#429

Committed 19 May 2025 12:15PM UTC coverage: 90.5% (+0.3%) from 90.169%

Build # #429

Build Type

Pull #258

github

Committed by

web-flow

Commit Message

Merge e4d91d01c into 810d1995c

Pull Request Pull Request #258: Support .ELN-style crates in all zip readers and writers

Run Details

224 of 245 new or added lines in 15 files covered. (91.43%)

3 existing lines in 3 files now uncovered.

2010 of 2221 relevant lines covered (90.5%)

0.9 hits per line

Source File
Press 'n' to go to next uncovered line, 'b' for previous

92.98

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

package edu.kit.datamanager.ro_crate.reader;

import com.fasterxml.jackson.databind.ObjectMapper;
import com.fasterxml.jackson.databind.node.ObjectNode;
import edu.kit.datamanager.ro_crate.entities.contextual.JsonDescriptor;
import edu.kit.datamanager.ro_crate.objectmapper.MyObjectMapper;
import java.io.File;
import java.io.FileOutputStream;
import java.io.IOException;
import java.io.InputStream;
import java.io.OutputStream;
import java.nio.file.Path;
import java.util.UUID;
import net.lingala.zip4j.io.inputstream.ZipInputStream;
import net.lingala.zip4j.model.LocalFileHeader;
import org.apache.commons.io.FileUtils;
import org.apache.commons.io.filefilter.FileFilterUtils;
import org.slf4j.Logger;
import org.slf4j.LoggerFactory;

/**
 * A ZIP file reader implementation of the StreamReaderStrategy interface.
 * This class handles reading and extraction of RO-Crate content from ZIP archives
 * into a temporary directory structure, which allows for accessing the contained files.
 *
 * @author jejkal
 */
public class ZipStreamStrategy implements GenericReaderStrategy<InputStream> {

    private static final Logger logger = LoggerFactory.getLogger(ZipStreamStrategy.class);
    protected final String ID = UUID.randomUUID().toString();
    protected Path temporaryFolder = Path.of(String.format("./.tmp/ro-crate-java/zipStreamReader/%s/", ID));
    protected boolean isExtracted = false;

    /**
     * Crates a ZipStreamReader with the default configuration as described in
     * the class documentation.
     */
    public ZipStreamStrategy() {
    }

    /**
     * Creates a ZipStreamReader which will extract the contents temporary to
     * the given location instead of the default location.
     *
     * @param folderPath the custom directory to extract content to for
     * temporary access.
     * @param shallAddUuidSubfolder if true, the reader will extract into
     * subdirectories of the given directory. These subdirectories will have
     * UUIDs as their names.
     */
    public ZipStreamStrategy(Path folderPath, boolean shallAddUuidSubfolder) {
        if (shallAddUuidSubfolder) {
            this.temporaryFolder = folderPath.resolve(ID);
        } else {
            this.temporaryFolder = folderPath;
        }
    }

    /**
     * @return the identifier which may be used as the name for a subfolder in
     * the temporary directory.
     */
    public String getID() {
        return ID;
    }

    /**
     * @return the folder (considered temporary) where the zipped crate will be
     * or has been extracted to.
     */
    public Path getTemporaryFolder() {
        return temporaryFolder;
    }

    /**
     * @return whether the crate has already been extracted into the temporary
     * folder.
     */
    public boolean isExtracted() {
        return isExtracted;
    }

    /**Read the crate metadata and content from the provided input stream.
     * 
     * @param stream The input stream.
     */
    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];

        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);
    }

    @Override
    public ObjectNode readMetadataJson(InputStream stream) throws IOException {
        if (!isExtracted) {
            this.readCrate(stream);
        }

