Testing Indexing Automation

Make sure to test your configuration thoroughly for different scenarios. Make sure to change split page counts when using the page count split method. Or try testing with different possible split values when using the break code method.

Scan in and keep a set of 2 or 3 multi-page TIFF files that represent different scenarios for testing. This will save you from having to scan documents each time. You can just drag and drop the documents onto the browser to test. 

The SPLITTING piece is the one that needs the most testing. The OCR does its job behind the scenes and there is nothing to configure.

Here is a typical testing plan:

1. Make sure all work items are set to include you during the testing. That way you won't have to log out and back in for each different step type.
2. Clean out all testing data to make sure you can tell which is the new test
3. Drag and drop a multi-page TIFF onto the main tool bar and drop it onto the "bucket" that represents the Record Type you created from previous steps.
4. Leave the default values that were set during configuration
5. Refresh your workflow every 30 seconds to see when the record is ready splitting.
6. Once available, click the work item so the split page loads 
7. Verify that the automatic split worked. If not proceed to problem list below.
8. Finish the split then load up the indexing screen via the work item(s) that will be created from the splitting
9. Finish then indexing then load up the classification screen via the work item(s) that will be created from the indexing
10. Make sure the proper Record Type is selected based on the @RECORDTYPE field value used in the configuration. If not proceed to problems section.
11. Clean out testing data when done

Possible Problems

Although there may be any number of issues (typos, triggers out of order, etc.) this section contains some of the more common problems that arise.

Problem: The Record Type that was set up for indexing does not show up as a "bucket" in the drag and drop operation

Resolution: Make sure your test account has rights to Read/Write in that record type. If not it won't show up.

Problem: The work item never shows up after dropping the documents into the indexing "bucket"

Resolution: Make sure your test account is assigned to the proper trigger for that step.

Problem: The split process is deleting pages when it should not

Resolution: Make sure you have a field named @Split_KeepBreakPage (use any header value you want) and make sure it is set to FIRST or LAST. It is usually a missing field or an incorrectly typed field name.

Problem: The split process is not always finding the break value

Resolution: Make sure break value specified in the @Split_BreakCode field actually exists on each document (not necessarily each page - each document). You can also bring up the page that is not breaking correctly in the viewer and use CTRL-SHIFT-O to view the OCR data., This will show you the actual data that came from OCR and Bar Codes that the system is using. If you see that your values are commonly being misread by OCR try to find another key value Either that or instruct the users to perform a manual split on any that were missed. This is why the split information is shown to users before proceeding to indexing - to catch these types of issues.

Problem: The incorrect Record Type is being select by default in the Indexing page

Resolution: Make sure you have a field named @RECORDTYPE (use any header value you want) and that the value in the field is the same as the name of the Record Type you wish this document to be indexed into. If they don't match exactly then this will not work properly.

Congratulations! At this point you should have a working system that is processing incoming documents! 

Next Step: Using in Production