From ec9c9b25f1fa1d0bd5f919b861c93d39a3e146f2 Mon Sep 17 00:00:00 2001
From: srosse <stephane.rosse@frentix.com>
Date: Fri, 26 Oct 2018 10:09:00 +0200
Subject: [PATCH] OO-3295: allow reloading qpool index if deleted, tune pdf
extractor
---
.../service/document/file/pdf/PdfBoxExtractor.java | 6 ++++--
.../document/file/pdf/PdfExternalExtractor.java | 2 +-
.../org/olat/search/service/indexer/JmsIndexer.java | 12 +++++++++---
3 files changed, 14 insertions(+), 6 deletions(-)
diff --git a/src/main/java/org/olat/search/service/document/file/pdf/PdfBoxExtractor.java b/src/main/java/org/olat/search/service/document/file/pdf/PdfBoxExtractor.java
index 82e9f0c4810..a73f2ffce94 100644
--- a/src/main/java/org/olat/search/service/document/file/pdf/PdfBoxExtractor.java
+++ b/src/main/java/org/olat/search/service/document/file/pdf/PdfBoxExtractor.java
@@ -64,7 +64,7 @@ public class PdfBoxExtractor implements PdfExtractor {
}
}
- private FileContent extractTextFromPdf(VFSLeaf leaf) throws IOException, DocumentAccessException {
+ private FileContent extractTextFromPdf(VFSLeaf leaf) throws IOException {
if (log.isDebug()) log.debug("readContent from pdf starts...");
try(BufferedInputStream bis = new BufferedInputStream(leaf.getInputStream());
@@ -72,6 +72,8 @@ public class PdfBoxExtractor implements PdfExtractor {
String title = getTitle(document);
if (log.isDebug()) log.debug("readContent PDDocument loaded");
PDFTextStripper stripper = new PDFTextStripper();
+ stripper.setSortByPosition(true);
+ stripper.setSuppressDuplicateOverlappingText(true);
LimitedContentWriter writer = new LimitedContentWriter(50000, FileDocumentFactory.getMaxFileSize());
stripper.writeText(document, writer);
writer.close();
@@ -88,7 +90,7 @@ public class PdfBoxExtractor implements PdfExtractor {
return new FileContent(leaf.getName(), writer.toString());
} catch(Exception e) {
log.error("", e);
- return null;
+ return new FileContent("", "");
}
}
diff --git a/src/main/java/org/olat/search/service/document/file/pdf/PdfExternalExtractor.java b/src/main/java/org/olat/search/service/document/file/pdf/PdfExternalExtractor.java
index e59b3af14fc..d9872f14a63 100644
--- a/src/main/java/org/olat/search/service/document/file/pdf/PdfExternalExtractor.java
+++ b/src/main/java/org/olat/search/service/document/file/pdf/PdfExternalExtractor.java
@@ -66,7 +66,7 @@ public class PdfExternalExtractor implements PdfExtractor {
return;
}
- List<String> cmds = new ArrayList<String>();
+ List<String> cmds = new ArrayList<>();
cmds.add(searchModule.getPdfExternalIndexerCmd());
cmds.add(((LocalFileImpl)document).getBasefile().getAbsolutePath());
cmds.add(bufferFile.getAbsolutePath());
diff --git a/src/main/java/org/olat/search/service/indexer/JmsIndexer.java b/src/main/java/org/olat/search/service/indexer/JmsIndexer.java
index 8f558ff03ee..82b7a6a02b6 100644
--- a/src/main/java/org/olat/search/service/indexer/JmsIndexer.java
+++ b/src/main/java/org/olat/search/service/indexer/JmsIndexer.java
@@ -334,9 +334,15 @@ public class JmsIndexer implements MessageListener, LifeFullIndexer, ConfigOnOff
}
private DirectoryReader getReader() throws IOException {
- DirectoryReader newReader = DirectoryReader.openIfChanged(reader);
- if(newReader != null) {
- reader = newReader;
+ if(reader == null) {
+ File tempIndexDir = new File(permanentIndexPath);
+ Directory indexPath = FSDirectory.open(tempIndexDir.toPath());
+ reader = DirectoryReader.open(indexPath);
+ } else {
+ DirectoryReader newReader = DirectoryReader.openIfChanged(reader);
+ if(newReader != null) {
+ reader = newReader;
+ }
}
return reader;
}
--
GitLab