http://www.cs.odu.edu/~extract/developersWiki/ 2011-07-14T16:18:25-04:00 http://www.cs.odu.edu/~extract/developersWiki/ http://www.cs.odu.edu/~extract/developersWiki/lib/images/favicon.ico text/html 2011-07-13T09:46:04-04:00 zeil http://www.cs.odu.edu/~extract/developersWiki/doku.php?id=extract:devguide:developers_guide&rev=1310564764&do=diff Basic information for developers on the support tools & procedures for the Extract project. Updated versions of this manual can be found at <http://www.cs.odu.edu/~extract/developersWiki/doku.php?id=extract:manuals> For developers working with this code, the overall plan is: text/html 2010-09-06T14:05:36-04:00 zeil http://www.cs.odu.edu/~extract/developersWiki/doku.php?id=extract:v4.11&rev=1283796336&do=diff September 6, 2010 * Improved ability to distinguish personal names from city & institutional names * Clicking on any entry in the “Validation” column will bring up the metadata editor instead of the previous “tree” view of the metadata XML. This forms a more convenient viewer for the metadata output. Note that, although it is possible to make changes in the output and to save those changes, the editor has yet not been integrated into the overall data flow process. Therefore the .cit/.txt f… text/html 2010-09-06T13:25:03-04:00 zeil http://www.cs.odu.edu/~extract/developersWiki/doku.php?id=extract:reporting_problems_with_specific_documents&rev=1283793903&do=diff Recommended procedure: * Make sure that Luratech is still running. Open the Luratech batch control window, and make sure that the OCR batch process is not shown as “stopped” or “paused”. * Remove all files from the OCR in, OCR out, extractInputs, and extractOutputs/Debug folders. * Start the Extract program, but before loading in any files, ... * In the Archive menu, turn on Debug Mode. * Load the PDF document(s) believed to cause the problem. * If the problem recurs, please send u… text/html 2010-09-06T11:51:11-04:00 zeil http://www.cs.odu.edu/~extract/developersWiki/doku.php?id=extract:software_releases&rev=1283788271&do=diff The development team rebuilds the system on a regular basis, with seldom more than a week elapsing between new internal releases. Internal releases with major changes in functionality are provided to DTIC staff via a virtual PC, extractxp3.seven.research.odu.edu. Contact the ODU Extract project leaders for information about accounts for access to this machine. text/html 2010-07-06T10:05:26-04:00 zeil http://www.cs.odu.edu/~extract/developersWiki/doku.php?id=extract:third-party_libraries_dependencies&rev=1278425126&do=diff Version 4.1, June 30 2010 (Download this report as a single file here.) Third party, open source libraries used by the Extract program are managed via the Apache Ivy dependency manager. We have configured our project to load from two sources: the Ivy/Maven repository at ibiblio and our own local repository. text/html 2010-06-30T21:39:13-04:00 zeil http://www.cs.odu.edu/~extract/developersWiki/doku.php?id=extract:v4.1releasedtic&rev=1277948353&do=diff June 30, 2010 * cit/txt file generation for DTIC internal use is now supported as a “demo”, though uncertainty about the format specification for the .cit files makes this untrustworthy as yet. * Various template additions and changes * More stable GUI for the form and non-form template editors * The metadata editor will no longer pop up when examining untrusted metadata. This behavior may be restored in future releases. * Significant improvements to grouping of characters into words… text/html 2010-06-30T15:55:43-04:00 zeil http://www.cs.odu.edu/~extract/developersWiki/doku.php?id=dtic:extract_program_customization_for_the_dtic_collection&rev=1277927743&do=diff June 29, 2010 Most of the code for the Extract program can be used with any document collection. The process of customizing Extract for use with a specific collection consists of * Writing form and non-form templates for document layouts in the collection * Providing code to standardize metadata output formats (post-processing) * Defining the validation rules to be used in evaluating the output and providing an appropriate statistical model * Providing for processing of output metad…