Documenti di Didattica
Documenti di Professioni
Documenti di Cultura
SAP® Solutions
Customizing Guide
®
OpenText™ Invoice Capture Center for SAP Solutions
Customizing Guide
Rev.: V7.5 SP1 (2015-01-27)
Open Text SA
40 Avenue Monterey, Luxembourg, Luxembourg L-2163
Tel: 35 2 264566 1
Open Text Corporation
275 Tompa Drive, Waterloo, Ontario, Canada, N2L 0A1
Tel: +1-519-888-7111
Toll-Free Canada/USA: +1-800-499-6544, International: +800-4996-5440
Fax: +1-519-888-0677
Support: http://support.opentext.com
For more information, visit: http://www.opentext.com
Copyright 2015 Open Text SA and/or Open Text ULC. All Rights Reserved.
Open Text is a trademark or registered trademark of Open Text SA and/or Open Text ULC. The list of trademarks is not exhaustive of
other trademarks, registered trademarks, product names, company names, brands and service names mentioned herein are property of
Open Text SA and/or Open Text ULC or other respective owners.
RecoStar is a registered trademark of Open Text Document Technologies GmbH
ACKNOWLEDGEMENTS
Portions copyright 2003-2014 AllMyPapers.
Portions copyright 2014 Artifex Software Inc.
This software is based in part on the work of the Independent JPEG group.
Portions copyright 2001 URW++. All rights reserved.
This work uses Boost C++ libraries.
SAP®, R/3® and SAP CRM® are registered trademarks of SAP AG.
Microsoft®, Microsoft SQL®, and Microsoft SharePoint® are either registered trademarks or trademarks of Microsoft Corporation in
the United States and/or other countries.
Artifex®, the Artifex logo, Ghostscript®, and the Ghostscript logo are registered trademarks of Artifex Software Inc.
PostScript® is a trademark of Adobe Systems Incorporated.
PCL® is a trademark of Hewlett Packard Company.
All other products or company names are used for identification purposes only, and are trademarks of their respective owners.
Disclaimer
No Warranties and Limitation of Liability
Every effort has been made to ensure the accuracy of the features and techniques presented in this publication. However, Open Text
Corporation and its affiliates accept no responsibility and offer no warranty whether expressed or implied, for the accuracy of this
publication.
Contents
1 Introduction 1
1.1 Functional Overview ................................................................................................... 1
1.2 Components of Invoice Capture Center ...................................................................... 3
1.3 Customizing................................................................................................................. 4
1.4 Hot Spots ..................................................................................................................... 5
1.5 Achieving Optimum Recognition Results ................................................................... 6
2 Basic Customizing 7
2.1 Prerequisites for ICC Interaction with VIM ................................................................ 7
2.2 Creating a Standard Use Case Application .................................................................. 8
2.2.1 Specify Name and Application Type .............................................................. 9
2.2.2 Specify Country .............................................................................................. 9
2.2.3 Add Archive Document Types ...................................................................... 10
2.2.4 Specify Company Code Detection ................................................................ 10
2.2.5 Specify Connection to SAP ERP .................................................................. 12
2.2.6 Generating the Application ........................................................................... 14
2.3 Configure and Start Your ICC Application for VIM ................................................. 15
2.4 Enable Validation - Export Application to SAP ERP ................................................ 18
2.5 Test Your Application ............................................................................................... 19
2.6 Changing Your Application....................................................................................... 19
3 Use Cases 21
3.1 More than One Receiving Country ............................................................................ 21
3.2 More than One Sending Country ............................................................................... 23
3.3 More than One Application ....................................................................................... 23
3.4 More than One SAP ERP System Involved .............................................................. 23
3.5 More than One Company Code ................................................................................. 24
3.6 More Than One Tax Rate and VAT Amount ............................................................ 24
3.7 High Recognition Capacity Needed .......................................................................... 24
3.8 More than One Validation Client .............................................................................. 25
3.9 Vendor IDs are Identical throughout Company Codes .............................................. 25
3.10 Vendors IDs are not Identical throughout Company Codes ...................................... 25
3.11 Shared Service Center ............................................................................................... 25
3.12 Intercompany Invoices .............................................................................................. 26
3.13 Custom Field Needed ................................................................................................ 26
3.14 Export of Custom Fields to VIM ............................................................................... 26
3.15 Not Purchase-Order-Related Invoices ....................................................................... 26
3.16 Non-Standard Syntax/Length of PO Numbers .......................................................... 26
3.17 Processing Invoices with Appendices ....................................................................... 28
3.18 Application for New Country .................................................................................... 28
3.19 Applications for China (Traditional), Japan, Korea, Thailand .................................. 29
3.20 Presetting ICC Fields ................................................................................................. 30
3.21 Processing Invoice-Related Documents .................................................................... 31
3.22 Processing Down Payment Invoices .......................................................................... 31
10 Index 199
Invoice Capture Center Recognition extracts the essential invoice data (date, amount, order number etc.) from the
scanned invoices in the archive. ICC Recognition uses a knowledge base that has been trained on samples of thousands
of invoices from multiple countries around the world. In addition, it uses fuzzy database matching: ICC Recognition
maps the read data – such as names, addresses, vendor numbers, purchase order numbers etc. – to SAP ERP vendor
master data and SAP ERP purchase order data during analysis.
The extracted information is passed on to Vendor Invoice Management.
Besides the facility to save an invoice application in SAP ERP (export), the Customizing Client also provides the facility
to load an invoice application from SAP ERP to the Recognition Server (import). Using the SAP ERP export and import
facilities the Customizing Client user can copy an invoice application from a development system to a test system or
from a test system to a production system.
To create an application, click New Application in the Invoice Capture Center application menu:
Note: When the application has been created, the county settings can be found in the Settings dialog box where
they are shown at the Receiving Countries subnode of the General node.
Note: For applications created with ICC 5.2 Service Pack 4 or earlier, the application name will be shown in the
field Field Description in SAP ERP, whereas JobClass<n> will be displayed as application name where
<n> is a number between one and ten.
Click Next to continue.
Add all Archive document types that should be processed by the new application. The archive document types have to be
the same as in VIM! If you test without VIM you can type in a dummy archive document type.
Note: You can process several archive document types in one application or you can process each archive
document type in a separate application.
Click Next to continue.
If you want to use more than one company code, select Use a company code per archive document type.
Enter the company code and the SAP ERP system (technical name).
When you confirm with OK, the specified company code and SAP ERP system name are shown:
To test ICC without VIM you can continue with Next without entering SAP ERP system data. Continue with
"Generating the Application" on page 14.
As soon as the application has been created, the following page will be shown.
Click Close.
The new invoice application is stored in a shared folder on the Recognition Server and can immediately be used to
process invoices for testing purposes on the Recognition Server and its Recognition Nodes. The new application is
opened automatically.
All Hot Spots are created with a default scheduling behavior. Each Hot Spot checks its input channel regularly for new
data according to its scheduling configuration. For each Hot Spot you can configure separately, whether it should check
its input channel once a day at a specified time, or regularly after waiting a specified interval.
To show the configuration of a Hot Spot, select the Hot Spot and click Properties.
It is important that you add all relevant pairs of SAP ERP system (technical name) and company codes. Otherwise the
Hot Spot will not retrieve all needed data, and the recognition rate will be diminished.
