From 83b55b7801cbf1c37971eb530ddf9b0f46a11536 Mon Sep 17 00:00:00 2001
From: Kensuke Matsuzaki <zakki@peppermint.jp>
Date: Sun, 25 Mar 2012 17:25:04 -0400
Subject: [PATCH] Add Japanese massages
---
src/com/gitblit/LuceneExecutor.java | 246 ++++++++++++++++++++++++++----------------------
1 files changed, 133 insertions(+), 113 deletions(-)
diff --git a/src/com/gitblit/LuceneExecutor.java b/src/com/gitblit/LuceneExecutor.java
index 489d308..cd7074b 100644
--- a/src/com/gitblit/LuceneExecutor.java
+++ b/src/com/gitblit/LuceneExecutor.java
@@ -25,7 +25,6 @@
import java.text.MessageFormat;
import java.text.ParseException;
import java.util.ArrayList;
-import java.util.Arrays;
import java.util.Collections;
import java.util.Comparator;
import java.util.HashMap;
@@ -138,35 +137,33 @@
private final Map<String, IndexSearcher> searchers = new ConcurrentHashMap<String, IndexSearcher>();
private final Map<String, IndexWriter> writers = new ConcurrentHashMap<String, IndexWriter>();
- private final Set<String> excludedExtensions = new TreeSet<String>(Arrays.asList("7z", "arc",
- "arj", "bin", "bmp", "dll", "doc", "docx", "exe", "gif", "gz", "jar", "jpg", "lib",
- "lzh", "odg", "pdf", "ppt", "png", "so", "swf", "xcf", "xls", "xlsx", "zip"));
-
+ private final String luceneIgnoreExtensions = "7z arc arj bin bmp dll doc docx exe gif gz jar jpg lib lzh odg odf odt pdf ppt png so swf xcf xls xlsx zip";
+ private Set<String> excludedExtensions;
+
public LuceneExecutor(IStoredSettings settings, File repositoriesFolder) {
this.storedSettings = settings;
this.repositoriesFolder = repositoriesFolder;
+ String exts = luceneIgnoreExtensions;
+ if (settings != null) {
+ exts = settings.getString(Keys.web.luceneIgnoreExtensions, exts);
+ }
+ excludedExtensions = new TreeSet<String>(StringUtils.getStringsFromValue(exts));
}
/**
- * Indicates if the Lucene executor can index repositories.
- *
- * @return true if the Lucene executor is ready to index repositories
- */
- public boolean isReady() {
- return storedSettings.getBoolean(Keys.lucene.enable, false);
- }
-
- /**
- * Run is executed by the gitblit executor service at whatever frequency
- * is specified in the settings. Because this is called by an executor
- * service, calls will queue - i.e. there can never be concurrent execution
- * of repository index updates.
+ * Run is executed by the Gitblit executor service. Because this is called
+ * by an executor service, calls will queue - i.e. there can never be
+ * concurrent execution of repository index updates.
*/
@Override
public void run() {
- if (!isReady()) {
+ if (!storedSettings.getBoolean(Keys.web.allowLuceneIndexing, true)) {
+ // Lucene indexing is disabled
return;
}
+ // reload the excluded extensions
+ String exts = storedSettings.getString(Keys.web.luceneIgnoreExtensions, luceneIgnoreExtensions);
+ excludedExtensions = new TreeSet<String>(StringUtils.getStringsFromValue(exts));
for (String repositoryName: GitBlit.self().getRepositoryList()) {
RepositoryModel model = GitBlit.self().getRepositoryModel(repositoryName);
@@ -188,7 +185,7 @@
* @param repository
* the repository object
*/
- protected void index(RepositoryModel model, Repository repository) {
+ private void index(RepositoryModel model, Repository repository) {
try {
if (shouldReindex(repository)) {
// (re)build the entire index
@@ -228,7 +225,16 @@
*
* @param repositoryName
*/
- public void close(String repositoryName) {
+ public synchronized void close(String repositoryName) {
+ try {
+ IndexSearcher searcher = searchers.remove(repositoryName);
+ if (searcher != null) {
+ searcher.getIndexReader().close();
+ }
+ } catch (Exception e) {
+ logger.error("Failed to close index searcher for " + repositoryName, e);
+ }
+
try {
IndexWriter writer = writers.remove(repositoryName);
if (writer != null) {
@@ -236,23 +242,14 @@
}
} catch (Exception e) {
logger.error("Failed to close index writer for " + repositoryName, e);
- }
-
- try {
- IndexSearcher searcher = searchers.remove(repositoryName);
- if (searcher != null) {
- searcher.close();
- }
- } catch (Exception e) {
- logger.error("Failed to close index searcher for " + repositoryName, e);
- }
+ }
}
/**
* Close all Lucene indexers.
