tests: add blacklist functionality for unsupported ELN files in tests

Author: Pfeil

src/main/java/edu/kit/datamanager/ro_crate/reader/ZipStreamStrategy.java

@@ -85,53 +85,49 @@ public boolean isExtracted() {
      * 
      * @param stream The input stream.
      */
-    private void readCrate(InputStream stream) {
-        try {
-            File folder = temporaryFolder.toFile();
-            // ensure the directory is clean
-            if (folder.exists()) {
-                if (folder.isDirectory()) {
-                    FileUtils.cleanDirectory(folder);
-                } else if (folder.isFile()) {
-                    FileUtils.delete(folder);
-                }
-            } else {
-                FileUtils.forceMkdir(folder);
+    private void readCrate(InputStream stream) throws IOException {
+        File folder = temporaryFolder.toFile();
+        // ensure the directory is clean
+        if (folder.exists()) {
+            if (folder.isDirectory()) {
+                FileUtils.cleanDirectory(folder);
+            } else if (folder.isFile()) {
+                FileUtils.delete(folder);
             }
+        } else {
+            FileUtils.forceMkdir(folder);
+        }
 
-            LocalFileHeader localFileHeader;
-            int readLen;
-            byte[] readBuffer = new byte[4096];
+        LocalFileHeader localFileHeader;
+        int readLen;
+        byte[] readBuffer = new byte[4096];
 
-            try (ZipInputStream zipInputStream = new ZipInputStream(stream)) {
-                while ((localFileHeader = zipInputStream.getNextEntry()) != null) {
-                    String fileName = localFileHeader.getFileName();
-                    File extractedFile = new File(folder, fileName).getCanonicalFile();
-                    if (!extractedFile.toPath().startsWith(folder.getCanonicalPath())) {
-                        throw new IOException("Entry is outside of target directory: " + fileName);
-                    }
-                    if (localFileHeader.isDirectory()) {
-                        FileUtils.forceMkdir(extractedFile);
-                        continue;
-                    }
-                    FileUtils.forceMkdir(extractedFile.getParentFile());
-                    try (OutputStream outputStream = new FileOutputStream(extractedFile)) {
-                        while ((readLen = zipInputStream.read(readBuffer)) != -1) {
-                            outputStream.write(readBuffer, 0, readLen);
-                        }
+        try (ZipInputStream zipInputStream = new ZipInputStream(stream)) {
+            while ((localFileHeader = zipInputStream.getNextEntry()) != null) {
+                String fileName = localFileHeader.getFileName();
+                File extractedFile = new File(folder, fileName).getCanonicalFile();
+                if (!extractedFile.toPath().startsWith(folder.getCanonicalPath())) {
+                    throw new IOException("Entry is outside of target directory: " + fileName);
+                }
+                if (localFileHeader.isDirectory()) {
+                    FileUtils.forceMkdir(extractedFile);
+                    continue;
+                }
+                FileUtils.forceMkdir(extractedFile.getParentFile());
+                try (OutputStream outputStream = new FileOutputStream(extractedFile)) {
+                    while ((readLen = zipInputStream.read(readBuffer)) != -1) {
+                        outputStream.write(readBuffer, 0, readLen);
                     }
                 }
             }
-            this.isExtracted = true;
-            // register deletion on exit
-            FileUtils.forceDeleteOnExit(folder);
-        } catch (IOException ex) {
-            logger.error("Failed to read crate from input stream.", ex);
         }
+        this.isExtracted = true;
+        // register deletion on exit
+        FileUtils.forceDeleteOnExit(folder);
     }
 
     @Override
-    public ObjectNode readMetadataJson(InputStream stream) {
+    public ObjectNode readMetadataJson(InputStream stream) throws IOException {
         if (!isExtracted) {
             this.readCrate(stream);
         }
@@ -152,17 +148,11 @@ public ObjectNode readMetadataJson(InputStream stream) {
                     .orElseThrow(() -> new IllegalStateException("No %s found in zip file".formatted(JsonDescriptor.ID)));
             jsonMetadata = firstSubdir.toPath().resolve(JsonDescriptor.ID).toFile();
         }
-
-        try {
-            return objectMapper.readTree(jsonMetadata).deepCopy();
-        } catch (IOException e) {
-            logger.error("Failed to deserialize crate metadata.", e);
-            return null;
-        }
+        return objectMapper.readTree(jsonMetadata).deepCopy();
     }
 
     @Override
-    public File readContent(InputStream stream) {
+    public File readContent(InputStream stream) throws IOException {
         if (!isExtracted) {
             this.readCrate(stream);
         }

src/test/java/edu/kit/datamanager/ro_crate/reader/ElnFileFormatTest.java

@@ -19,6 +19,20 @@ public interface ElnFileFormatTest<
         >
         extends TestableReaderStrategy<SOURCE_T, READER_STRATEGY>
 {
+    /**
+     * Some readers may not be able to read a subset of eln files,
+     * e.g. because a zip file may not be readable in streaming mode.
+     * <p>
+     * An implementation test may use this methode to provide a subset of the
+     * test cases where an IOException is expected.
+     *
+     * @param input the input to test for presence in the blacklist
+     * @return true if the input is in the blacklist, false otherwise
+     */
+    default boolean isInBlacklist(String input) {
+        return false;
+    }
+
     /**
      * ELN Crates are zip files not fully compatible with the Ro-Crate standard
      * in the sense that they must contain a single subfolder in the zip file
@@ -46,9 +60,14 @@ default void testReadElnCrates(String urlStr, @TempDir Path tmp) throws IOExcept
         FileUtils.copyURLToFile(url, elnFile.toFile(), 10000, 10000);
         assertTrue(elnFile.toFile().exists());
 
-        // Read the crate from the downloaded file
-        Crate read = this.readCrate(elnFile);
-        assertNotNull(read);
-        assertFalse(read.getAllDataEntities().isEmpty());
+        if (!isInBlacklist(urlStr)) {
+            // Read the crate from the downloaded file
+            Crate read = this.readCrate(elnFile);
+            assertNotNull(read);
+            assertFalse(read.getAllDataEntities().isEmpty());
+        } else {
+            // If the file is in the blacklist, we expect an IOException
+            assertThrows(IOException.class, () -> this.readCrate(elnFile));
+        }
     }
 }

src/test/java/edu/kit/datamanager/ro_crate/reader/ZipStreamReaderTest.java

@@ -5,6 +5,7 @@
 
 import java.io.*;
 import java.nio.file.Path;
+import java.util.Set;
 
 import static org.junit.jupiter.api.Assertions.*;
 
@@ -13,6 +14,19 @@ class ZipStreamReaderTest implements
         CommonReaderTest<InputStream, ZipStreamStrategy>,
         ElnFileFormatTest<InputStream, ZipStreamStrategy>
 {
+    /**
+     * At the point of writing this test,
+     *  these files are in a zip format which cannot be read in streaming mode
+     */
+    @Override
+    public boolean isInBlacklist(String input) {
+        return Set.of(
+                "https://github.com/TheELNConsortium/TheELNFileFormat/raw/refs/heads/master/examples/kadi4mat/records-example.eln",
+                "https://github.com/TheELNConsortium/TheELNFileFormat/raw/refs/heads/master/examples/kadi4mat/collections-example.eln"
+        )
+                .contains(input);
+    }
+
     @Override
     public void saveCrate(Crate crate, Path target) throws IOException {
         Writers.newZipStreamWriter().save(crate, new FileOutputStream(target.toFile()));