        ObjectMapper objectMapper = MyObjectMapper.getMapper();
        File jsonMetadata = temporaryFolder.resolve(JsonDescriptor.ID).toFile();
        if (!jsonMetadata.isFile()) {
            // Try to find the metadata file in subdirectories
            File firstSubdir = FileUtils.listFilesAndDirs(
                            temporaryFolder.toFile(),
                            FileFilterUtils.directoryFileFilter(),
                            null
                    )
                    .stream()
                    .limit(50)
                    .filter(file -> file.toPath().toAbsolutePath().resolve(JsonDescriptor.ID).toFile().isFile())
                    .findFirst()
                    .orElseThrow(() -> new IllegalStateException("No %s found in zip file".formatted(JsonDescriptor.ID)));
            jsonMetadata = firstSubdir.toPath().resolve(JsonDescriptor.ID).toFile();
        }
        return objectMapper.readTree(jsonMetadata).deepCopy();
    }

    @Override
    public File readContent(InputStream stream) throws IOException {
        if (!isExtracted) {
            this.readCrate(stream);
        }
        return temporaryFolder.toFile();
    }
}

1	package edu.kit.datamanager.ro_crate.reader;
2
3	import com.fasterxml.jackson.databind.ObjectMapper;
4	import com.fasterxml.jackson.databind.node.ObjectNode;
5	import edu.kit.datamanager.ro_crate.entities.contextual.JsonDescriptor;
6	import edu.kit.datamanager.ro_crate.objectmapper.MyObjectMapper;
7	import java.io.File;
8	import java.io.FileOutputStream;
9	import java.io.IOException;
10	import java.io.InputStream;
11	import java.io.OutputStream;
12	import java.nio.file.Path;
13	import java.util.UUID;
14	import net.lingala.zip4j.io.inputstream.ZipInputStream;
15	import net.lingala.zip4j.model.LocalFileHeader;
16	import org.apache.commons.io.FileUtils;
17	import org.apache.commons.io.filefilter.FileFilterUtils;
18	import org.slf4j.Logger;
19	import org.slf4j.LoggerFactory;
20
21	/**
22	* A ZIP file reader implementation of the StreamReaderStrategy interface.
23	* This class handles reading and extraction of RO-Crate content from ZIP archives
24	* into a temporary directory structure, which allows for accessing the contained files.
25	*
26	* @author jejkal
27	*/
28	public class ZipStreamStrategy implements GenericReaderStrategy<InputStream> {
29
30	private static final Logger logger = LoggerFactory.getLogger(ZipStreamStrategy.class);	1✔
31	protected final String ID = UUID.randomUUID().toString();	1✔
32	protected Path temporaryFolder = Path.of(String.format("./.tmp/ro-crate-java/zipStreamReader/%s/", ID));	1✔
33	protected boolean isExtracted = false;	1✔
34
35	/**
36	* Crates a ZipStreamReader with the default configuration as described in
37	* the class documentation.
38	*/
39	public ZipStreamStrategy() {	1✔
40	}	1✔
41
42	/**
43	* Creates a ZipStreamReader which will extract the contents temporary to
44	* the given location instead of the default location.
45	*
46	* @param folderPath the custom directory to extract content to for
47	* temporary access.
48	* @param shallAddUuidSubfolder if true, the reader will extract into
49	* subdirectories of the given directory. These subdirectories will have
50	* UUIDs as their names.
51	*/
52	public ZipStreamStrategy(Path folderPath, boolean shallAddUuidSubfolder) {	1✔
53	if (shallAddUuidSubfolder) {	1✔
54	this.temporaryFolder = folderPath.resolve(ID);	1✔
55	} else {
56	this.temporaryFolder = folderPath;	1✔
57	}
