Merge pull request #32 from richardrodgers/alg-copy-fix

Cleanup of Adapter bugs - prepare for release

README.md

@@ -118,11 +118,20 @@ If automatic generation is not desired, an API call disables it.
 Another extra is _sealed_ bags. Bags created by Loaders are immutable, meaning they cannot be altered via the API.
 But we typically _can_ gain access to the backing bag storage, which we can of course then
 change at will. However, if a bag is created as _sealed_ (a method on the Loader), all
-method calls that expose the underlying storage will throw IllegalAccess exceptions. So, for example,
-we would be _unable_ to obtain a File reference, but _could_ get an I/O stream to the same content.
+method calls that expose the underlying storage will throw IllegalAccess exceptions. So, for example, we would
+be _unable_ to obtain a File reference to a payload file, but _could_ get an I/O stream to the same content.
 In other words, the content can be accessed, but the underlying representation cannot be altered, and
 to this degree the bag contents are _tamper-proof_.
 
+Finally, the library bundles several convenience (helper) methods for common operations on bags, using
+the _Adapter_ class. One illustrative use case would be to upgrade an existing bag to a better checksum algorithm.
+One could code:
+
+    Bag oldbag = Loader.load(oldBagZip);
+    Path newbagZip = Adapter.copy(newbag, oldbag, "SHA-512").toPackage();
+
+Additional methods will be added as other common use-cases are identified.
+
 ## Bagger on the command line ##
 
 The library bundles a very simple command-line tool called _Bagger_ that exposes much of the API.

src/main/java/edu/mit/lib/bagit/Adapter.java

@@ -10,6 +10,7 @@
 import java.net.URISyntaxException;
 import java.nio.file.Files;
 import java.nio.file.Path;
+import java.util.Arrays;
 import java.util.HashMap;
 import java.util.HashSet;
 import java.util.Map;
@@ -38,10 +39,10 @@ public class Adapter {
      */
     public static Filler copy(Bag bag) throws IOException, IllegalAccessException, URISyntaxException {
         var csAlgs = bag.csAlgorithms();
-        var algArray = new String[csAlgs.size()];
-        csAlgs.toArray(algArray);
-        var filler = new Filler(Files.createTempDirectory("bag"), bag.tagEncoding(), bag.lineSeparator(), true, algArray);
-        copyBag(filler, bag, csAlgs, algArray[0]);
+        var filler = new Filler(Files.createTempDirectory("bag"),
+                                bag.tagEncoding(), bag.lineSeparator(), true,
+                                csAlgs.toArray(String[]::new));
+        copyBag(filler, bag, csAlgs);
         return filler;
     }
 
@@ -60,10 +61,9 @@ public static Filler copy(Bag bag) throws IOException, IllegalAccessException, U
      */
     public static Filler copy(Path base, Bag bag) throws IOException, IllegalAccessException, URISyntaxException {
         var csAlgs = bag.csAlgorithms();
-        var algArray = new String[csAlgs.size()];
-        csAlgs.toArray(algArray);
-        var filler = new Filler(base, bag.tagEncoding(), bag.lineSeparator(), false, algArray);
-        copyBag(filler, bag, csAlgs, algArray[0]);
+        var filler = new Filler(base, bag.tagEncoding(), bag.lineSeparator(), false,
+                                csAlgs.toArray(String[]::new));
+        copyBag(filler, bag, csAlgs);
         return filler;
     }
 
@@ -82,14 +82,21 @@ public static Filler copy(Path base, Bag bag) throws IOException, IllegalAccessE
      * @throws URISyntaxException when invalid URIs encountered
      */
     public static Filler copy(Path base, Bag bag, String ... csAlgorithms) throws IOException, IllegalAccessException, URISyntaxException {
+        // ensure we can handle refs - bail if not
+        var algSet = new HashSet<String>(Arrays.asList(csAlgorithms));
+        if (! bag.payloadRefs().isEmpty() && ! bag.csAlgorithms().containsAll(algSet)) {
+            throw new IllegalStateException("Fetch file checksums absent for requested algorithms");
+        }
         var filler = new Filler(base, bag.tagEncoding(), bag.lineSeparator(), false, csAlgorithms);
-        copyBag(filler, bag, Set.of(csAlgorithms), csAlgorithms[0]);
+        copyBag(filler, bag, Set.of(csAlgorithms));
         return filler;
     }
 
-    private static void copyBag(Filler filler, Bag bag, Set<String> csAlgs, String csAlg) 
+    private static void copyBag(Filler filler, Bag bag, Set<String> newAlgs) 
         throws IOException, IllegalAccessException, URISyntaxException  {
         var manifs = new HashMap<String, Map<String, String>>();
+        var csAlgs = bag.csAlgorithms();
+        var csAlg = csAlgs.iterator().next();
         var refs = bag.payloadRefs();
         var autoGs = autoGens(csAlgs);
         var autoMs = autoMetas();
@@ -113,10 +120,10 @@ private static void copyBag(Filler filler, Bag bag, Set<String> csAlgs, String c
             }
         }
         for (String relPath : refs.keySet()) {
-            var refParts = refs.get(relPath).split(" ");
+            var refParts = refs.get(relPath).split("\\s+");
             var size = refParts[0].equals("-") ? -1L : Long.parseLong(refParts[0]);
             var checksums = new HashMap<String, String>();
-            for (String alg : csAlgs) {
+            for (String alg : newAlgs) {
                 checksums.put(alg, manifs.get(alg).get(relPath));
             }
             filler.payloadRefUnsafe(relPath, size, new URI(refParts[1]), checksums);
@@ -129,7 +136,6 @@ private static void copyBag(Filler filler, Bag bag, Set<String> csAlgs, String c
                 }
             }
         }
-
     }
 
     private static Set<String> autoGens(Set<String> csAlgs) {

src/main/java/edu/mit/lib/bagit/Bag.java

@@ -52,7 +52,7 @@ public class Bag {
     static final String LIB_VSN = "1.0";
     static final String DFLT_FMT = "zip";
     static final String TGZIP_FMT = "tgz";
-    static final String SPACER = "  ";
+    static final String SPACER = " ";
     // required file and directory names
     static final String MANIF_FILE = "manifest-";
     static final String TAGMANIF_FILE = "tagmanifest-";

src/main/java/edu/mit/lib/bagit/Filler.java

@@ -402,7 +402,7 @@ public Filler payloadRef(String relPath, InputStream in, URI uri) throws IOExcep
         FlatWriter refWriter = getWriter(REF_FILE);
         var destDir = Files.createTempDirectory("null");
         var destFile = destDir.resolve("foo");
-        long size = digestCopy(in, destFile, relPath, manWriters);
+        long size = digestCopy(in, destFile, DATA_PATH + relPath, manWriters);
         var sizeStr = (size > 0L) ? Long.toString(size) : "-";
         refWriter.writeLine(uri.toString() + " " + sizeStr + " " + DATA_PATH + relPath);
         Files.delete(destFile);
@@ -428,7 +428,7 @@ public Filler payloadRefUnsafe(String relPath, long size, URI uri, Map<String, S
             throw new IllegalStateException("Payload file already exists at: " + relPath);
         }
         if (! uri.isAbsolute()) {
-            throw new IOException("URI must be absolute");
+            throw new IOException("URI must be absolute: '" + uri.toString() + "'");
         }
         if (! checksums.keySet().equals(csAlgs)) {
             throw new IOException("checksums do not match bags");