ReleaseNotes

2013.10.14 ([r480])

TagCloudSetup updated.

A full dump of release 480 is available at:

• http://sobre.arquivo.pt/~pwa/PWA-TechnologiesSourceCodeDump22-11-2013/
• https://drive.google.com/folderview?id=0B9fSuJY2vxOESmltSmlIUTJoWDA&usp=sharing
• https://www.dropbox.com/sh/qtdfrw4xu3qxiqn/Whva8XSvfB
• http://visibilidade.net/awp/PWA-TechnologiesSourceCodeDump22-11-2013/
• FCCN address (DVDs and Tape backups, check www.fccn.pt): Campus do Laboratório Nacional de Engenharia Civil, LNEC, Av. do Brasil n.º 101, 1700-066 LISBOA, Portugal
• FCCN data center: backup at server t1.tomba.fccn.pt:/home/nutchwax/pwa/PWA-TechnologiesSourceCodeDump22-11-2013

2013.08.05 ([r478])

Homepage was updated with new values of archived data.

2013.05.17 ([r475])

New:

• Added inspector utility for debugging purposes

Change:

• Corrected jar files dependencies
• Fixed XSS vulnerability in spell checker project
• Minor fixes from code quality platform
• Changed the location of hunspell

2012.07.16 ([r433])

Change:

• Changed information of the number of documents available for searching
• Added link to see more highlights

2012.07.10 ([r411])

New:

• Path to blacklist file is configurable. To allow adding new documents to blacklist, we added possibility of changing the path to the blacklist.

Change:

• Added web analytics tracking when displaying archived pages
• Only translate the datepicker to PT when browsing in that language.

2012.3.30 ([r401])

Changes:

• A wrong configuration was presenting the result page as the intro page.

2012.3.20 ([r400])

Changes:

• Added information for testing

2012.3.16 ([r399])

New:

• "PwaLogsMiner" project added. This code is used for processing the access logs
• Apply TFxIDF and BM25 functions over all index fields.
• Added a new temporal feature for the ranking algorithm. It uses the number of versions and time span between changed versions.

Changes:

• Changed the threshold for the minimum number of terms that occur only in the anchor field when matching a document.
• Adding dependencies no longer available on repositories.
• Updated the Portuguese texts to respect the new Portuguese Ortographic Agreement.
• Increased the number of matched documents returned by Nutch to 2K.
• Several bugfixes in the TFxIDF code.
• Uped the number of anchor terms used by the ranking algorithm.
• Fixed the project building that failed due to missing libs and classes.
• Created an in-project maven repository to handle deprecated/old dependencies.
• Index servers response timeout set to 60 seconds.
• The memcache timeout set to 5 seconds.
• Updated information and resources links for project resources in Maven file.
• Other fixes.

Removed:

• Remove the WERA sub-project. We don't use it and seems deprecated.

2012.1.10 ([r346])

Changes:

• Fixed a problem where queries with quotes returned zero results.

2012.1.5 ([r345])

New:

• Added the code of the "PwaProcessor" project. This project facilitates the processing of archived files in the ARC format.

Changes:

• The grid of versions on the history page was incorrectly being filled.
• One of the i18n files wasn't being called.
• Remove old, uneeded UI files.
• One of the i18n files wasn't being called.

2012.1.2 ([r337])

New:

• New graphical layout
• Added several new term-based and temporal ranking features (BM25, TFxIDF, PageRank, URLDepth, Age, BoostNewer, BoostOlder, URLSlash, BoostNewerOlder, LuceneSimilarity, MinSpan, LinInlinks, NumberVersions, SPanVersions, TimePointDivergence)
• Added memcached/membase support for global data sharing between all collections.
• Added the code of the spellchecker(PwaSpellchecker)

Changes:

• A wrong date was being displayed for search results.
• Fixed cases where specific types of queries changed the language of the UI.
• Fixed several issues where the UI language reverted to the default language.
• Fix XHTML validation errors on several pages.
• Lucene and Nutch similarity function normalized with an exponential decay.
• Ranking functions updated in configuration file.
• Fix integer concatenation of floats division on the scorer.
• Several bugfixes on ranking features and algorithms.
• Caches for URL version and flags of new version were removed.
• Change spellchecker to use Hunspell instead of Aspell.
• Changes and updated in the footer and sponsors.
• Force the request encoding before it is read/used.
• Deprecating taglibs-i18n on the OpenSearch JSP in favor of taglibs-fmt.
• Fixed missing dependencies.
• I18n updates and fixes.
• Several UI fixes.
• Some code refactoring.
• other bugfixes .

Removed:

• Removed unused code.
• Remove non-working wayback and spellchecker configurations.
• Removed lucene cache. It has not been used.
• Removed old parameter "search.max.hits" in favor of "max.fulltext.matches.ranked".

2011.12.06 ([r207])

Changes:

• Fixed several problems in the opensearch response that were not respecting

the specification.

• PWA xsd, for opensearch, changed.

2011.10.20 ([r199])

Changes:

• Switch old MCTES logo for the temporary logo of the MEC(Portuguese Ministry of

Science, Technology and Higher Education).

2011.10.13 ([r192])

Changes:

• Fix unescaped ampersand and wrong MIME-type for the OpenSearch RSS

2011.10.11 ([r191])

New:

• Added OpenSearch auto-discovery by browsers ( issue 1 ).
• Added XML Schema for the PWA namespace.

Changes:

• Updated OpenSearch code to spec 1.1.
• Added OpenSearch Time extension namespace.
• Added RFC3339 to opensearch time extension.
• Remove old and commented code of the opensearch service.
• OpenSearch bug fixes.
• Read global parameters for the ranking algorithms instead of those being hard-coded
• Update the OpenSearch description to allow the OpenSearch time extension for RSS+XML ( issue 2 ).

2011.7.29 ([r151])

New:

• Import of the Portuguese Web Archive code
• Added resource files and management scripts

Change:

• Added info to the project configuration files
• Fix an overwrite with old configuration files and compiling the project.

Removed:

• Deprecated classes and libraries