diff options
Diffstat (limited to 'pdf-as-lib/src/main/java/at/gv/egiz/pdfas/framework/input')
7 files changed, 413 insertions, 0 deletions
diff --git a/pdf-as-lib/src/main/java/at/gv/egiz/pdfas/framework/input/DataSource.java b/pdf-as-lib/src/main/java/at/gv/egiz/pdfas/framework/input/DataSource.java new file mode 100644 index 0000000..f834622 --- /dev/null +++ b/pdf-as-lib/src/main/java/at/gv/egiz/pdfas/framework/input/DataSource.java @@ -0,0 +1,70 @@ +/**
+ * <copyright> Copyright 2006 by Know-Center, Graz, Austria </copyright>
+ * PDF-AS has been contracted by the E-Government Innovation Center EGIZ, a
+ * joint initiative of the Federal Chancellery Austria and Graz University of
+ * Technology.
+ *
+ * Licensed under the EUPL, Version 1.1 or - as soon they will be approved by
+ * the European Commission - subsequent versions of the EUPL (the "Licence");
+ * You may not use this work except in compliance with the Licence.
+ * You may obtain a copy of the Licence at:
+ * http://www.osor.eu/eupl/
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the Licence is distributed on an "AS IS" basis,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the Licence for the specific language governing permissions and
+ * limitations under the Licence.
+ *
+ * This product combines work with different licenses. See the "NOTICE" text
+ * file for details on the various modules and licenses.
+ * The "NOTICE" text file is part of the distribution. Any derivative works
+ * that you distribute must include a readable copy of the "NOTICE" text file.
+ */
+package at.gv.egiz.pdfas.framework.input;
+
+import java.io.InputStream;
+
+/**
+ * The input document data source.
+ *
+ * <p>
+ * Usually this is a PdfDataSource, but it may be a TextDataSource as well.
+ * </p>
+ *
+ * @author wprinz
+ *
+ */
+public interface DataSource
+{
+ /**
+ * Creates a new InputStream that allows to read out the document's binary
+ * data from the beginning.
+ *
+ * @return Returns the InputStream with the binary data.
+ */
+ public InputStream createInputStream();
+
+ /**
+ * Returns the length (number of bytes) of the stream.
+ *
+ * @return Returns the length (number of bytes) of the stream.
+ */
+ public int getLength();
+
+ /**
+ * Returns the data of this DataSource as a byte array.
+ *
+ * <p>
+ * Calling this method indicates that you need a byte array for random <strong>read</strong> access.
+ * The DataSource implementation should of course cache this byte array to avoid too much memory usage.
+ * </p>
+ * <p>
+ * Performance analysis has shown that the libraries internally convert the streams to byte arrays and
+ * that file system access is very slow.
+ * </p>
+ *
+ * @return
+ */
+ public byte [] getAsByteArray();
+}
diff --git a/pdf-as-lib/src/main/java/at/gv/egiz/pdfas/framework/input/ExtractionStage.java b/pdf-as-lib/src/main/java/at/gv/egiz/pdfas/framework/input/ExtractionStage.java new file mode 100644 index 0000000..19065b0 --- /dev/null +++ b/pdf-as-lib/src/main/java/at/gv/egiz/pdfas/framework/input/ExtractionStage.java @@ -0,0 +1,90 @@ +/**
+ * <copyright> Copyright 2006 by Know-Center, Graz, Austria </copyright>
+ * PDF-AS has been contracted by the E-Government Innovation Center EGIZ, a
+ * joint initiative of the Federal Chancellery Austria and Graz University of
+ * Technology.
+ *
+ * Licensed under the EUPL, Version 1.1 or - as soon they will be approved by
+ * the European Commission - subsequent versions of the EUPL (the "Licence");
+ * You may not use this work except in compliance with the Licence.
+ * You may obtain a copy of the Licence at:
+ * http://www.osor.eu/eupl/
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the Licence is distributed on an "AS IS" basis,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the Licence for the specific language governing permissions and
+ * limitations under the Licence.
+ *
+ * This product combines work with different licenses. See the "NOTICE" text
+ * file for details on the various modules and licenses.
+ * The "NOTICE" text file is part of the distribution. Any derivative works
+ * that you distribute must include a readable copy of the "NOTICE" text file.
