Metacat

Merge 1.9.2 changes back into the trunk

Modified readFromMetacat() to pass most exceptions up the call stack, which
allows creation of new entry points for calling reads. Still need to
continue factoring out the HTTPServletResponse that is passed in in order to
make entrypoints that are not servlet based possible. Problem now is in...

Beef up log messages

do a quick retry if building path index fails with a SQL error

Make sure buildIndex throws an exception if it has a sqlexception. That way the indexing object will be added to the indexing queue and reprocessed.

Pass the doc xml as a string to docImpl.write and writeRepication. This is so a reader can be create for the parsing and for the write to disk. Also created a db access class for xml query result deletion.

Fix the bug of http://bugzilla.ecoinformatics.org/show_bug.cgi?id=4645. handleGetRevisionAndDocTypeAction will search both xml_documents and xml_revisions table.
It also changed some constrain in AccessionNumber when user update a document, of which all previous versions are in xml_revisions table.

Change log levels where appropriate and add class/method name to output.

beef up error log messages

special handling for RDF namespace documents (semtools project)

Move access control source to it's own directory.

Change location of PropertyService to properties directory

Move document specific utilities to DocumentUtil from MetacatUtil. This makes it easier to define a layer between the core metacat services and the rest of the code.

Create database and shared directories for database management code and shared code respectively.

Add archival read funtionality (jar/kar/war files)

Introduce replication user. Use the fileutil writer methods instead of writing directly.

Beef up exception handling from file utilities. Move UtilException to MetacatUtilException to eliminate conflict with similar exception in utility package.

Do not read or write zero length documents to/from disk

Format indexPaths in metacat.properties. Remove from build.properties and build.xml. Move indexPath list getter from MetacatUtil to SystemUtil.

Fix typos in error messages.

Renamed MetaCatUtil to MetacatUtil

move readerToString function to StringUtil class in utilities module.

Add debug statements

Rename LDAPUtil to AuthUtil

Move the code to write metadata to disk into documentImpl

Add some generics typing. Separate the code that strips inline data from document files to have a different strategy for 2.0.X versus 2.1.X documents.

Look for schemaLocations in the document while initializing parser. If full schema validation is turned on in metacat.properties, and at least one schema is not registered locally, then turn on full schema validation in the parser.

Fix workgroup description

add closing </distribution> tag to regex for inline data.

Created a method in DocumentImpl to read a document file from disk. Created a support method to strip inline data for users that don't have read access.

Move the DBAdaptor accessor into a DatabaseService class

qualify xml and eml properties with an xml. prefix

Continue to qualify property names

Fully qualify spatial and replication properties

Use qualified properties

Add sql debug statements

Append context url onto system id instead of server url.

Merge 1.9 changes into Head

fixed bug 3403

Remove an obsolete private method getPartNodeRecordList.

Use dev=1 replace dev like '1'. since postgres 8.3 doesn't support it.

Add debug info for special charater.

Modified a sql command from "like" to "=". It dramatically improves the peformance of build index.

Add debug information.

Add code to expire cached query result when metacat have a insert, update or delete action.

Fix a bug that move node from xml_nodes to xml_nodes_revisions.

Decreased the debug message priority to 'debug' level for messages when reading an XML document.

Somehow the change went to the head rather than branch. So i rollback the change in head.

This commit is for branch. In this commit the correct ip and user name will be stored in access_log table in replication event. However, it only for Forcereplication. And the test isn't completed yet. This commit is only for future use.

Change some log info.

One more patch for bug 2469:
Although the correct parentid values were being indexed in xml_path_index
for leaf node xpaths, they were still incorrect for relative and absolute
paths. This patch modifies traverseParents() and changes the parent node id
to be indexed to that of the leaf node, no matter if the path is a leaf,...

As a continued fix for http://bugzilla.ecoinformatics.org/show_bug.cgi?id=2469,
I've fixed the indexing implementation in both buildIndex() and
traverseParents(). Duane pointed out that the incorrect parent node ids
were being indexed in xml_path_index, causing some stylesheets to render...

Fixed the implementation of the buildIndex function which was not
working for new document insertions. A previous fix in updatePathIndex
for ATTRIBUTE data inadvertantly caused a foreign key duplication
exception for insertions of ELEMENT nodes when multiple relative paths...

I'm fixing a compile problem under jdk 1.4.2, where the get() method in
HashMap needs an Object as a parameter, not a primitive data type. I changed
the long to a Long as the lookup key.