*
*/
- public void close() {
+ public synchronized void close() {
// close all writers
for (String writer : writers.keySet()) {
try {
@@ -266,7 +263,7 @@
// close all searchers
for (String searcher : searchers.keySet()) {
try {
- searchers.get(searcher).close();
+ searchers.get(searcher).getIndexReader().close();
} catch (Throwable t) {
logger.error("Failed to close Lucene searcher for " + searcher, t);
}
@@ -283,18 +280,9 @@
*/
public boolean deleteIndex(String repositoryName) {
try {
- // remove the repository index writer from the cache and close it
- IndexWriter writer = writers.remove(repositoryName);
- if (writer != null) {
- writer.close();
- writer = null;
- }
- // remove the repository index searcher from the cache and close it
- IndexSearcher searcher = searchers.remove(repositoryName);
- if (searcher != null) {
- searcher.close();
- searcher = null;
- }
+ // close any open writer/searcher
+ close(repositoryName);
+
// delete the index folder
File repositoryFolder = new File(repositoriesFolder, repositoryName);
File luceneIndex = new File(repositoryFolder, LUCENE_DIR);
@@ -358,7 +346,7 @@
* @return tree
* @throws IOException
*/
- protected RevTree getTree(final RevWalk walk, final RevCommit commit)
+ private RevTree getTree(final RevWalk walk, final RevCommit commit)
throws IOException {
final RevTree tree = commit.getTree();
if (tree != null) {
@@ -398,7 +386,7 @@
* @param repository
* @return true of the on-disk index format is different than INDEX_VERSION
*/
- protected boolean shouldReindex(Repository repository) {
+ private boolean shouldReindex(Repository repository) {
try {
FileBasedConfig config = getConfig(repository);
config.load();
@@ -420,7 +408,7 @@
* @return IndexResult
*/
public IndexResult reindex(RepositoryModel model, Repository repository) {
- IndexResult result = new IndexResult();
+ IndexResult result = new IndexResult();
if (!deleteIndex(model.name)) {
return result;
}
@@ -553,7 +541,7 @@
// index the blob content
if (StringUtils.isEmpty(ext) || !excludedExtensions.contains(ext)) {
ObjectLoader ldr = repository.open(blobId, Constants.OBJ_BLOB);
- InputStream in = ldr.openStream();
+ InputStream in = ldr.openStream();
int n;
while ((n = in.read(tmp)) > 0) {
os.write(tmp, 0, n);
@@ -615,8 +603,8 @@
// commit all changes and reset the searcher
config.setInt(CONF_INDEX, null, CONF_VERSION, INDEX_VERSION);
config.save();
- resetIndexSearcher(model.name);
writer.commit();
+ resetIndexSearcher(model.name);
result.success();
} catch (Exception e) {
logger.error("Exception while reindexing " + model.name, e);
@@ -645,7 +633,7 @@
IndexWriter writer = getIndexWriter(repositoryName);
for (PathChangeModel path : changedPaths) {
// delete the indexed blob
- deleteBlob(repositoryName, branch, path.path);
+ deleteBlob(repositoryName, branch, path.name);
// re-index the blob
if (!ChangeType.DELETE.equals(path.changeType)) {
@@ -678,8 +666,17 @@
}
}
writer.commit();
-
- Document doc = createDocument(commit, null);
+
+ // get any annotated commit tags
+ List<String> commitTags = new ArrayList<String>();
+ for (RefModel ref : JGitUtils.getTags(repository, true, -1)) {
+ if (ref.isAnnotatedTag() && ref.getReferencedObjectId().equals(commit.getId())) {
+ commitTags.add(ref.displayName);
+ }
+ }
+
+ // create and write the Lucene document
+ Document doc = createDocument(commit, commitTags);
doc.add(new Field(FIELD_BRANCH, branch, Store.YES, Index.ANALYZED));
result.commitCount++;
result.success = index(repositoryName, doc);
@@ -757,7 +754,7 @@
* @param repository
* @return IndexResult
*/
- protected IndexResult updateIndex(RepositoryModel model, Repository repository) {
+ private IndexResult updateIndex(RepositoryModel model, Repository repository) {
IndexResult result = new IndexResult();
try {
FileBasedConfig config = getConfig(repository);
@@ -931,8 +928,8 @@
try {
IndexWriter writer = getIndexWriter(repositoryName);
writer.addDocument(doc);
- resetIndexSearcher(repositoryName);
writer.commit();
