From 3d982813b34f6f230baf4a467cdc37ec92a77595 Mon Sep 17 00:00:00 2001
From: netconomy <netconomy@7b5415b0-85f9-ee4d-85bd-d5d0c3b42d1c>
Date: Fri, 17 Aug 2007 06:10:56 +0000
Subject: Performance

git-svn-id: https://joinup.ec.europa.eu/svn/pdf-as/trunk@167 7b5415b0-85f9-ee4d-85bd-d5d0c3b42d1c
---
 .../framework/vfilter/VerificationFilter.java      | 52 +++++++++++++++
 .../vfilter/VerificationFilterParameters.java      | 76 ++++++++++++++++++++++
 2 files changed, 128 insertions(+)
 create mode 100644 src/main/java/at/gv/egiz/pdfas/framework/vfilter/VerificationFilter.java
 create mode 100644 src/main/java/at/gv/egiz/pdfas/framework/vfilter/VerificationFilterParameters.java

(limited to 'src/main/java/at/gv/egiz/pdfas/framework/vfilter')

diff --git a/src/main/java/at/gv/egiz/pdfas/framework/vfilter/VerificationFilter.java b/src/main/java/at/gv/egiz/pdfas/framework/vfilter/VerificationFilter.java
new file mode 100644
index 0000000..1633b09
--- /dev/null
+++ b/src/main/java/at/gv/egiz/pdfas/framework/vfilter/VerificationFilter.java
@@ -0,0 +1,52 @@
+/**
+ * 
+ */
+package at.gv.egiz.pdfas.framework.vfilter;
+
+import java.util.List;
+
+import at.gv.egiz.pdfas.exceptions.framework.VerificationFilterException;
+import at.gv.egiz.pdfas.framework.input.PdfDataSource;
+import at.gv.egiz.pdfas.framework.input.TextDataSource;
+
+/**
+ * Extracts all signatures from a given PDF document or text.
+ * 
+ * @see VerificationFilterParameters
+ * 
+ * @author wprinz
+ */
+public interface VerificationFilter
+{
+
+  /**
+   * Extracts the signatures from the given PDF.
+   * 
+   * @param pdf
+   *          The PDF.
+   * @param blocks
+   *          The List of Incremental Update blocks. Usually this comes from a
+   *          preprocessing step.
+   * @param parameters
+   *          The algorithm parameters.
+   * @return Returns a List of SignatureHolders containing the signatures. May
+   *         be empty in case no signatures have been found.
+   * @throws VerificationFilterException
+   *           Thrown if something goes wrong.
+   */
+  public List extractSignatureHolders(PdfDataSource pdf, List blocks, VerificationFilterParameters parameters) throws VerificationFilterException;
+
+  /**
+   * Extracts the text signatures from the given free-text.
+   * 
+   * @param text
+   *          The free-text.
+   * @param parameters
+   *          The algorithm parameters.
+   * @return Returns a List of SignatureHolders containing the signatures. May
+   *         be empty in case no signatures have been found.
+   * @throws VerificationFilterException
+   *           Thrown if something goes wrong.
+   */
+  public List extractSignaturHolders(TextDataSource text, VerificationFilterParameters parameters) throws VerificationFilterException;
+}
diff --git a/src/main/java/at/gv/egiz/pdfas/framework/vfilter/VerificationFilterParameters.java b/src/main/java/at/gv/egiz/pdfas/framework/vfilter/VerificationFilterParameters.java
new file mode 100644
index 0000000..c518fef
--- /dev/null
+++ b/src/main/java/at/gv/egiz/pdfas/framework/vfilter/VerificationFilterParameters.java
@@ -0,0 +1,76 @@
+/**
+ * 
+ */
+package at.gv.egiz.pdfas.framework.vfilter;
+
+/**
+ * The parameters of the VerificationFilter algorithm.
+ * 
+ * @author wprinz
+ */
+public interface VerificationFilterParameters
+{
+
+  /**
+   * Tells the VerificationFilter to extract binary signatures only.
+   * 
+   * <p>
+   * Not scanning for textual signatures allows the algorithm to skip text
+   * extraction and signature extraction, which are both time and memory
+   * intensive processes.
+   * </p>
+   * 
+   * @return Returns true if the VerificationFilter should extract binary
+   *         signatures only.
+   */
+  public boolean extractBinarySignaturesOnly();
+
+  /**
+   * Tells the VerificationFilter to assume that there are only singatures (and
+   * their Incremental Update blocks) younger than the original document.
+   * 
+   * <p>
+   * This is equivalent to saying that the document was not updated using an
+   * Incremental update block other than a signature after being singed. The
+   * incremental update blocks after the original document contain only
+   * signatures (either text or binary).
+   * </p>
+   * <p>
+   * This is equivalent to saying that there exists no Incremental Update block
+   * that would render a text signature before it invalid.
+   * </p>
+   * <p>
+   * Under this assumption, the process of finding all text signatures
+   * simplifies to one text extraction of the whole document and one signature
+   * extraction. This is of course a massive performance gain.
+   * </p>
+   * <p>
+   * Actually the algorithm performs a text extraction of the whole document not
+   * including trailing binary signature Incremental Update blocks. This means
+   * that if a the last n Incremental Update blocks of a document are binary,
+   * there is no use extract text from them.
+   * </p>
+   * <p>
+   * Note that if there are Incremental Update blocks with text after a
+   * signature thus this assumption does not hold the signatures older than this
+   * block will break.
+   * </p>
+   * 
+   * @return Returns true if the Verification filter should assume that there
+   *         are only signature blocks after the original document.
+   */
+  public boolean assumeOnlySignatureUpdateBlocks();
+
+  /**
+   * Tells the VerificationFilter so scan for old signatures in the rest text.
+   * 
+   * <p>
+   * The rest text is the text of the oldest text signature or the original
+   * document text if there is no text signature.
+   * </p>
+   * 
+   * @return Returns true if the VerificationFilter should scan for old text
+   *         signatures in the rest text.
+   */
+  public boolean scanForOldSignatures();
+}
-- 
cgit v1.2.3