As part of a patch fix for:

http://bugzilla.ecoinformatics.org/show_bug.cgi?id=2469

I've changed DocumentImpl.java in three locations:

buildIndex()
  traverseParents()
  updatePathIndex()

This patch modifies buildIndex(). Like the prior two patches, it changes...

As part of a patch fix for:

http://bugzilla.ecoinformatics.org/show_bug.cgi?id=2469

I've changed DocumentImpl.java in three locations:

buildIndex()
traverseParents()
updatePathIndex()

This patch modifies traverseParents(). It changes pathsFoundForIndexing from...

As part of a patch fix for:

http://bugzilla.ecoinformatics.org/show_bug.cgi?id=2469

I've changed DocumentImpl.java in three locations:

buildIndex()
traverseParents()
updatePathIndex()

This patch modifies updatePathIndex() by changing the pathsFound input...

Ensure that spatial operations are only triggered when runSpatialOption is set to true.

Removed the release ant token from all files in 'src'.

Changed float object handling in spatial components in order to compile on jdk 1.4.2

initial import of changes to metacat source and config files to support spatial option

added exception handler for normalize function. if StringIndexOutOfBoundsException is thrown then it is reported and nodedata value is forwarded without normalization.

Allow admin to update any document

Modified code so that database is queried just once when a document is read. Currently the read is being done in the constructor and then again in
the function.

1. Modified buildIndex() so that only one connection is used for both indexing nodes and indexing paths.
2. Modified the indexing algo so that indexing of paths is done while nodes are being indexed. Results in a much faster indexing algo
3. Replaced access to NodeRecord.<variable> with NodeRecord.<getVariable>

Fixed the bug that replicate data file failed that the documents records wouldn't be deleted.

Changes in calls made to IndexingQueue class

Removed the call to indexing function in DBSAXHandler and added a call to indexing function IndexingQueue

Modified the debug levels of debug statements

Add code to handle catalog is null when inserting xml_revisions table.

Add code to handle write record into xml_documents when catalog id is null.

Bugfix to the code which sets parentnodeid in xml_path_index. (Use setInt instead of setFloat)

Changes in log levels of some commands mentioned in the previos commit.

Replacing MetaCatUtil.debugMessage or MetaCatUtil.logMetacat call with logMetacat (private Logger object) call

Fixed bug which can cause eml beta6 couldn't run access control.

Fixed bug couldn't delete xml_nodes when moving to xml_nodes_revision in replication.

Change the code to make sure DocumentImpl class's constructor should have rev attached.

Add code to transfer string to date type.

Add function transfer string to date.

Rewrite replication for revision document.

Fixed some bug in replication.

Add code to handle replicate the deleted document.

Add default constructor and change some viriable to protected.

Added code to start threads which initiate indexing once insertion to xml_nodes is complete.

Fix for bug 2060.

Moved the call to starting of indexing thread from endDocument to DocumentImpl after commit has been done. This way when ever a document is indexed it has already been entered in xml_nodes and xml_documents

Added new methods to MetaCatUtil: isAdministrator and isModerator.

These methods are used to check if a given username is part of the admin and moderator list specified in metacat.properties

Modified DocumentImpl and MetaCatServlet to use these functions.

Bugfix to the previous commit.

Also made changes so that while indexing the path in buildIndex(), instead of seperate database calls to the database for each path, one call is made with all the paths in it.

Modified code to enter/remove data from xml_path_index and enter data into xml_nodes_revision when action=insert, update and delete are performed.

Added code to delete all entries in xml_queryresult table when buildIndex for a docid is called.

Adding the call to normalize function for now -- too many calls to normalize function are made and this needs to be looked into. Removing call to normalize function caused trouble in text like this "A&B"

Removing call to normalize from getNodeRecordList()

Removed irrelevant code from previous commit

Removed code which entered value for nodedata in xml_index

Added code so that metacat administrator can delete any document.

Added code for updating xml_queryresult for action=delete and action=update

For both actions, the entries in xml_queryresult are deleted.
This works for action=update because deletion of entry is simple and the entries will be created again when the docid is part of a search result next time.

Modifying code so that nodedata is stored in xml_index table next to the paths.
This helps in making the search faster.

Modified code to fix bug # 1850

Modified buildIndex() to now include an '@' sign in the path for ATTRIBUTE
nodes. Removed a bunch of debugging information. Fixed the BuildIndexTest
so that it would work on any machine (removed hardcoded paths).

Whitespace changes that fix a few formatting problems after Jing's commit.

Add new feature that delete can be broadcasted by force replication.