+ resetIndexSearcher(repositoryName);
return true;
} catch (Exception e) {
logger.error(MessageFormat.format("Exception while incrementally updating {0} Lucene index", repositoryName), e);
@@ -940,8 +937,10 @@
return false;
}
- private SearchResult createSearchResult(Document doc, float score) throws ParseException {
+ private SearchResult createSearchResult(Document doc, float score, int hitId, int totalHits) throws ParseException {
SearchResult result = new SearchResult();
+ result.hitId = hitId;
+ result.totalHits = totalHits;
result.score = score;
result.date = DateTools.stringToDate(doc.get(FIELD_DATE));
result.summary = doc.get(FIELD_SUMMARY);
@@ -964,7 +963,7 @@
private synchronized void resetIndexSearcher(String repository) throws IOException {
IndexSearcher searcher = searchers.remove(repository);
if (searcher != null) {
- searcher.close();
+ searcher.getIndexReader().close();
}
}
@@ -1017,19 +1016,21 @@
*
* @param text
* if the text is null or empty, null is returned
- * @param maximumHits
- * the maximum number of hits to collect
+ * @param page
+ * the page number to retrieve. page is 1-indexed.
+ * @param pageSize
+ * the number of elements to return for this page
* @param repositories
* a list of repositories to search. if no repositories are
* specified null is returned.
* @return a list of SearchResults in order from highest to the lowest score
*
*/
- public List<SearchResult> search(String text, int maximumHits, List<String> repositories) {
+ public List<SearchResult> search(String text, int page, int pageSize, List<String> repositories) {
if (ArrayUtils.isEmpty(repositories)) {
return null;
}
- return search(text, maximumHits, repositories.toArray(new String[0]));
+ return search(text, page, pageSize, repositories.toArray(new String[0]));
}
/**
@@ -1037,15 +1038,17 @@
*
* @param text
* if the text is null or empty, null is returned
- * @param maximumHits
- * the maximum number of hits to collect
+ * @param page
+ * the page number to retrieve. page is 1-indexed.
+ * @param pageSize
+ * the number of elements to return for this page
* @param repositories
* a list of repositories to search. if no repositories are
* specified null is returned.
* @return a list of SearchResults in order from highest to the lowest score
*
- */
- public List<SearchResult> search(String text, int maximumHits, String... repositories) {
+ */
+ public List<SearchResult> search(String text, int page, int pageSize, String... repositories) {
if (StringUtils.isEmpty(text)) {
return null;
}
@@ -1082,14 +1085,15 @@
searcher = new IndexSearcher(reader);
}
Query rewrittenQuery = searcher.rewrite(query);
- TopScoreDocCollector collector = TopScoreDocCollector.create(maximumHits, true);
+ TopScoreDocCollector collector = TopScoreDocCollector.create(5000, true);
searcher.search(rewrittenQuery, collector);
- ScoreDoc[] hits = collector.topDocs().scoreDocs;
+ int offset = Math.max(0, (page - 1) * pageSize);
+ ScoreDoc[] hits = collector.topDocs(offset, pageSize).scoreDocs;
+ int totalHits = collector.getTotalHits();
for (int i = 0; i < hits.length; i++) {
int docId = hits[i].doc;
Document doc = searcher.doc(docId);
- // TODO identify the source index for the doc, then eliminate FIELD_REPOSITORY
- SearchResult result = createSearchResult(doc, hits[i].score);
+ SearchResult result = createSearchResult(doc, hits[i].score, offset + i + 1, totalHits);
if (repositories.length == 1) {
// single repository search
result.repository = repositories[0];
@@ -1121,42 +1125,82 @@
*/
private String getHighlightedFragment(Analyzer analyzer, Query query,
String content, SearchResult result) throws IOException, InvalidTokenOffsetsException {
- content = content == null ? "":StringUtils.escapeForHtml(content, false);
-
+ if (content == null) {
+ content = "";
+ }
+
+ int fragmentLength = SearchObjectType.commit == result.type ? 512 : 150;
+
QueryScorer scorer = new QueryScorer(query, "content");
- Fragmenter fragmenter;
-
- // TODO improve the fragmenter - hopefully on line breaks