58	}	1✔
59
60	/**
61	* @return the identifier which may be used as the name for a subfolder in
62	* the temporary directory.
63	*/
64	public String getID() {
65	return ID;	1✔
66	}
67
68	/**
69	* @return the folder (considered temporary) where the zipped crate will be
70	* or has been extracted to.
71	*/
72	public Path getTemporaryFolder() {
73	return temporaryFolder;	1✔
74	}
75
76	/**
77	* @return whether the crate has already been extracted into the temporary
78	* folder.
79	*/
80	public boolean isExtracted() {
81	return isExtracted;	1✔
82	}
83
84	/**Read the crate metadata and content from the provided input stream.
85	*
86	* @param stream The input stream.
87	*/
88	private void readCrate(InputStream stream) throws IOException {
89	File folder = temporaryFolder.toFile();	1✔
90	// ensure the directory is clean
91	if (folder.exists()) {	1✔
92	if (folder.isDirectory()) {	1✔
93	FileUtils.cleanDirectory(folder);	1✔
NEW 94	} else if (folder.isFile()) {	×
NEW 95	FileUtils.delete(folder);	×
96	}
97	} else {
98	FileUtils.forceMkdir(folder);	1✔
99	}
100
101	LocalFileHeader localFileHeader;
102	int readLen;
103	byte[] readBuffer = new byte[4096];	1✔
104
105	try (ZipInputStream zipInputStream = new ZipInputStream(stream)) {	1✔
106	while ((localFileHeader = zipInputStream.getNextEntry()) != null) {	1✔
107	String fileName = localFileHeader.getFileName();	1✔
108	File extractedFile = new File(folder, fileName).getCanonicalFile();	1✔
109	if (!extractedFile.toPath().startsWith(folder.getCanonicalPath())) {	1✔
NEW 110	throw new IOException("Entry is outside of target directory: " + fileName);	×
111	}
112	if (localFileHeader.isDirectory()) {	1✔
113	FileUtils.forceMkdir(extractedFile);	1✔
114	continue;	1✔
115	}
116	FileUtils.forceMkdir(extractedFile.getParentFile());	1✔
117	try (OutputStream outputStream = new FileOutputStream(extractedFile)) {	1✔
118	while ((readLen = zipInputStream.read(readBuffer)) != -1) {	1✔
119	outputStream.write(readBuffer, 0, readLen);	1✔
120	}
121	}
122	}	1✔
123	}
124	this.isExtracted = true;	1✔
125	// register deletion on exit
126	FileUtils.forceDeleteOnExit(folder);	1✔
127	}	1✔
128
129	@Override
130	public ObjectNode readMetadataJson(InputStream stream) throws IOException {
131	if (!isExtracted) {	1✔
132	this.readCrate(stream);	1✔
133	}
134
135	ObjectMapper objectMapper = MyObjectMapper.getMapper();	1✔
136	File jsonMetadata = temporaryFolder.resolve(JsonDescriptor.ID).toFile();	1✔
137	if (!jsonMetadata.isFile()) {	1✔
138	// Try to find the metadata file in subdirectories
139	File firstSubdir = FileUtils.listFilesAndDirs(	1✔
140	temporaryFolder.toFile(),	1✔
141	FileFilterUtils.directoryFileFilter(),	1✔
142	null
143	)
144	.stream()	1✔
145	.limit(50)	1✔
146	.filter(file -> file.toPath().toAbsolutePath().resolve(JsonDescriptor.ID).toFile().isFile())	1✔
147	.findFirst()	1✔
148	.orElseThrow(() -> new IllegalStateException("No %s found in zip file".formatted(JsonDescriptor.ID)));	1✔
149	jsonMetadata = firstSubdir.toPath().resolve(JsonDescriptor.ID).toFile();	1✔
150	}
151	return objectMapper.readTree(jsonMetadata).deepCopy();	1✔
152	}
153
154	@Override
155	public File readContent(InputStream stream) throws IOException {
156	if (!isExtracted) {	1✔
UNCOV 157	this.readCrate(stream);	×
158	}
159	return temporaryFolder.toFile();	1✔
160	}
161	}

kit-data-manager / ro-crate-java / #429

Source File Press 'n' to go to next uncovered line, 'b' for previous

Source File
Press 'n' to go to next uncovered line, 'b' for previous