (o == null || getClass() != o.getClass()) return false;
         WarcDigest digest = (WarcDigest) o;
-        return Objects.equals(algorithm, digest.algorithm) &&
-                Objects.equals(value, digest.value);
+        return Objects.equals(algorithm(), digest.algorithm()) &&
+                Objects.equals(value(), digest.value());
     }
 
     @Override
     public int hashCode() {
-        return Objects.hash(algorithm, value);
+        return Objects.hash(algorithm(), value());
     }
 
     /**
@@ -223,6 +250,6 @@ public static MessageDigest getDigester(String algorithm) throws NoSuchAlgorithm
     }
 
     public MessageDigest getDigester() throws NoSuchAlgorithmException {
-        return getDigester(algorithm);
+        return getDigester(algorithm());
     }
 }

src/org/netpreserve/jwarc/WarcTargetRecord.java

@@ -54,14 +54,6 @@ public URI targetURI() {
     public Optional<WarcDigest> payloadDigest() {
         return headers().sole("WARC-Payload-Digest").map(WarcDigest::new);
     }
-
-    /**
-    * Return the payload digest value directly
-    */
-   public Optional<String> payloadDigestUnchanged() {
-       return headers().sole("WARC-Payload-Digest");
-   }
-
 
     /**
      * A content-type that was identified by an independent check (not just what the server said).

src/org/netpreserve/jwarc/cdx/CdxFields.java

@@ -41,12 +41,9 @@ public static String format(byte field, WarcCaptureRecord record, boolean digest
         try {
             switch (field) {
                 case CHECKSUM:
-                    if (digestUnchanged) {
-                        return record.payloadDigestUnchanged().get();        
-                    }
-                    else {
-                      return record.payloadDigest().map(WarcDigest::base32).orElse("-");
-                    }
+                    return record.payloadDigest()
+                            .map(digestUnchanged ? WarcDigest::raw : WarcDigest::base32)
+                            .orElse("-");
                 case DATE:
                     return DATE_FORMAT.format(record.date());
                 case MIME_TYPE:

test/org/netpreserve/jwarc/WarcDigestTest.java

@@ -11,13 +11,20 @@
 import java.security.NoSuchAlgorithmException;
 import java.util.Arrays;
 
-import static org.junit.Assert.assertEquals;
-import static org.junit.Assert.assertTrue;
+import static org.junit.Assert.*;
 
 public class WarcDigestTest {
 
     private byte[] contentBytes = "hello world".getBytes();
 
+    @Test
+    public void testParsing() {
+        WarcDigest digest = new WarcDigest("Sha1:FKXGYNOJJ7H3IFO35FPUBC445EPOQRXN");
+        assertEquals("Sha1:FKXGYNOJJ7H3IFO35FPUBC445EPOQRXN", digest.raw());
+        assertEquals("sha1", digest.algorithm());
+        assertEquals("sha1:FKXGYNOJJ7H3IFO35FPUBC445EPOQRXN", digest.prefixedBase32());
+    }
+
     @Test
     public void testSha1() throws NoSuchAlgorithmException {
         MessageDigest md = MessageDigest.getInstance("SHA-1");