+ */
+package at.gv.egiz.pdfas.framework.input;
+
+import java.util.List;
+
+import org.apache.commons.logging.Log;
+import org.apache.commons.logging.LogFactory;
+
+import at.gv.egiz.pdfas.exceptions.ErrorCode;
+import at.gv.egiz.pdfas.exceptions.framework.VerificationFilterException;
+import at.gv.egiz.pdfas.framework.DataSourceHolder;
+import at.gv.egiz.pdfas.framework.vfilter.VerificationFilter;
+import at.gv.egiz.pdfas.framework.vfilter.VerificationFilterParameters;
+import at.gv.egiz.pdfas.impl.input.IncrementalUpdateParser;
+import at.gv.egiz.pdfas.impl.vfilter.VerificationFilterImpl;
+import at.knowcenter.wag.egov.egiz.exceptions.PDFDocumentException;
+import at.knowcenter.wag.egov.egiz.exceptions.PresentableException;
+
+/**
+ * Extracts all signatures from a given input DataSource.
+ *
+ * @author wprinz
+ */
+public class ExtractionStage
+{
+ /**
+ * The log.
+ */
+ private static final Log log = LogFactory.getLog(ExtractionStage.class);
+
+ public List extractSignatureHolders(final DataSourceHolder dataSource, VerificationFilterParameters parameters) throws PresentableException
+ {
+ if (dataSource.getDataSource() instanceof PdfDataSource)
+ {
+ PdfDataSource pdfDataSource = (PdfDataSource) dataSource.getDataSource();
+
+ PdfDataSourceHolder pdsh = new PdfDataSourceHolder(pdfDataSource);
+ List blocks = parsePdfIntoBlocks(pdsh);
+ dataSource.setDataSource(pdsh.getDataSource());
+ parameters.setBeenCorrected(pdsh.hasChanged());
+
+ VerificationFilter vf = new VerificationFilterImpl();
+ List signatures = vf.extractSignatureHolders(pdsh.getDataSource(), blocks, parameters);
+
+ return signatures;
+ }
+
+ if (dataSource.getDataSource() instanceof TextDataSource)
+ {
+ TextDataSource textDataSource = (TextDataSource) dataSource.getDataSource();
+
+ VerificationFilter vf = new VerificationFilterImpl();
+ List signatures = vf.extractSignaturHolders(textDataSource, parameters);
+
+ return signatures;
+ }
+
+ String msg = "The input DataSource is neither pdf nor text. class.name = " + dataSource.getClass().getName();
+ log.error(msg);
+ throw new VerificationFilterException(ErrorCode.DOCUMENT_CANNOT_BE_READ, msg);
+ }
+
+ protected List parsePdfIntoBlocks(PdfDataSourceHolder pdfDataSource) throws PDFDocumentException
+ {
+ List blocks = IncrementalUpdateParser.parsePdfIntoIUBlocks(pdfDataSource);
+ return blocks;
+ }
+}
diff --git a/pdf-as-lib/src/main/java/at/gv/egiz/pdfas/framework/input/PdfDataSource.java b/pdf-as-lib/src/main/java/at/gv/egiz/pdfas/framework/input/PdfDataSource.java new file mode 100644 index 0000000..ce3d658 --- /dev/null +++ b/pdf-as-lib/src/main/java/at/gv/egiz/pdfas/framework/input/PdfDataSource.java @@ -0,0 +1,41 @@ +/**
+ * <copyright> Copyright 2006 by Know-Center, Graz, Austria </copyright>
+ * PDF-AS has been contracted by the E-Government Innovation Center EGIZ, a
+ * joint initiative of the Federal Chancellery Austria and Graz University of
+ * Technology.
+ *
+ * Licensed under the EUPL, Version 1.1 or - as soon they will be approved by
+ * the European Commission - subsequent versions of the EUPL (the "Licence");
+ * You may not use this work except in compliance with the Licence.
+ * You may obtain a copy of the Licence at:
+ * http://www.osor.eu/eupl/
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the Licence is distributed on an "AS IS" basis,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the Licence for the specific language governing permissions and
+ * limitations under the Licence.
+ *
+ * This product combines work with different licenses. See the "NOTICE" text
+ * file for details on the various modules and licenses.
+ * The "NOTICE" text file is part of the distribution. Any derivative works
+ * that you distribute must include a readable copy of the "NOTICE" text file.
+ */
+package at.gv.egiz.pdfas.framework.input;
+
+
+/**
+ * Represents the binary data of a PDF document.
+ *
+ * <p>
+ * This interface allows Pdf data to be handled in an abstract way so that the
+ * storage (byta array, disk etc.) of pdf documents can be separated from the
+ * algorithms.