Please check whether all relevant archive document types have been added during application creation. If not, add them
at the Document types property.
After you have configured the Hot Spots, start those that are used by your application. Select a Hot Spot and click Start.
Tip: To run a Hot Spot immediately, for example to provide the download data for testing, start the Hot Spot, and
then click Run now.
For more information, see "Hot Spots (SAP Links)" on page 117.
Click OK.
If you use a different syntax or length for purchase order numbers, you have to change PO number settings. Mark Use
Regular Expression, and then click Edit in order to enter a regular expression:
For more information about regular expressions, see "Using Regular Expression" on page 67.
The official Microsoft homepage for regular expressions is: http://msdn.microsoft.com/en-us/library/hs600312.aspx.
As long as the Use Regular Expression check box is selected, the standard format and previously specified number
ranges are not checked. Therefore the number ranges are not shown and you can’t add or remove a number range in this
case.
In the example code, fixed strings are assigned for the two fields. Alternatively, the application could use barcodes. In
this case you would use add a custom field of type BarCode and use scripting to extract system and company code from
the result of this field.
The fields must not be empty. Therefore take care to provide a default value if necessary. Also take care to assign a value
to both fields; setting only one of the fields will not take effect.
The value of the CompanyCode field must comprise four digits. The value of the System field is a string.
The following example code shows the scripting code needed to preset a value for the field InvoiceVendorNumber:
The example code assigns a fixed value. In a real application, a value would be supplied in a variable by preceding code.
Take notice that a vendor ID supplied by scripting will still be checked against downloaded vendor master data. If the
supplied vendor ID is not present in the master data, it will not take effect and the InvoiceVendorNumber field will
remain empty.
The following example shows scripting code that could be used to preset a value for the field ListPO:
The example code assigns the value of a field PreListPO, which could have been extracted using a custom
automation method. The field in the document PresetInvoice must have the name PreListPO; it may contain
additional PO’s as alternatives. If the field PreListPO exists in the document PresetInvoice, its content is
used in ICC exclusively. This is also true if PreListPO is empty.
The following example shows scripting code that could be used to preset a value for the field ListDN:
The example code assigns the value of a field PreListDN, which could have been extracted using a custom
automation method. The field in the document PresetInvoice must have the name PreListDN; it may contain
additional DN’s as alternatives. If the field PreListDN exists in the document PresetInvoice, its content is
used in ICC exclusively. This is also true if PreListDN is empty.
Note: For download you should use a SAP ERP system with current vendor and purchase order data in order to
achieve good recognition results on matching invoices.
4. Click the ellipsis button at the Index Column list. The Manage Lookup List dialog box opens. Select the
Show internal tables check box.
5. Select the list entry Vendors, and then click Reload. In the appearing file dialog box select the file that you
want to load (Lieferanten.txt).
4. Click the ellipsis button at the Index Column list. The Manage Lookup List dialog box opens. Select the
Show internal tables check box.
5. Select the list entry Positions, and then click Reload. In the appearing file dialog box select the file that you
want to load (PO_Header.txt).
6. Wait until a dialog box indicates that import has been finished. Close the dialog box with OK.
You can now check the recognition results. Fields that could be extracted correctly and empty fields that are
optional are highlighted with green color; empty mandatory fields are highlighted with red color.
4. Click Submit or Cancel to close the validation view. The recognition results will be deleted.
6.1.1 Structure
In the Customizing Client you can switch between two views showing special aspects of your Invoice Capture Center
system, and you can open several tools needed for application development.
The views and the tools are accessible in the ribbon of the Customizing Client: There are buttons to open views, and
buttons to start development tools.
If you click a development tool button, the tool is opened in a new window. If you click a view button, the view is
opened within the Customizing Client window; the according view button in the ribbon stays pressed to indicate which
view is currently open in the Customizing Client.
The view buttons and tool buttons are provided on different tabs of the ribbon. View buttons and tool buttons that belong
together in a way are presented on the same tab. A tab of the ribbon corresponds to a scope of the ICC system.
Selecting a different tab in the ribbon will automatically change the view.
During application development you will work mainly with the Application view.
Note that the Application tab does not provide view buttons. There is only one view for the Application tab. Selecting
the Application tab will automatically open the respective view.
The General tab provides monitoring facilities to the Customizing Client user. On the General tab, you can switch
between several technical monitoring views: HotSpot view, Jobs view, and Cluster view. In addition, the General
The views and tools of the General tab are needed for the technical configuration and for monitoring of an Invoice
Capture Center production system. You might need these facilities during testing of your profiles.
All tools of the Application tab work on the currently opened application whereas the General tab permits to show
global data of the Recognition Server – independently of the opened application.
The tools and views of the General tab and the Validation tools of the Application tab are not discussed in detail in
this guide. For more information about monitoring and technical configuration, see Invoice Capture Center -
Administrator`s Guide; for more information about Validation, see Invoice Capture Center - User`s Guide. This guide
concentrates on application development.
New Application
This command is used to create a new application. An assistant starts. Follow the instructions of the assistant
and then click Create. Application creation will take some seconds. When the application has been created,
close the assistant by confirming with Close. For more information, see "Creating an Application" on page
47.
Note: Normalization has been modified incompatibly with ICC 6.0. If you open an application created with ICC
5.2 that uses scripting or custom normalization, you will be asked whether the application should be migrated. If
you answer with No, the application will not be upgraded and the documentation of ICC 5.2 is still valid with
regard to normalization for this application. If you answer with Yes, you will have to adapt your scripting code.
Applications that use neither scripting nor custom normalization will be always be upgraded.
For countries of the second list the application is preconfigured with regard to character recognition, culture settings
controlling data conversion and formatting, and VAT rates are specified. For some of these countries, data extraction will
also yield good results if the respective country uses the same keywords and phrases as one of the countries of the first
list. For some countries, additional customizing will be needed to achieve high recognition rates. For more information,
see "Application for New Country" on page 28. The second country list contains the following countries:
For applications of type Custom documents, the country setting only affects settings for character recognition and
cultures. For this application type, there is no difference between countries of the first and of the second country list.
Select the required countries and click Add until the list on the right side contains all required receiving countries.
On this page you specify the vendor languages, that is, the languages occurring on incoming documents, and
corresponding data formats. Select the required items in the list on the left side and click Add until the list on the right
side contains all needed languages.
Some countries have a special date format. For example in the United States a date starts with the month whereas in
other countries with language English, the date starts with the day. Therefore some items in the list show a country in
parentheses.
Due to possible ambiguities, order in which formats are checked can be important. Therefore you can change the order of
the items in the list on the right side. To move an item, select it and then click Move up or Move down.
When the language list is correct, click Next to proceed to the Archive document types page.
Note: You can modify the archive document types later using the Settings dialog box (for more information, see
"Configuring an Application" on page 59).
Tip: If you process invoices for one company code only, choose Use a fixed company code.
For each company code detection method, the company code is connected with a SAP ERP target system. At runtime, if
Invoice Capture Center has detected the company code of an invoice, it also knows the SAP ERP target system for the
invoice. When Invoice Capture Center exports the recognition results to Vendor Invoice Management, it also exports the
found target SAP ERP system. VIM will manage the transfer to the respective SAP ERP system. Thus, Invoice Capture
Center supports SAP ERP multi-backend systems.
Click Next to proceed to the next page.