- if (SearchObjectType.commit == result.type) {
- fragmenter = new SimpleSpanFragmenter(scorer, 1024);
- } else {
- fragmenter = new SimpleSpanFragmenter(scorer, 150);
- }
+ Fragmenter fragmenter = new SimpleSpanFragmenter(scorer, fragmentLength);
// use an artificial delimiter for the token
- String termTag = "<!--[";
- String termTagEnd = "]-->";
+ String termTag = "!!--[";
+ String termTagEnd = "]--!!";
SimpleHTMLFormatter formatter = new SimpleHTMLFormatter(termTag, termTagEnd);
Highlighter highlighter = new Highlighter(formatter, scorer);
highlighter.setTextFragmenter(fragmenter);
-
- String [] fragments = highlighter.getBestFragments(analyzer, "content", content, 5);
+
+ String [] fragments = highlighter.getBestFragments(analyzer, "content", content, 3);
if (ArrayUtils.isEmpty(fragments)) {
if (SearchObjectType.blob == result.type) {
return "";
}
- return "<pre class=\"text\">" + content + "</pre>";
+ // clip commit message
+ String fragment = content;
+ if (fragment.length() > fragmentLength) {
+ fragment = fragment.substring(0, fragmentLength) + "...";
+ }
+ return "<pre class=\"text\">" + StringUtils.escapeForHtml(fragment, true) + "</pre>";
}
+
+ int contentPos = 0;
StringBuilder sb = new StringBuilder();
for (int i = 0, len = fragments.length; i < len; i++) {
String fragment = fragments[i];
-
+ String tag = "<pre class=\"text\">";
+
// resurrect the raw fragment from removing the artificial delimiters
- String raw = fragment.replace(termTag, "").replace(termTagEnd, "");
- sb.append(getPreTag(result, raw, content));
+ String raw = fragment.replace(termTag, "").replace(termTagEnd, "");
+
+ // determine position of the raw fragment in the content
+ int pos = content.indexOf(raw, contentPos);
+
+ // restore complete first line of fragment
+ int c = pos;
+ while (c > 0) {
+ c--;
+ if (content.charAt(c) == '\n') {
+ break;
+ }
+ }
+ if (c > 0) {
+ // inject leading chunk of first fragment line
+ fragment = content.substring(c + 1, pos) + fragment;
+ }
+
+ if (SearchObjectType.blob == result.type) {
+ // count lines as offset into the content for this fragment
+ int line = Math.max(1, StringUtils.countLines(content.substring(0, pos)));
+
+ // create fragment tag with line number and language
+ String lang = "";
+ String ext = StringUtils.getFileExtension(result.path).toLowerCase();
+ if (!StringUtils.isEmpty(ext)) {
+ // maintain leading space!
+ lang = " lang-" + ext;
+ }
+ tag = MessageFormat.format("<pre class=\"prettyprint linenums:{0,number,0}{1}\">", line, lang);
+
+ // update offset into content
+ contentPos = pos + raw.length() + 1;
+ }
+ sb.append(tag);
+
// replace the artificial delimiter with html tags
- String html = fragment.replace(termTag, "<span class=\"highlight\">").replace(termTagEnd, "</span>");
+ String html = StringUtils.escapeForHtml(fragment, false);
+ html = html.replace(termTag, "<span class=\"highlight\">").replace(termTagEnd, "</span>");
sb.append(html);
sb.append("</pre>");
if (i < len - 1) {
@@ -1164,31 +1208,7 @@
}
}
return sb.toString();
- }
-
- /**
- * Returns the appropriate tag for a fragment. Commit messages are visually
- * differentiated from blob fragments.
- *
- * @param result
- * @param fragment
- * @param content
- * @return an html tag appropriate for the fragment
- */
- private String getPreTag(SearchResult result, String fragment, String content) {
- String pre = "<pre class=\"text\">";
- if (SearchObjectType.blob == result.type) {
- int line = StringUtils.countLines(content.substring(0, content.indexOf(fragment)));
- int lastDot = result.path.lastIndexOf('.');
- if (lastDot > -1) {
- String ext = result.path.substring(lastDot + 1).toLowerCase();
- pre = MessageFormat.format("<pre class=\"prettyprint linenums:{0,number,0} lang-{1}\">", line, ext);
- } else {
- pre = MessageFormat.format("<pre class=\"prettyprint linenums:{0,number,0}\">", line);
- }
- }
- return pre;
- }
+ }
/**
* Simple class to track the results of an index update.
--
Gitblit v1.9.1