+ * </p>
+ *
+ * @author wprinz
+ */
+public interface PdfDataSource extends DataSource
+{
+ // jsut a marker interface
+}
diff --git a/pdf-as-lib/src/main/java/at/gv/egiz/pdfas/framework/input/PdfDataSourceHolder.java b/pdf-as-lib/src/main/java/at/gv/egiz/pdfas/framework/input/PdfDataSourceHolder.java new file mode 100644 index 0000000..14a4ff4 --- /dev/null +++ b/pdf-as-lib/src/main/java/at/gv/egiz/pdfas/framework/input/PdfDataSourceHolder.java @@ -0,0 +1,49 @@ +/**
+ * <copyright> Copyright 2006 by Know-Center, Graz, Austria </copyright>
+ * PDF-AS has been contracted by the E-Government Innovation Center EGIZ, a
+ * joint initiative of the Federal Chancellery Austria and Graz University of
+ * Technology.
+ *
+ * Licensed under the EUPL, Version 1.1 or - as soon they will be approved by
+ * the European Commission - subsequent versions of the EUPL (the "Licence");
+ * You may not use this work except in compliance with the Licence.
+ * You may obtain a copy of the Licence at:
+ * http://www.osor.eu/eupl/
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the Licence is distributed on an "AS IS" basis,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the Licence for the specific language governing permissions and
+ * limitations under the Licence.
+ *
+ * This product combines work with different licenses. See the "NOTICE" text
+ * file for details on the various modules and licenses.
+ * The "NOTICE" text file is part of the distribution. Any derivative works
+ * that you distribute must include a readable copy of the "NOTICE" text file.
+ */
+package at.gv.egiz.pdfas.framework.input;
+
+public class PdfDataSourceHolder {
+ private PdfDataSource dataSource;
+ private boolean hasChanged = false;
+
+ public PdfDataSourceHolder(PdfDataSource dataSource) {
+ this.dataSource = dataSource;
+ }
+
+ public PdfDataSource getDataSource() {
+ return dataSource;
+ }
+
+ public void setDataSource(PdfDataSource dataSource) {
+ this.dataSource = dataSource;
+ this.hasChanged = true;
+ }
+
+ public boolean hasChanged() {
+ return this.hasChanged;
+ }
+
+
+
+}
diff --git a/pdf-as-lib/src/main/java/at/gv/egiz/pdfas/framework/input/TextDataSource.java b/pdf-as-lib/src/main/java/at/gv/egiz/pdfas/framework/input/TextDataSource.java new file mode 100644 index 0000000..7623896 --- /dev/null +++ b/pdf-as-lib/src/main/java/at/gv/egiz/pdfas/framework/input/TextDataSource.java @@ -0,0 +1,39 @@ +/**
+ * <copyright> Copyright 2006 by Know-Center, Graz, Austria </copyright>
+ * PDF-AS has been contracted by the E-Government Innovation Center EGIZ, a
+ * joint initiative of the Federal Chancellery Austria and Graz University of
+ * Technology.
+ *
+ * Licensed under the EUPL, Version 1.1 or - as soon they will be approved by
+ * the European Commission - subsequent versions of the EUPL (the "Licence");
+ * You may not use this work except in compliance with the Licence.
+ * You may obtain a copy of the Licence at:
+ * http://www.osor.eu/eupl/
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the Licence is distributed on an "AS IS" basis,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the Licence for the specific language governing permissions and
+ * limitations under the Licence.
+ *
+ * This product combines work with different licenses. See the "NOTICE" text
+ * file for details on the various modules and licenses.
+ * The "NOTICE" text file is part of the distribution. Any derivative works
+ * that you distribute must include a readable copy of the "NOTICE" text file.
+ */
+package at.gv.egiz.pdfas.framework.input;
+
+/**
+ * Represents a free-text input text to be processed.
+ *
+ * @author wprinz
+ */
+public interface TextDataSource extends DataSource
+{
+ /**
+ * Returns the text to be processed.
+ * @return Returns the text to be processed.
+ */
+ public String getText();
+
+}
diff --git a/pdf-as-lib/src/main/java/at/gv/egiz/pdfas/framework/input/correction/Corrector.java b/pdf-as-lib/src/main/java/at/gv/egiz/pdfas/framework/input/correction/Corrector.java new file mode 100644 index 0000000..71f3ded --- /dev/null +++ b/pdf-as-lib/src/main/java/at/gv/egiz/pdfas/framework/input/correction/Corrector.java @@ -0,0 +1,61 @@ +/**
+ * <copyright> Copyright 2006 by Know-Center, Graz, Austria </copyright>
+ * PDF-AS has been contracted by the E-Government Innovation Center EGIZ, a
+ * joint initiative of the Federal Chancellery Austria and Graz University of
+ * Technology.
+ *
+ * Licensed under the EUPL, Version 1.1 or - as soon they will be approved by
+ * the European Commission - subsequent versions of the EUPL (the "Licence");
+ * You may not use this work except in compliance with the Licence.