Note: You can modify company code detection later using the Settings dialog box (for more information, see
"Configuring an Application" on page 59). Take notice that the company code field is shown in Validation only if
automatic company code detection has been specified. This setting can’t be modified, that is, selecting or clearing
the Visible check box on the Presentation tab in the Settings dialog box will not take effect for the company
code field.
Click Check login to test whether you have specified all needed values correctly.
Tip: If a saplogon.ini file is present, you can select a suitable SAP ERP system in the System list of the
Connection to SAP page before you click Advanced. The SAP Connection dialog box will then show the
parameters of the selected SAP ERP system, and you can modify values instead of typing everything in.
When you have entered all values correctly, type a suitable name in the System name box and then click Apply to
close the dialog box. The new SAP ERP connection will now appear in the System list. Take notice that the new
connection is used for the Hot Spot connections of the current application but that its data are not stored separately. If
you switch to another connection and close the Customizing Client, the connection specified in the SAP Connection
dialog box will no longer appear in the System list afterward and the corresponding data will get lost.
Click Next to proceed to the next page.
Note: You can modify the customizing settings using the Settings dialog box (for more information, see
"Configuring an Application" on page 59).
Note: Export of an application always comprises the recipient data, which have been imported for automatic
company code detection (see "Specifying Company Code Detection" on page 72), and the phrase value files that
have been created for SnapMatch custom fields (for expert use only). The database containing vendor master data
and purchase order data downloaded from SAP ERP (see "Vendor Master Data and Purchase Order Data" on page
152), and the databases containing learning information for adaptive recognition fields and adaptive classification
fields (see "Training" on page 122) are NOT exported to SAP ERP.
Tip: By clearing the Settings and Download data check boxes in the Select Application dialog box, you can
also use this command to export only the learning data to a file.
The available check boxes depend on the data contained in the selected archive file, that is, a check box is available if the
corresponding check box was selected in the Select Application dialog box during export. If you want to import an
application and the Settings check box is not present, you have selected an unsuitable archive file; in this case
terminate import by clicking Cancel and try again with a suitable archive file.
If one of the applications on the Recognition Server has the same name as the application in the archive file, a message
will be shown. In this case you can either cancel the import, or overwrite the application on the Recognition Server with
the application from the archive file.
Tip: The Import Application command is also used to import learning data for an existing application. Either
you import from an archive that contains only learning data, or you clear the Settings and the Download data
check boxes if they are present. If you clear the Settings check box or if it is not present, you can select the target
application by clicking the desired application in the application list at the top of the dialog box.
Important
The application will be deleted locally on the Recognition Server. It won’t be deleted in SAP ERP. For
information about deleting an application in SAP ERP, see “Deleting an ICC application from SAP tables” in
OpenText Vendor Invoice Management for SAP Solutions - Administration Guide.
Note: The Settings button is only available if an application is open. The current application is shown in the title
bar of the Customizing Client. To open an application, click Open Application on the application menu, and
then click the appropriate application.
On the left side, the Settings dialog box shows all configurable items of the current application in a tree view. To
modify an item, you open the respective nodes and click the desired item. The right side of the dialog box will then show
the properties of the selected item.
At the uppermost level, four nodes are shown:
The General node permits to modify name and description of the application at its Name subnode.
The Configuration node permits to specify archive document types and to control training in the Validation
Client.
The Fields node permits to specify all settings controlling data extraction.
The Advanced node permits to enter scripting code, and to control which pages should be excluded from data
extraction.
Note: For applications created with ICC 5.2 Service Pack 4 or earlier, changing the application name in ICC will
only change the field description in SAP ERP, whereas the SAP ERP application name will still be
JobClass<n> (<n>=1..10).
Important
For Chinese applications, only the Recognition Engine and Document Extraction can be used. Don't try to use
Adaptive Recognition; it is not supported for Chinese applications.
The list on the right side shows all specified archive document types.
To add a type, click Add, type the name of an existing SAP ArchiveLink document type in the appearing dialog box and
confirm with OK.
To modify a defined type, select it and click Edit. You can then modify the name in the appearing dialog box. Be sure to
use the name of an existing ArchiveLink document type in SAP ERP.
To remove a type, select it and click Remove.
For information about configuring ArchiveLink document types in SAP ERP, see "Configuring Archive Link" in
OpenText Vendor Invoice Management for SAP Solutions- Configuration Guide.
Note: During application creation, three Hot Spots are automatically created and configured. The archive
document types entered during application creation are used for the configuration of the SAP Extraction Link Hot
Spot (for more information, see "Adapting the SAP Extraction Link" on page 119). The list of document types in
the application settings and the list of document types in the SAP Extraction Link configuration should always be
identical. If you change the document types list in the Settings dialog box, the document types list of the SAP
Extraction Links won’t be updated automatically. Therefore, after changing the archive document types
configuration of the application, be sure to adapt the configuration of the SAP Extraction Links (see "Configuring
Hot Spots" on page 117) .
To add an email address, click Add, type a complete email address (john.smith@company.com), or an entire
domain (@company.com) in the appearing dialog box, and confirm with OK.
To modify an entry, select it and click Edit. You can then modify the email address in the appearing dialog box.
To remove an entry, select it and click Remove.
To extract additional information contained in the line items, such as additional costs and discounts, select the Enable
line item qualifying and filtering check box.
By default, line items will be extracted. To switch off line item processing, clear the Extract line items check box.
If purchase order data will be downloaded, and the invoice line items should be matched against these data, select the
Check line items against downloaded data check box.
Note: PO line number, PO number, and delivery note number are not returned by the recognition step. Therefore
these fields will be empty in the invoice item data if the Check line items against downloaded data check
box is not selected.
Note: PO number, PO item number, and delivery note number (optional field, see "Activating Optional Fields" on
page 80) are only supplied by line item mapping; these items are not extracted from the invoice. Therefore these
fields will be empty if line item mapping is not enabled, or doesn’t succeed.
Note: Download comprises items with ITEMTYPE 3 (planned delivery costs), and ITEMTYPE 4 (goods receipt
for planned delivery costs). These items are not used at the moment.
If the sum of the recognized item amounts equals the recognized invoice amount and only a single invoice item could not
be mapped to a single remaining purchase order item, this item will be mapped without regard to the field values.
If the unit price and /or the amount of an invoice item on the invoice differ(s) slightly from the unit price and / or amount
values in the corresponding purchase order item in the downloaded data, line item mapping won’t find the corresponding
purchase order item in the downloaded data by default. To allow mapping despite of slightly different values, enter a
tolerance value under Set line item mapping tolerance. To specify an absolute tolerance, accept the default option
Absolute; otherwise select Percent. If the difference is in the specified tolerance range, line item mapping will find
the corresponding purchase order item.
By default, only PO numbers found on the invoice will be returned. If you select the If no PO number found,
determine PO number from DB check box, a PO number found in the database will be returned if the PO number
could not be found on the invoice.
To prevent high runtimes or erroneous results, line item mapping will not be performed if no PO number could be found
and more than 200 line item candidates were detected.
By default, item rows for additional costs and discounts will be returned, whereas found rows for the other amount types
in the Qualifier list will be removed from the InvoiceItems field result. ICC uses keywords and phrases to determine
the different kinds of expenses and discounts and searches them not only in the items table area but also below where
they may occur as a separate block on some invoices.
