Problems Resolved in this Fix Pack

1776996 – Edit Training Document Window Causes Errors. The Edit Training Document window is now displayed as expected each time the window is opened.

Problems Resolved in previous Fix Packs

1699663 – Patch code not correctly recognized. By default, Patch codes should be recognized with the default FineReader recognition settings. However, if FineReader fails to recognize a Patch bar code using the default settings, you can improve the chances of bar code recognition by making use of the UseSpeedRecognition behavior via scripting. If you want to recognize Patch bar codes, you can add the following code to the BeforeLocate event.

1671463 – Perform check in Document Review for pending changes before closing a batch. It is no longer possible to close a batch in Document Review without a prompted to save any unconfirmed changes.

1662118 – OCR results decreased from FineReader engine 11.1 to 12.2 in KTM 6.3. It is now possible to use the FineReader “ProhibitModelAnalysis” property in script to improve OCR Results in some cases. This option is available via script only and does not appear on the user interface. The value for this option is set to True by default. When this option is set to False, the recognition engine uses model analysis to improve recognition in some cases.

Note: This property can be saved in the project. This means that the script does not need to be executed each time a project is processed. Execute the script from Project Builder and then save the project. The setting is then set when the project is used in production. Use the following script sample to set this option, modifying it for your specific needs.

1659361 – Retain Open Batch Windows Position for Dual Monitor Display. When using dual monitors, it is now possible to save the location of the Open Batch window. The next time you open a user module, the window appears in the same location.
1644370 – ValidationDocking3.xml can get into blocked state after starting many instances of Validation quickly one after the other. It is now possible to successfully open several back-to-back instances of the Validation module from the Batch Manager.
1640962 – Mixed Print Recognition Profile Crashes Project Builder and KTM Server On some Images. Project Builder and Server no longer crash when processing a project that uses a Mixed Print Recognition profile.
1635715 – Getting object reference error while synchronizing the project in KC having unmapped fields. An error is no longer encountered when a project is synchronized in Tungsten Capture.

1617050 – Folder fields are not populated to Tungsten Capture and are not preserved after suspend. Folder fields that have been changed by script in batch editing events are now successfully preserved and passed to Tungsten Capture.
1617420 – Several registry settings like AutoNextBatchOpen, ConfirmBatchClosing, Pollintervall, PollToOpenBatch and ShowOnlylnvalidDocs of for Validation are stored in the registry under “ProjectBuilder”, not “Validation”. The affected registry settings are now listed under Validation rather than Project Builder.

1619033 – Server error: “SeparatePages failed, StructuredException. An error is no longer encountered when a user separates a document.
1604360 – The Verify flag in Tungsten Capture Index Fields changes from True to False whenever a project is synchronized and the corresponding transformation project field does not require synchronization. The Verify flag only changes to True if the corresponding transformation project field requires verification.
1604355 – Layout Classifier “Enable automatic rotation” not recognized when used with TDS (Trainable Document Separation). The “Enable automatic rotation” setting for the Layout Classifier is now successfully executed as needed, when using Trainable Document Separation.

1588378 – Specific PDF document throws error message in Project Builder during extraction when using a barcode locator. An error is no longer encountered for larger PDFs that use a Barcode Locator during extraction.
1588101 – file.delete(folderpath) operation results with error Could not open Folder.xfd. It is now possuble to use the file.delete(folderpath) operation without encountering an error.

1566817 – Batch in dialog not opened after refresh. When a user refreshes the list of batches in the Open Batch window and then double-clicks on one of the batches, that batch opens successfully without error.
1564878 – Install from 6.3.1.9 breaks the access KSMS Standalone Server from within KTM Project. Accessing a KSMS standalone server from a project no longer causes an error.
1562713 – Enhance classification locator to provide more details. The full classification results are now attached to each alternative of the Classification Locator. If the Classification Mode setting is set to the Complete document (hierarchical) value, this information is available to scripting for further analysis as needed.

1546243 – In Validation Auto foldering with merge of documents results in index error. Merging documents using auto foldering in the Validation module no longer causes an error.
1561034 – ‘Open Batch’ operation failed with message ‘only one batch may be open at a time’ in Document Review module. An error is no longer encountered when a user opens a batch.