+ * You may obtain a copy of the Licence at:
+ * http://www.osor.eu/eupl/
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the Licence is distributed on an "AS IS" basis,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the Licence for the specific language governing permissions and
+ * limitations under the Licence.
+ *
+ * This product combines work with different licenses. See the "NOTICE" text
+ * file for details on the various modules and licenses.
+ * The "NOTICE" text file is part of the distribution. Any derivative works
+ * that you distribute must include a readable copy of the "NOTICE" text file.
+ */
+package at.gv.egiz.pdfas.framework.input.correction;
+
+import at.gv.egiz.pdfas.exceptions.framework.CorrectorException;
+import at.gv.egiz.pdfas.framework.input.PdfDataSource;
+
+/**
+ * Interface for PDF corretors.
+ *
+ * <p>
+ * Often PDF documents generated by various document to PDF converters have an
+ * invalid structure that upsets PDF-AS. The correction mechanism allows these
+ * documents to be corrected before being signed.
+ * </p>
+ * <p>
+ * A PDF corrector takes an incorrect PDF document and transforms it into a
+ * correct one.
+ * </p>
+ * <p>
+ * Note that correcting a document destroys all signatures in that document, so
+ * never correct an already signed document.
+ * </p>
+ *
+ * @author wprinz
+ */
+public interface Corrector
+{
+ /**
+ * Corrects the given PDF document to a form that PDF-AS can use.
+ *
+ * @param document
+ * The (incorrect) PDF document.
+ * @return Returns the corrected PDF document.
+ * @throws CorrectorException
+ * Exception thrown if the document couldn't be corrected.
+ */
+ public PdfDataSource correctDocument(PdfDataSource document) throws CorrectorException;
+
+}
diff --git a/pdf-as-lib/src/main/java/at/gv/egiz/pdfas/framework/input/correction/CorrectorFactory.java b/pdf-as-lib/src/main/java/at/gv/egiz/pdfas/framework/input/correction/CorrectorFactory.java new file mode 100644 index 0000000..8c954f7 --- /dev/null +++ b/pdf-as-lib/src/main/java/at/gv/egiz/pdfas/framework/input/correction/CorrectorFactory.java @@ -0,0 +1,63 @@ +/**
+ * <copyright> Copyright 2006 by Know-Center, Graz, Austria </copyright>
+ * PDF-AS has been contracted by the E-Government Innovation Center EGIZ, a
+ * joint initiative of the Federal Chancellery Austria and Graz University of
+ * Technology.
+ *
+ * Licensed under the EUPL, Version 1.1 or - as soon they will be approved by
+ * the European Commission - subsequent versions of the EUPL (the "Licence");
+ * You may not use this work except in compliance with the Licence.
+ * You may obtain a copy of the Licence at:
+ * http://www.osor.eu/eupl/
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the Licence is distributed on an "AS IS" basis,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the Licence for the specific language governing permissions and
+ * limitations under the Licence.
+ *
+ * This product combines work with different licenses. See the "NOTICE" text
+ * file for details on the various modules and licenses.
+ * The "NOTICE" text file is part of the distribution. Any derivative works
+ * that you distribute must include a readable copy of the "NOTICE" text file.
+ */
+package at.gv.egiz.pdfas.framework.input.correction;
+
+import at.gv.egiz.pdfas.impl.input.correction.ExternalCorrector;
+import at.gv.egiz.pdfas.impl.input.correction.InternalCorrector;
+import at.knowcenter.wag.egov.egiz.cfg.SettingsReader;
+import at.knowcenter.wag.egov.egiz.exceptions.SettingsException;
+
+/**
+ * Factory for creating Correctors.
+ *
+ * @author wprinz
+ */
+public class CorrectorFactory
+{
+ public static final String INTERNAL_CORRECTOR = "internal";
+
+ public static final String EXTERNAL_CORRECTOR = "external";
+
+ public static final String CORRECTOR_KEY = "corrector";
+
+ public static Corrector createCorrector(String id) throws SettingsException
+ {
+ if (id.equals(INTERNAL_CORRECTOR))
+ {
+ return new InternalCorrector();
+ }
+ if (id.equals(EXTERNAL_CORRECTOR))
+ {
+ return new ExternalCorrector();
+ }
+ throw new SettingsException("The connector id '" + id + "' is not a valid corrector id.");
+ }
+
+ public static Corrector createCorrector() throws SettingsException
+ {
+ String id = SettingsReader.getInstance().getSetting(CORRECTOR_KEY, INTERNAL_CORRECTOR);
+ return createCorrector(id);
+ }
+
+}
|