To optimize detection of expenses and discounts, you can add phrases, for example if you customize an application for a
new country, or you can remove phrases that cause problems on special invoices. To modify a phrase list, click the
ellipsis button in the Phrases column for the respective item. A Phrase dialog box opens where you can add, edit, or
remove phrases.
If found rows for additional qualifier types should also be returned, select the check box in the Deliver column for the
respective item.
You can even define additional application-specify qualifier types. Right-click into the table, click Add key word row,
type a name in the Qualifier column, and add phrases to the phrase list of the new item. If the detected items should be
removed from the result, clear the check box in the Deliver column.
The ExclusionPhrases list contains phrases that are similar to some item qualifiers but may occur in other types of
items lines. Therefore they must not be found to avoid confusion.
Caution
Don't use the Delete key word row command on the context menu on the predefined items; otherwise the
preconfigured phrase list will get lost.
When you create a new application, the configuration will already contain some typical purchase order number ranges
shown in the PO number ranges list. To add a range, click Add, type the first two digits of the purchase order
number range, and confirm with OK.
To modify an entry, select it and click Edit. You can then modify the first two digits in the appearing dialog box.
To remove an entry, select it and click Remove.
If your purchase order numbers have a different structure that cannot be determined by number ranges you can use a
regular expression to determine the purchase order numbers (see below).
Note: The PO number ranges (and the regular expression) only affect recognition. Their purpose is to prevent that
other numbers found on the invoices, such as phone numbers, are interpreted as PO numbers. If PO numbers are
checked against downloaded data, or during line item mapping, PO numbers may be returned that are contained in
downloaded data but are outside the specified number ranges (or don’t match the regular expression).
Note: This check box only affects the returned header PO number. To get PO numbers for line items, you need to
select the Check line items against downloaded data check box at the Extraction subnode of the
InvoiceItems field node.
Note: To check the purchase order numbers, not all entries of the downloaded purchase order data are used.
Invoice Capture Center only uses the entries whose VENDORID field value matches the found vendor number,
whose COMPANYCODE field value matches the found company code, and whose SYSTEM field value matches
the found SAP ERP target system.
Click Edit to enter the regular expression. The Enter Regular Expression dialog box opens.
In the Regular expression box, type the regular expression. To test your regular expression, type a test string in the
Sample input data box and click Test. If the test string matches the regular expression, the Result color box will
If the checkbox Use PO number for vendor ID detection is not selected, the vendor ID will be determined by
matching vendor data on the image against downloaded vendor master data. Good results can be expected if the invoices
bear bank data and if the vendor master data are sufficiently filled for ICC recognition.
If the checkbox Use PO number for vendor ID detection is selected, the downloaded purchase order data will be
used to determine the vendor ID. Invoice Capture Center selects the purchase order entries, whose company code field
value matches the current company code and whose system field value matches the current SAP ERP system, and
searches for these purchase order numbers on the invoice image. If a number is found on the image, Invoice Capture
Center reads the vendor ID in the according purchase order entry and puts the read value into the vendor ID field.
If no PO number can be found on the image (for example if the invoice is a non-PO invoice), the vendor ID is
determined by using the vendor master data download (see above).
The usage of downloaded PO data for vendor ID detection is useful if the invoices bear no bank data or if the
downloaded vendor master data are not sufficiently filled for ICC recognition purposes.
The check box Ignore company code and SAP system at vendor detection allows to find a vendor who does
not have a matching entry in the downloaded vendor master data for the current company code and SAP ERP system.
The list on the right side has an item for each currency supported by ICC; each currency is represented by its three-letter
ISO code. At the currencies that are preconfigured for the current application the Active check box is selected; these
currencies are shown on top.
To modify the set of valid currencies, select or clear Active check boxes as needed.
To show the supported notations, click the ellipsis button in the respective currency row. A Phrase dialog box opens
where you can add, modify, and delete notations as needed.
Tip: To specify the default currency, go to the InvoiceCurrency field, and click a currency in the Default value
list on the Data tab.
To add a new currency, right-click the item unit list, click Add translation line, type a name in the appearing new line,
click the ellipsis button to open the Phrase dialog and specify phrases for the new currency. Take notice that new
currencies will only bring about that the specified values will be accepted in the Validation Client; specifying additional
currencies will not affect Recognition.
Take notice that only selecting and clearing of the Active check box for standard currencies will affect Recognition.
Modification of notations and added new currencies will only affect manual input and Single Click Entry.
During application creation, the current VAT rates of the specified country will be added to the application configuration
automatically. Therefore you need to modify the VAT rates only if the VAT rates of the respective country change, or if
you want to process invoices from an additional country with different VAT rates.
To add a VAT rate, click Add, type the VAT rate in the appearing dialog box and confirm with OK. As VAT rate, you
can enter an integer, or a decimal with up to two decimal places and a dot as decimal separator. The value must be
between 0 and 100.
To modify a VAT rate, select the respective entry, click Edit, and type the new value in the appearing dialog box.
To remove a VAT rate, select the respective entry and click Remove.
With country setting Canada, the VAT rates box looks different and permits to enter valid tax rates of the different tax
types (Provincial Sales Taxes, Goods and Services Tax, and Harmonized Sales Tax):
Note: The VAT Rates node is not available for applications with country setting USA.
Four different methods are available to determine the company code. You can use a single, fixed company code for all
documents of the current application, a separate company code for each archive document type, or you can use automatic
company code detection either using only recipient data or also using the PO number.
To use a fixed company code, select the click Use a fixed company code in the Company code
detection list, enter the company code, and the name of the target SAP ERP system all documents shall be
transferred to.
To use different company codes for the different archive document types, click Use a company code per
archive document type in the Company code detection list. The dialog box will then show a list of all
archive document types you have specified when you created the application. Select the first archive document
type, and click Edit. In the appearing dialog box enter the company code and target SAP ERP system name for
the document type, and confirm with OK. Repeat this for all shown archive document types.
Note: If you process invoices for a single company code only, choose Use a fixed company code.
Note: Take notice that the company code field is shown in Validation only if automatic company code detection
has been specified. This setting can’t be modified, that is, selecting or clearing the Visible check box on the
Presentation tab of the Settings dialog box will not take effect for this field.
For each company code detection method, the company code is connected with a SAP ERP target system. At runtime, if
Invoice Capture Center has detected the company code of an invoice, it also knows the SAP ERP target system for the
invoice. When Invoice Capture Center exports the recognition results to Vendor Invoice Management, it also exports the
found target SAP ERP system. Thus Invoice Capture Center supports SAP ERP multi-backend systems.
To be able to use automatic company code detection, you need to prepare a database in text file format (see "Recipient
Data"), which contains the address data of the recipients together with their company codes and SAP ERP target
systems. If you select one of the two automatic company detection methods, an Import recipients button appears,
which permits to import the recipients data file.
To prepare the recipients data file, you can use transaction /OPT/VIM_COMP_DN and report
/OPT/VIM_COMPCODE_DOWNLOAD of Vendor Invoice Management which will create a file in CSV format
containing address data for all company codes. To optimize company code detection, insert additional lines with all
address variants that may occur on invoices for each company code manually before you import the file.
When you later export the application settings to a file or to SAP ERP, the recipient data will be included automatically
so that they will be present when you import an application on the production system.
