boxtec
/
yacy_search_server
mirror of https://github.com/yacy/yacy_search_server
4
1
Fork 0
Code Issues Projects Releases Wiki Activity

update to pdf parser

Browse Source
pull/1/head
Michael Peter Christen 12 years ago
parent 4a9182ae16
commit 95712fdc8b
12 changed files with 21 additions and 21 deletions
Show Stats Download Patch File Download Diff File

10
.classpath
Unescape View File

@ -46,9 +46,6 @@
<classpathentry kind="lib" path="lib/servlet-api-2.5-20081211.jar"/>
<classpathentry kind="lib" path="lib/jetty-6.1.26-patched-JETTY-1340.jar"/>
<classpathentry kind="lib" path="lib/jetty-util-6.1.26-patched-JETTY-1340.jar"/>
<classpathentry kind="lib" path="lib/fontbox-1.7.0.jar"/>
<classpathentry kind="lib" path="lib/pdfbox-1.7.0.jar"/>
<classpathentry kind="lib" path="lib/jempbox-1.7.0.jar"/>
<classpathentry kind="lib" path="lib/jaudiotagger-2.0.4-20111207.115108-15.jar"/>
<classpathentry kind="lib" path="lib/commons-codec-1.7.jar"/>
<classpathentry kind="lib" path="lib/jcl-over-slf4j-1.6.4.jar"/>
@ -59,17 +56,20 @@
<classpathentry kind="lib" path="lib/apache-solr-solrj-4.0.0.jar" sourcepath="/Users/admin/Development/sourcecode/apache-solr-4.0.0-src/src/java"/>
<classpathentry kind="lib" path="lib/lucene-analyzers-common-4.0.0.jar"/>
<classpathentry kind="lib" path="lib/lucene-analyzers-phonetic-4.0.0.jar"/>
<classpathentry kind="lib" path="lib/lucene-core-4.0.0.jar"/>
<classpathentry kind="lib" path="lib/lucene-core-4.0.0.jar" sourcepath="/Users/admin/Development/sourcecode/apache-solr-4.0.0-src/src/java"/>
<classpathentry kind="lib" path="lib/lucene-grouping-4.0.0.jar"/>
<classpathentry kind="lib" path="lib/lucene-highlighter-4.0.0.jar"/>
<classpathentry kind="lib" path="lib/lucene-memory-4.0.0.jar"/>
<classpathentry kind="lib" path="lib/lucene-misc-4.0.0.jar"/>
<classpathentry kind="lib" path="lib/lucene-misc-4.0.0.jar" sourcepath="/Users/admin/Development/sourcecode/apache-solr-4.0.0-src/src/java"/>
<classpathentry kind="lib" path="lib/lucene-queries-4.0.0.jar"/>
<classpathentry kind="lib" path="lib/lucene-queryparser-4.0.0.jar"/>
<classpathentry kind="lib" path="lib/lucene-spatial-4.0.0.jar"/>
<classpathentry kind="lib" path="lib/lucene-suggest-4.0.0.jar"/>
<classpathentry kind="lib" path="lib/zookeeper-3.3.6.jar"/>
<classpathentry kind="lib" path="lib/spatial4j-0.3.jar"/>
<classpathentry kind="lib" path="lib/pdfbox-1.7.1.jar"/>
<classpathentry kind="lib" path="lib/fontbox-1.7.1.jar"/>
<classpathentry kind="lib" path="lib/jempbox-1.7.1.jar"/>
<classpathentry kind="con" path="org.eclipse.jdt.junit.JUNIT_CONTAINER/4"/>
<classpathentry kind="lib" path="lib/icu4j-core.jar"/>
<classpathentry kind="lib" path="lib/htmllexer.jar"/>

6
addon/YaCy.app/Contents/Info.plist
Unescape View File

@ -53,7 +53,7 @@
<string>$JAVAROOT/lib/commons-jxpath-1.3.jar</string>
<string>$JAVAROOT/lib/commons-lang-2.6.jar</string>
<string>$JAVAROOT/lib/commons-logging-1.1.1.jar</string>
<string>$JAVAROOT/lib/fontbox-1.7.0.jar</string>
<string>$JAVAROOT/lib/fontbox-1.7.1.jar</string>
<string>$JAVAROOT/lib/geronimo-stax-api_1.0_spec-1.0.1.jar</string>
<string>$JAVAROOT/lib/guava-r05.jar</string>
<string>$JAVAROOT/lib/htmllexer.jar</string>
@ -66,7 +66,7 @@
<string>$JAVAROOT/lib/jaudiotagger-2.0.4-20111207.115108-15.jar</string>
<string>$JAVAROOT/lib/jcifs-1.3.15.jar</string>
<string>$JAVAROOT/lib/jcl-over-slf4j-1.6.4.jar</string>
<string>$JAVAROOT/lib/jempbox-1.7.0.jar</string>
<string>$JAVAROOT/lib/jempbox-1.7.1.jar</string>
<string>$JAVAROOT/lib/jena-2.6.4.jar</string>
<string>$JAVAROOT/lib/jetty-6.1.26-patched-JETTY-1340.jar</string>
<string>$JAVAROOT/lib/jetty-util-6.1.26-patched-JETTY-1340.jar</string>
@ -87,7 +87,7 @@
<string>$JAVAROOT/lib/lucene-suggest-4.0.0.jar</string>
<string>$JAVAROOT/lib/metadata-extractor-2.4.0-beta-1.jar</string>
<string>$JAVAROOT/lib/mysql-connector-java-5.1.12-bin.jar</string>
<string>$JAVAROOT/lib/pdfbox-1.7.0.jar</string>
<string>$JAVAROOT/lib/pdfbox-1.7.1.jar</string>
<string>$JAVAROOT/lib/poi-3.6-20091214.jar</string>
<string>$JAVAROOT/lib/poi-scratchpad-3.6-20091214.jar</string>
<string>$JAVAROOT/lib/sax-2.0.1.jar</string>