1553366 – Enhance classification locator to provide more details. The full classification results are now attached to each alternative of the Classification Locator. This information is available to scripting for further analysis as needed.
1532329 – Vendor Locator incorrect database vs fields mapping. An error in a Vendor Locator no longer occurs when the Vendor Name 2 predefined field is mapped to a database column.
1530665 – Dateformatter fails with month dictionary and dates in YYYY-DD-MM format. The YYYY-DD-MM date format is now supported.
1524087 – Filtering feature within the Open Batch method in KTM Document Review the “values” button does not populate. When filtering by Batch Field, a list of values is now successfully generated for Document Review, Correction, and Verification.

1494664 – Wrong GUID when copying OMR profiles. The GUID is no longer copied to the destination OMR profile when cloning.

1466122 – NLP does not work with Japanese documents. The NLP engine now successfully extracts Japanese named entities successfully.
1460797 – When processing batches through Tungsten Transformation Modules the documents are not being classified correctly because they are being locked. Due to a new retry mechanism, documents that are temporarily locked are now successfully processed once that lock is released.

1456194 – Auto Folder feature no longer initializing fields on new folders. When used, the Auto Folder feature now initializes fields on new folders as expected.
1444843 – When users attempt to use the Filtering feature within the Open Batch window, the “values” button does not populate with any data for users to select from when a ‘Batch Field’ item is selected. Filtering the Open Batch window now behaves as expected.
1402724 – Validation Fuzzy Database Lookup sorting issue. When using a database that has one or more columns that are not imported in the fuzzy index, the column sorting of of the search results now works as expected.

1441234 – Barcode value is empty once checksum option is enabled. By default, Code39 bar codes that contains asterisks should be recognized with the default FineReader recognition settings. However, if FineReader fails to recognize a Code29 bar code. Using the default settings, you can use scripting to improve the chances of bar code recognition. This is possible because the IsCode39WithOutAsterisk behavior is now available for editing via script. If you want to recognize code 39 bar codes with asterisks when checksum is enabled, you can add the following code to the BeforeLocate event.

1404073 – Calling CopyDocument() between MergeDocument() calls causes Out Of Range Exception in SynchronizeXRootFolderToKCBatch. When batch restructuring, the combination of merging and copying documents now performs as expected.
1398810 – Document routing fails in Validation as PDF file is still locked by another process. A PDF can be successfully routed in Validation without causing an error.
1379752 – “Object reference not set to an instance of an object” on opening a batch in Validation – get_IsViewerDocked. This error no longer occurs because the Docking XML is now protected from corruption.
1342366 – Bad performance with Bar Code Locator using FineReader when extracting QR code on color image. By default, QR bar codes should be recognized with the default FineReader recognition settings. However, if FineReader fails to recognize a QR bar code using the default settings, you can use scripting to improve the chances of bar code recognition. This is possible because the AdvancedExtractionMode behavior is now available for editing via script. If you want to recognize QR bar codes, you can add the following code to the BeforeLocate event.