The Phrase list tab contains phrases in different languages, which are used to detect a credit memo. To optimize
recognition you could remove phrases if invoices of the respective language will not occur at your application, or you
could add phrases of countries that are not supported by Invoice Capture Center.
On the Exclusion list tab you can add phrases that could result in an erroneous detection of a credit memo. For
example, the word Credit appears in the company names of several banks. If such a company name occurs on an
invoice, it would be erroneously classified as credit memo. This can be prevented by adding the respective company
names to the exclusion list.
Caution
Don't modify format specifications in the Formats tab of the Region and Language dialog box and the
Customize Format dialog box of the Windows Control Panel because this would affect checks and
formatting within Invoice Capture Center.
For each group, the dialog box shows a rectangle highlighted with the currently used color.
For each group, the color used for highlighting is a lighter version of a dark base color. Below the colored rectangle of
each group, a slider controls the color used for highlighting. If you move the slider to the left, the color becomes lighter
until it becomes white if you move the slider to the left end. If you move the slider to the right, the color becomes darker
until it equals the dark base color if you move the slider to the right end.
To select a different base color for a group, click the respective rectangle. The Color dialog box opens. In this dialog
box you can select a color from a set of basic colors or specify your own color.
The specified colors will be used in the Validation Client as long as the Validation user doesn't modify colors. If the
Validation user has modified colors and then clicks Restore Default in the Colorize Field Groups dialog box, the
application-specific colors again take effect.
Take notice that several fields can't be found on the topmost level. For example, to access fields of the amount group,
open the Amounts node and its Elements subnode.
Note: The Settings button is only available if an application is open. The current application is shown in the title
bar of the Customizing Client. To open an application, click Open Application on the application menu, and
then click the appropriate application.
Caution
It is recommended to change field settings of an application only in ways described in this guide. Changing
field settings without expert knowledge may deteriorate recognition results.
These fields are adaptive recognition fields. The Server and the Recognition Nodes must learn how to extract data for
these fields. The learning data for these fields have to be provided by training (see "Training of Adaptive Recognition
Fields" on page 122).
For Austria, Germany, and Switzerland, Supply Date is a standard field, which is read by a special automation method.
To make an optional field visible, expand the nodes Fields and Predefined Fields, and click the respective field to
show its properties.
On the Presentation tab, select the Visible check box:
In Validation (within the Customizing Client or Validation Client) the fields will be shown in the order defined by the
Tab order values.
Note: The invoice items table is always shown as last field without regard to its tab order value. Therefore it is not
meaningful to assign a lower tab order value to the InvoiceItems field.
In the same way you can change the column order for the invoice items table. Go to the InvoiceItems field in the field
list, expand its Elements subnode to show the columns, and modify the tab order values of the column subfields as
needed.
To improve the recognition result, you have to perform training. For more information, see "Training of Adaptive
Recognition Fields" on page 122.
If the limits are set to a negative value, only date values will be found and returned that are smaller (that is earlier) than
the date of processing.
If Lower limit is set to an absolute date, only date values will be found and returned that are greater (that is later) than
or equal to the specified date. If Upper limit is set to an absolute date, only date values will be returned that are smaller
(that is earlier) than or equal to the specified date.
The Lookup list column determines the assigned database column. Don't modify this property.
This property takes effect if Deliver OCR results or Deliver database entry checked by OCR is
selected. If you type a value in the Column confidence box, the column value will only be returned if the
value read from the invoice matched the database column value with a higher confidence. If you set it to 100,
the value will only be returned if the read value was identical to the column value. For some standard fields, this
property has the default value 100.
If Deliver OCR results is selected, the field will return the data that were read from the invoice for this
column if it could be found. If Deliver database entry is selected, the value of the respective database
column is returned. If Deliver database entry checked by OCR is selected, the database column value is
returned if the value could be found on the invoice; this will make sure that the value is returned in the correct
format for ICC/VIM. For some fields, Deliver database entry checked by OCR is selected by default,
whereas for most fields Deliver OCR results is selected.
If the Refresh from base field check box is selected, the field value will be updated if another database entry
is being selected by the Validation user. If the field value should not be updated during Validation, clear the
check box.
By default, the field will be positioned within a group together with the corresponding base field and its tab
order will be modified accordingly. If the additional field should be the last field in tab order, clear the Place
near base field check box. If you select the check box again later, modify the tab order to position the field as
desired.
4. In the Field name box, type the name that identifies the field.
5. Enter a default value if necessary; otherwise the field will be empty if it is not captured.
6. Select the Field cannot be empty check box if the field is mandatory. Validation will then indicate an error
if the user tries to submit the invoice while the field is still empty.
2. If the field should not be shown in the Validation Client, clear the Visible check box.
3. In the Display name box, type a title for the field that should be shown in the Validation Client.
Otherwise the field name will be used.
4. The field will appear as last field in the field list and by default it will also be the last field in the tab order.
If you want that the field should be activated earlier if the Validation user captures fields or uses the TAB
key to proceed through the fields, enter a value for Tab order that lies between that of the desired
predecessor and successor.
5. Select the Read only check box if the Validation user should not be able to modify the field value.
6. At the Control property, specify the appearance of the field. With the default TextBox, a simple text box
will be displayed where the user can capture text using the keyboard or the mouse:
If you select ComboBox, the dialog box will change, and will show an empty value list, and buttons Add
and Remove:
Click Add to add a value; a text box appears where you can enter the text. The list shows the already
specified values. To remove a value, select it and click Remove.
For the values, you can use uppercase and lowercase letters, but during processing lowercase letters will be
used. If you add the same name with different spelling, this variant will be ignored. Captured data may use
any spelling, that is, any combination of uppercase and lowercase letters. The returned result value will use the
spelling of the value you specified in the value list.
LookupList
Select LookupList to configure database support for the field.
ListBox
If you select ListBox, the field value must be a comma-separated list of strings which is shown as a list in
the Validation Client:
If values are captured for such a field, the field remains active in contrast to a textbox field. Thus the
Validation operator can capture several list items using the mouse or the keyboard. In addition, list items can
be selected and removed.
Button
If you select Button, a button with the current field value as label will be shown:
When the button is pressed, the code defined at the scripting entry point Validation (Field Changed) will
be called.
If the button should show a fixed label, specify this text at the Default value property on the Data tab.
If you configure database support for the first time, the list will be empty. In this case, click the ellipsis button behind the
Lookup list column list. The Manage Lookup List dialog box opens:
With this dialog box you can configure a list of all database files used by the current application. To add a database, click
Import. The Import a Table dialog box opens:
When you click Import, the database file will be imported, and will then appear in the file list of the Manage Lookup
List dialog box:
If want to remove a file, select it and click Delete. To import a newer version of a database file, select it and click
Reload.
Tip: If your database file changes frequently, an automatic reload can be configured using the program
LookupDatabaseUpdater.exe. For more information, see section "Configuring Lookup Lists in" the chapter
“Technical Operations” in Invoice Capture Center - Administrator’s Guide..
The Lookup list column list on the Presentation tab will show all available columns of all loaded database files as
<filename>.<columnname>. In the following example, a database file MyDatabase.txt has been
imported:
To assign a database column to the current field, click the desired column.
If you specify LookupList as control for further fields and assign columns of the same database file, the fields will be
shown as a group in the Validation Client. As first control in the group, a search box is shown, which is used to
determine the matching database record manually. If you type a search string in this box, or capture an according value
from the document, the list of the control will show all matching database records. As soon as you select an item in the
list, all fields of the group will be filled with the respective column values.