6
build.xml
Unescape View File

@ -172,7 +172,7 @@
<pathelement location="${lib}/commons-jxpath-1.3.jar" />
<pathelement location="${lib}/commons-lang-2.6.jar" />
<pathelement location="${lib}/commons-logging-1.1.1.jar" />
<pathelement location="${lib}/fontbox-1.7.0.jar" />
<pathelement location="${lib}/fontbox-1.7.1.jar" />
<pathelement location="${lib}/geronimo-stax-api_1.0_spec-1.0.1.jar" />
<pathelement location="${lib}/guava-r05.jar" />
<pathelement location="${lib}/htmllexer.jar" />
@ -185,7 +185,7 @@
<pathelement location="${lib}/jaudiotagger-2.0.4-20111207.115108-15.jar" />
<pathelement location="${lib}/jcifs-1.3.15.jar" />
<pathelement location="${lib}/jcl-over-slf4j-1.6.4.jar" />
<pathelement location="${lib}/jempbox-1.7.0" />
<pathelement location="${lib}/jempbox-1.7.1" />
<pathelement location="${lib}/jena-2.6.4.jar" />
<pathelement location="${lib}/jetty-6.1.26-patched-JETTY-1340.jar" />
<pathelement location="${lib}/jetty-util-6.1.26-patched-JETTY-1340.jar" />
@ -207,7 +207,7 @@
<pathelement location="${lib}/lucene-suggest-4.0.0.jar" />
<pathelement location="${lib}/metadata-extractor-2.4.0-beta-1.jar" />
<pathelement location="${lib}/mysql-connector-java-5.1.12-bin.jar" />
<pathelement location="${lib}/pdfbox-1.7.0.jar" />
<pathelement location="${lib}/pdfbox-1.7.1.jar" />
<pathelement location="${lib}/poi-3.6-20091214.jar" />
<pathelement location="${lib}/poi-scratchpad-3.6-20091214.jar" />
<pathelement location="${lib}/sax-2.0.1.jar" />

0
lib/apache-solr-core-3.6.1.License → lib/apache-solr-core-4.0.0.License
Unescape View File

0
lib/apache-solr-solrj-3.6.1.License → lib/apache-solr-solrj-4.0.0.License
Unescape View File

0
lib/fontbox-1.7.0.License → lib/fontbox-1.7.1.License
Unescape View File

BIN
lib/fontbox-1.7.0.jar → lib/fontbox-1.7.1.jar
View File

Binary file not shown.

0
lib/jempbox-1.7.0.License → lib/jempbox-1.7.1.License
Unescape View File

BIN
lib/jempbox-1.7.0.jar → lib/jempbox-1.7.1.jar
View File

Binary file not shown.

0
lib/pdfbox-1.7.0.License → lib/pdfbox-1.7.1.License
Unescape View File

BIN
lib/pdfbox-1.7.0.jar → lib/pdfbox-1.7.1.jar
View File

Binary file not shown.

20
source/net/yacy/document/parser/pdfParser.java
Unescape View File

@ -33,16 +33,6 @@ import java.io.FileNotFoundException;
import java.io.IOException;
import java.io.InputStream;
import net.yacy.cora.document.MultiProtocolURI;
import net.yacy.document.AbstractParser;
import net.yacy.document.Document;
import net.yacy.document.Parser;
import net.yacy.kelondro.data.meta.DigestURI;
import net.yacy.kelondro.io.CharBuffer;
import net.yacy.kelondro.logging.Log;
import net.yacy.kelondro.util.FileUtils;
import net.yacy.kelondro.util.MemoryControl;
import org.apache.pdfbox.cos.COSName;
import org.apache.pdfbox.exceptions.CryptographyException;
import org.apache.pdfbox.pdmodel.PDDocument;
@ -53,6 +43,16 @@ import org.apache.pdfbox.pdmodel.encryption.StandardDecryptionMaterial;
import org.apache.pdfbox.pdmodel.font.PDFont;
import org.apache.pdfbox.util.PDFTextStripper;
import net.yacy.cora.document.MultiProtocolURI;
import net.yacy.document.AbstractParser;
import net.yacy.document.Document;
import net.yacy.document.Parser;
import net.yacy.kelondro.data.meta.DigestURI;
import net.yacy.kelondro.io.CharBuffer;
import net.yacy.kelondro.logging.Log;
import net.yacy.kelondro.util.FileUtils;
import net.yacy.kelondro.util.MemoryControl;
public class pdfParser extends AbstractParser implements Parser {

Write Preview
Loading…
Cancel
Save