1399311 – The Document_Validated is not thrown anymore if Fixpack 16 is installed and reclassify is used. If using Reclassify in AfterFieldConfirmed the Document_Validated event is now successfully raised again when a field is confirmed.
1399142 – TDS Model not working after KTM 6.3 upgrade. Trainable document separation is now successfully classifying and separating documents without issue.
1397124 – If a field is confirmed with holding down the enter key and the next batch is opened automatically then an Object reference error occurs. It is now possible to confirm a field by holding down the Enter key and when the next batch is opened automatically, there is no error.
1375245 – Field Count data missing from “Field Recognition Accuracy Per Day” report. Field Count data is no longer missing from the “Field Recognition Accuracy Per Day” report.
1381523 – Getting error during Validation when trying to confirm a field “… DocumentLoaded: Reclassify is not permitted in the current script function…”. It is now possible to use the Reclassify call in script without this error.
1358251 – Table Evaluation not possible in KTM. Table Evaluation is now performed successfully.
1378131 – Debug DLL Fixes Validation Module “Object Reference” Errors. An error no longer occurs in the Validation Module.
1375723 – Child set reread setting is ignored and parent class setting is used. When a child class has a different locator reread setting than its parent class, the parent reread settings are no longer executed during extraction.
1358297 – Paste grayed out after doing Copy (CTRL+V) once KTM Validation, so you cannot continue doing copy/paste. It is now possible to copy and then paste using the Ctrl-based keyboard shortcuts or the clipboard, as many times as necessary.
1370292, 1352443 – Object reference error is shown when confirming the field that reclassifies the document in script. An error no longer occurs when a document is reclassified in script using the AfterFieldConfirmed event.
1369734 – Datamatrix Barcode is not recognized in KTM anymore. By default, Datamatrix bar codes should be recognized with the default FineReader recognition settings. However, if FineReader fails to recognize a datamatrix bar code using the default settings, you can use scripting to improve the chances of bar code recognition. This is possible because the DetectInvertedZones behavior is now available for editing via script. If you want to recognize Datamatrix bar codes, you can add the following code to the BeforeLocate event.

1353640 – PDF file shows rendered without font (squares instead font) in Project Builder and Validation. Squares instead of the built-in PDF font are no longer displayed when rendering a PDF file.
1351693, 1349336 – Delete button grayed out in Validation when selecting multiple images. The Delete button is no longer greyed out in Validation when selecting multiple images.
1343422 – Inherited Advanced Evaluators Cause Crash in Project Builder. Project Builder no longer crashes when a Script Locator is defined before an Advanced Evaluator in the locator hierarchy of an inherited class.
1330287, 1327445 – Document routing fails with PDF document if Batch.CopyDocument is used. It is now possible to successfully route PDF documents.
1329796, 1328209 – Server fails batches: Error in function ProcessBatch: Conversion from string “” to type ‘Integer’ is not valid. An error no longer occurs when an empty or non-existent custom storage string is encountered.
1328691 – Latency with Inbuilt Database script, delay in Validation module. In order to avoid this latency, a project Administrator needs to provide the database connection type to ODBC window in script. A new property called ‘ConnectionType’ is now available.Here are the possible options for this new property.
– EnumConnectionType_Unknown
– EnumConnectionType_Odbc
– EnumConnectionType_OdbcSql
– EnumConnectionType_OdbcOracle
– EnumConnectionType_SqlServer
When you provide the specific ‘ConnectionType,’ this bypasses any additional calls to the database that check the database schema. Sample Script:

‍
1321559, 1320959 – NLP: All entities have confidence 1. The confidence of an alternative is now the average confidence of all of the located mentions.
1321411, 1320885 – NLP: Locator is throwing error “Zone index out of range”. It is now possible to draw a region on a page and then test or use that region without it causing an error.
1317047, 1317053 – Named Entity Locator does not highlight the best mention. The best mention is now highlighted when viewing the Named Entity Locator results.
1317027, 1317030 – Named Entity Locator loses highlighting when you click TEST again. When testing the Named Entity Locator, all locator alternatives are displayed in the Document Viewer each time the Test button is pressed on the Test Results tab.
1317021, 1317033 – Missing entities in Named Entity Locator. When extraction is performed for the Named Entity Locator, a complete list of entities is now extracted as expected.
1317020, 1317037 – Missing scrollbar in Named Entity Locator. A scrollbar is now displayed when needed.
1315675, 1246992 – There are two processing windows coming for Training Extraction. If there are no documents in a document set, the progress bar window is no longer displayed if Extraction training is performed for that document set.
1313696, 1320629 – Failed to extract entities with an error no language is specified. The Test button on the locator is disabled if no language is configured at the class-level.
1310150, 1305677 – Validation crashing on reclassifying via script. Validation no longer crashes when a user reclassifies via script and then closes the module.
1280195 – Table Locator Extracting Lines Past End of Table Dictionary. The Table Locator now ignores lines extracted after the end of the table is found.
1266814, 1254229 – Name of representation is not changed after overwriting the recognition results. The name of a representation is now updated correctly when OCR is performed on a document that already has extraction results.