If you want to specify automation for such fields, you must use the automation method Generic SnapMatch for all
fields using columns of the same database file. SnapMatch will then use the respective database file to find a database
record that matches the current document, that is, where several column values of the result record could be found on the
document.
Take notice that it is of no relevance for SnapMatch, for which columns of the database you configure fields. Just
configure fields for those columns that should be exported in the invoice result. If you want to export a column whose
content is of no use for the validation operator, you can clear the Visible check box on the Presentation tab for this
field. The field will then be filled and exported, but it will not be shown in the Validation Client.
The Method list contains four generic automation methods, which can be used for custom fields:
Generic DatabaseColumn method: this method permits to export additional recipient or vendor data.
Generic Snapmatch method: this method permits to find a record in a database that matches data on the
invoice. Therefore this method is best suited if you have a database and expect that some data of one of its
records should be present on the invoice. For more information, see "SnapMatch Automation Method" on page
95.
Generic Adaptive recognition method: with this method you first perform training in order to store typical
invoices and the position of the index field on these invoices as templates in a database. During extraction, a
matching sample is determined for the current document, and the field is then extracted at the position where it
was trained for this sample. This method will be suitable for many kinds of fields. For more information, see
"Adaptive Recognition Automation Method" on page 96.
Generic Adaptive classification method: this method compares the whole textual content of the invoice
with previously trained samples in a database and returns a class value for the best matching samples in the
database. This method is suitable for fields that should determine a property of the whole invoice; for example
it would be suitable to determine the language the invoice is written in. For more information, see "Adaptive
Classification Automation Method" on page 97.
In addition, you can create your own rule-based method using the Document Extraction Design Studio. There a
lot of field types are available. For example you can search date or amount values, regular expressions in a
specified area of the invoice, search a field relative to the position of another field, and much more. For more
information, see "Rule-Based Recognition Automation Methods" on page 98. The respective method names
are preceded by Custom.
Note: For applications with language setting China (Traditional), Japan, Korea, and Thailand, Adaptive
Recognition is also used for extraction of the line item table.
Adaptive Recognition can't be used for:
line item fields,
fields using download data, such as vendor and recipient fields,
amount fields of applications with country setting Canada, China, or India,
fields occurring at different positions on invoices of the same vendor; therefore extraction of amount fields will
not benefit from training for vendors where the amount is positioned behind the last line item row so that its
position will depend on the number of line items.
Note: Take notice that for the standard amount fields values found by Adaptive Recognition will only be taken
into account if the standard rule-based method didn't return a result.
For more information about rule-based methods, see "Creating Rule-Based Methods with Document Extraction" on page
98 and OpenText Capture Document Extraction - Developer’s Guide.
Note: Modifications of the Document Extraction project will not take effect until you close the Settings dialog
box with OK afterward. It is not sufficient to save the project in the Document Extraction Design Studio.
Note: Modifications of the Document Extraction project will not take effect until you close the Settings
dialog box with OK afterwards. It is not sufficient to save the project in the Document Extraction Design Studio.
If you add your custom method to this node, it will replace the standard method only for documents of the respective
document type whereas the standard method will take effect for all other documents.
You can even replace the standard field for several user-defined document types by using the same field name at all
document type link nodes.
4. Type the desired column name in the Field name box on the Data tab.
4. Invoice Capture Center will take care of the specified VAT rates only for the standard fields. Therefore
take care to specify the valid VAT rates at the Vat Rates property of the Invoice Items Custom field.
For applications with country setting Hungary or India, special preconfigured Invoice Items Custom fields have to be
used, which are delivered as custom operator files that have to be loaded as global fields in the Design Studio.
For applications with country setting Hungary or India:
1. In the Document Extraction Design Studio, click the Globals tab in the Project Explorer view.
2. Right-click the view, and click Add Custom Operator on the context menu:
3. A file dialog box opens. Go to the CustomItems subfolder of the ICC installation folder. For Hungary
select the file CeItems_HU.ics, for India select the file CeItems_IN.ics, and then click Open.
Note: Modifications of the Document Extraction project will not take effect until you close the Settings dialog
box with OK afterwards. It is not sufficient to save the project in the Document Extraction Design Studio.
2. A new column node appears. Replace the default name with the column name you entered in the Settings
dialog box of the Customizing Client; it is important that these names are identical!
3. Open the new column node, select its index field subnode, and modify the field properties as needed. For more
information about field properties, see OpenText Capture Document Extraction - Developer’s Guide or the
online help of the Design Studio. In the following example, a Regular Expression column with the name
MyCustomColumn has been added; the node has been selected, and the Properties view shows the field
properties:
When you click OK to close the Settings dialog box, the application will be updated.
Note: Modifications of the Document Extraction project will not take effect until you close the Settings
dialog box with OK afterwards. It is not sufficient to save the project in the Document Extraction Design Studio.
4. The Lookup list column list will show all database columns of the current vendor and recipient database
files. Click the desired item in the list.
5. The Column confidence property takes effect if Deliver OCR results or Deliver database entry
checked by OCR is selected. If you specify a value at the Column confidence box, the column value will
only be returned if the value read from the invoice matched the database column value with a higher
confidence. If you set it to 100, the value will only be returned if the read value was identical to the column
value.
6. By default, the field will return the data that were read from the invoice for this column if it could be found. To
return the value of the respective database column, click Deliver database entry. To deliver the database
column value only if the value could be found on the invoice, click Deliver database entry checked by
OCR; this will make sure that the value is returned in the correct format for ICC/VIM.
Important
The code entered in the text editor must not contain a method header! Simply enter the method body as shown
in "Scripting Examples" on page 114.
Note that the scripting code will be executed even if you won’t have compiled it. Therefore always make sure that the
scripting code is correct and executable when you close the Settings dialog box.
To specify the namespaces used by your scripting code, click Namespaces…. The Specify Namespaces dialog
box opens.
Enter the namespaces for the current scripting code and click OK.
6.11.3.4 IScriptHelper
An IScriptHelper object is passed as parameter to each Validation scripting method. You can use this object to set the
status of a field to error and to show a message in the status line of Validation.
6.11.3.5 Tracing
An ITrace object is passed as parameter to each scripting method. You can use this object to write trace messages to the
Invoice Capture Center trace folder.
For more information about tracing, see OpenText Invoice Capture Center - Administrator's Guide.
6.11.3.6 Cultures
Recognition and Validation use an internal format for dates, amounts, and numbers:
dates: de-DE
amounts and numbers: en-US with decimal separator (.), but without thousands separator (,)
Be sure to use these internal formats in the scripting entry points.
Note: If the user selects an entry in a Lookup List control, the scripting method will be called for each data field
belonging to the Lookup List.
Note: The scripting method won`t be called for data fields belonging to a Lookup List because these fields never
have the focus and never lose the focus.
//Parameters:
//DataPool data
//ITrace trace
Important
For all SAP ERP downloads, the data size is restricted to 5 GB if you work with VIM 5.2 SP3 or higher. In this
case, a download is terminated as soon as 5 GB data have been downloaded.
After creation of the application, all Hot Spots are disabled. After configuring the Hot Spots, start those Hot Spots that
are used by your application.
All Hot Spots are created with a default scheduling behavior. Each Hot Spot checks its input channel regularly for new
data according to its scheduling configuration. For each Hot Spot you can configure separately whether it should check
its input channel once a day at a specified time, or regularly after waiting a specified interval.
Both types of SAP Hot Spots fetch their data directly from SAP ERP using a specified SAP ERP user account. For each
Hot Spot you can configure the SAP ERP user account separately. By default, the SAP ERP configuration of the Hot
Spots is set to the SAP ERP user account data that the user has entered during application creation.
Configuration of a SAP Extraction Link comprises the archive document types processed by the application. By default
the archive document type configuration of the Hot Spots is set to the document types that the user has entered during
application creation.
The configuration of a SAP Download Link provides additional settings specifying the download data.
If you later modify archive document types using the Settings dialog box, or if SAP ERP login data change, you must
adapt the Hot Spot configurations manually.
Caution
Within Invoice Capture Center, Hot Spots are always created and deleted automatically together with the
corresponding application. You should not create Hot Spots manually, and you should never delete a Hot
Spot of an Invoice Capture Center application manually.
Tip: If a Hot Spot is running, its properties can’t be changed and therefore the Properties command will not be
available. In this case click Stop. The state will then change to Disabled, and the Properties command will
become available.
The SAP Extraction Link will download 10 invoices per cluster node if less than 5 invoices per cluster node are in
progress so that about 5 to 15 invoices per cluster node will be in progress. Therefore you can let the default value
unchanged so that the Hot Spot will be activated every minute. To prevent timeouts in VIM, you should set the allowed
time for Extraction to 7200 seconds and the retry count to three; for more information see “Maximum allowed time for
Extraction and Validation” in OpenText Vendor Invoice Management for SAP Solutions - Configuration Guide.
Two SAP Download links are created for an application. The first is configured to run one per week performing a full
download including all data. The second is configured to run every day performing a delta download excluding the
vendor table.
For SAP Download links, use a fixed time for all days or for workdays, because master data downloads should not occur
more than once a day. The scheduled times must be coordinated between all Hot Spots and the download batch programs
on SAP ERP side. On a day, the SAP ERP download should run first, let’s say at 3 o’clock a.m. Then schedule the
download for the first Hot Spot not before 5 o’clock. For the next Hot Spot, schedule the download at least half an hour
later, and so on.
To specify several document types, separate them with a semicolon followed by a space.
For a SAP Download Link, click the ellipsis button behind the System box to modify the SAP ERP login data.
To control whether you specified correct login data, click Check login.
The list of document types should be identical to the list of document types defined for the application in the
Application settings dialog box on the Archive document types tab (see "Specifying Archive Document Types"
on page 61).
If you don’t want to download vendor tables, purchase order tables, or purchase order items, clear the respective check
box. If you want to download only changes, click Delta download.
The download has to be configured on the SAP ERP side (for more information, see Vendor Invoice Management for
SAP Solutions - Administration Guide) and on ICC Hot Spot side (for more information about ICC Hot Spots, see
"Configuring Hot Spots" on page 117). By default, ICC creates two SAP Download Link HotSpots when you create an
application. The first is configured for weekly full download, whereas the second is configured for daily delta-download
excluding the vendor table. The full download is useful to synchronize the tables, especially if you are using the Open
only option in the PO download program on SAP ERP side. The delta-download performed each day reduces the data
volume that has to be downloaded and ensures that downloaded data are up-to-date.
Caution
Do not use daily full download on SAP ERP side, and daily delta download on ICC side. This would lead to
fast growth of the database on the ICC side with many duplicate entries.
Note: The vendor table is always downloaded completely. If you switch off download of the purchase order table,
enabling of the option Delta-download has no effect.
In the Company codes area, you enter the name of the SAP ERP systems in the Logical system box, and the
corresponding company codes for which data should be downloaded into the Company code column. Whereas
Vendor Input Management collects data for all available SAP ERP systems and company codes, only data for the
specified logical systems and company codes will be downloaded to the Recognition Server.
It is important that you either leave the subsystem and/or company code empty (empty means “all”), or add all relevant
combinations of SAP ERP system (logical system name like for example D01CLNT800) and company codes. Otherwise
the Hot Spot will not retrieve all relevant data and recognition quality be reduced.
When the download operation has been completed, the temporary download files will be deleted immediately by default.
If these files are needed for diagnostic purposes, select the Keep data for diagnostics check box. In this case the
download files will be available in the cache until the respective cache entry is deleted later on.
Tip. To determine how long a download takes, start the Trace Viewer, go to the DOKuStar Load Manager
section of the trace file list and look into traces files whose name starts with DOKuStarClusterNode. There
trace records permit to determine how long the three steps of a download took: downloading the data, writing the
data to the database, and updating the vendor index file.
A newly configured Hot Spot is always in state Disabled. If you create a new application, all its Hot Spots will be in
this state. In this case configure each Hot Spot of the new application and start it afterwards. The state will then change
to Scheduled.
To modify configuration of a Hot Spot later, you will have to stop the respective Hot Spot temporarily.
To start or stop a Hot Spot, select it, and then click the respective command on the toolbar, or on the context menu of the
Hot Spot.
If a Hot Spot should run before it’s appointed scheduling time has arrived, select it, and then click Run now on the
toolbar, or on the context menu of the Hot Spot. This will sometimes be useful during testing. The Run now command
will only be available if the Hot Spot is in state Scheduling.
Error Description
Wrong field To specify a field at a wrong logical position is a severe error during training, because this will
yield an extraction error each time the sample is used for extraction. Adaptive Recognition will
always extract the field at a wrong position.
Field not present on the Some fields may not be present on all invoices of a vendor. If a candidate doesn’t contain such a
training document field, it is not suitable for training. Wait for a better candidate that contains all fields where
Adaptive Recognition is used.
Wrong logical field size Sometimes the tooltip will contain an additional character or word that is printed closely to the
field value. For example a VAT rate could contain a following percent sign, or a value could
include a following comma. In this case, hold CTRL down while you draw a rectangle around the
field value so that the captured value will only contain the desired characters.
Adaptive Recognition is used for some optional fields (Supply Date, Freight Amount, or Handling Charges).
These fields are not used by default. They have to be activated manually (see "Activating Optional Fields" on page 80).
Adaptive Recognition is also used for any of the standard fields of applications with country setting China (Traditional),
Japan, Korea, and Thailand.
In addition, Adaptive Recognition can be used for custom fields (see "Adaptive Classification Automation Method" on
page 97).
Finally, Adaptive Recognition can be activated for some standard invoice fields (see "Using Adaptive Recognition for
Standard Invoice Fields" on page 33).
Adaptive Recognition fields have to be trained (see "Training of Adaptive Recognition Fields" on page 122). Learning
data are kept in different repositories within a database. There is a single Adaptive Recognition (ART) repository for all
fields of an application that use the Adaptive Recognition automation method.
If you don’t use Adaptive Recognition fields in an application, the ART repository of the application will be empty.
Adaptive Classification can also be used for custom fields (see "Adaptive Classification Automation Method" on page
97). Adaptive Classification is not used by default. Adaptive Classification fields have to be trained (see "Training of
Adaptive Classification Fields" on page 124). Learning data are kept in different repositories within a database. There is
a separate Adaptive Classification repository (ACT) for each field of an application that uses the Adaptive Classification
automation method.
If you don’t use Adaptive Classification fields in an application, there won’t be ACT repositories to administer.
On the left side of the Adaptive dialog box you find a tree structure presenting the ART and ACT repositories of the
current application. Select the respective tree node to display the corresponding repository on the right side of the
Adaptive dialog box.
For each sample (trained invoice) of Adaptive Recognition, the trained field values are shown in addition to
administrative data.
As soon as a sample has been used for data extraction, the Usage Count column indicates how often it has been used,
whereas the Last Usage Date column shows when it was used last time. If you click Reset usage counters ( )
in the toolbar, last usage date and usage count of all samples will be reset.
The state of a sample, indicated by the State column, affects recognition and training.
A sample in state confirmed will never be removed automatically.
State unconfirmed is the default state for newly trained samples. A sample in state unconfirmed may be
removed during training of a new sample if it has been trained more than 30 days ago, its usage count is zero,
and repository size has reached its maximum size of 2000 samples.
A sample in state excluded will not be used for data extraction. In addition, it will not be removed
automatically.
Newly trained samples always have state unconfirmed. The first three buttons of the toolbar permit to assign a
different state to selected samples. You can use this, to prevent removal of samples by setting their state to confirmed
after you have checked that they are useful and have been trained correctly.
It is recommended to let the state unconfirmed unchanged for all samples. Thus samples that are no longer needed can
be removed automatically during training if the repository has reached its maximum size. If you want to change the state
of a sample to confirmed, you should be sure that the sample has been trained correctly and that it will not become
obsolete in the future.
Note: Take notice that samples that have been trained with ICC 6.0 or earlier, will show state confirmed.
When you select samples, the context menu and corresponding toolbar buttons permit to delete the selected samples from
the repository, or to save the images to a folder.
For easier identification, samples can be renamed. To rename a sample, right-click the sample row and then click
Rename. A dialog box opens where you can enter the new name.
If you selected several samples, you can use the arrow buttons to browse through the samples.
On the left side, the trained field values are shown. When you select a field result, and the respective box coordinates are
available, the result is highlighted on the document. When you position the pointer on a field, additional data, such as
state and confidence value are shown.
You can’t correct a sample in this view. If a sample has been trained incorrectly, click Delete on the toolbar to remove
the sample from the repository.
For all samples in the repository it shows, amongst others, the identification number, which is used to identify the
sample, the date when it was added to the repository, the file name of the image file used for training, and the class.
Each possible field value of an ACT field is a class. When you select a field value for an ACT field in the Validation
Client or in the Adaptive dialog box, you classify the invoice.
The context menu of an entry permits to open or remove a sample, or to change its document class (that is, the field
value of the ACT field) directly:
Delete will remove the sample from the repository. Use this command to delete incorrectly trained samples.
Change Class permits to change the document class of the selected samples directly. A Change Class dialog box
opens where you can select the new document class. Use this command to correct incorrectly trained samples.
The toolbar of the view permits to refresh data, and to start training.
For each document of the repository, the Analysis tab displays, how this document would have been classified if it
were not trained. Thus this view shows how Adaptive Classification would classify unknown documents using the
documents in the repository. Thus, if you have a set of documents, there is no need to divide it into a set of training
candidates and a second set you use to test classification afterwards. You can train all your documents and still you can
then test classification and get a detailed analysis about the quality of classification for the different classes.
Test classification may yield different results: For example, a document may not be classified at all if there are no similar
documents in the repository. It may be rejected if its confidence value is below a threshold, or it may be classified to its
trained class (correct) or to another class. These different states are represented by traffic light colors:
Green: documents that have been classified to their trained class (correct)
Yellow: documents that have been rejected or couldn’t be classified at all (reject)
Red: documents that have been classified to another class (incorrect)
Note: If only a single document has been trained for a class, it will necessarily be classified to a different class in
this test classification. But you should train additional samples for such a class anyway.
The Analysis tab shows the classification results of all documents in two ways:
The bar on the right visualizes the state of all documents in the repository.
In the center, the states are displayed in a map per result class. The result classes are sorted by the number of documents
assigned to the respective class. The width of a column indicates the number of documents.
Thus every area unit in the map represents the same number of documents. Thus it provides a view to the contents of the
repository.
The map can be zoomed, if necessary.
Now horizontal lines appear in the reject and incorrect areas, separating areas of documents rejected or misclassified to
another class. In green yellow, you can also see the documents that are correct, but rejected.
The documents in these areas can be selected in the same way as described above.
Refresh
The first button of this group starts a refresh of the data shown in the view. Use this button if you modified the
repository.
The second button permits to restrict the analysis to the last trained samples. When you click the button, a dialog box
opens where you can specify a date. When you confirm the dialog box with OK, the view will be refreshed with the
samples that were trained after the specified date. If new samples are trained, after your application has already been
released and is running in production, you can see how the new samples fit the classifier. If you select all substitutions,
you will see only the problematic samples added since the last released version of the repository.
Zooming
With these buttons you can enlarge or reduce the representation shown in the view horizontally or vertically. As soon as
you enlarge, a corresponding scroll bar will appear, which permits to scroll to parts of the view that are not visible. With
the last button you can return to the default representation with a single click.
Mode Selection
The first button shows the distribution of rejects and incorrect samples. If the button is not pressed, all rejects and
substitutions for a class are shown as a single area. If the button is pressed, the yellow areas will tell not only the rate of
documents that have been classified to the wrong class, but which classes were substituted how often. The green-yellow
areas will show, how many samples were classified correctly, but were rejected as unreliable.
The second button toggles between two different representations of the classification results in the Analysis tab. The
view either shows classification rates per trained class, or classification rates per result class.
The first representation will indicate the quality of the different trained classes. The second representation will show you
how reliable the result is if a specific class has been found.
Selecting
The first two buttons of this group permit to select all documents that have been classified to the wrong class during
analysis, or the documents that could not be classified at all, respectively. The last button removes all previous
selections.
Color
This button toggles between color and black-and-white representation of the view.
Edit Samples
Opens all selected samples, that is, samples of the areas shown hatched in the view, on a Training tab so that
they can be edited.
Show Bars
The Analysis tab indicates the number of samples trained for a class through the size of the respective area.
By default the bars for the different classes all have full height but different width according to the number of
samples. If this command is marked, all bars are shown with same width but different height.
Save Report
This command permits to save the result statistics of the data currently shown in the view to a text file. All
data items in the files are separated by tabs to make it easy to import the data into other programs. The
following example shows a report file:
Report on 2010.04.26 09.39.51 Trained class distribution per result class
Important
Normalization has been modified incompatibly with ICC 6.0. If you open an application created with ICC 5.2
that uses scripting or custom normalization, you will be asked whether the application should be migrated. If
you answer with No, the application will not be upgraded and the documentation of ICC 5.2 is still valid with
regard to normalization for this application. If you answer with Yes, you will have to adapt your scripting
code. Applications that use neither scripting nor custom normalization will be always be upgraded.
After data extraction, the Extraction action point can be used to specify custom rules that take effect before
the Custom Extraction scripting entry point and ICC Processing. Then, two scenarios have to be distinguished.
After the Data Extraction scripting entry point, the Extraction step will provide data for export to SAP ERP in
case that the invoice will not be sent to Validation. Here, at an action point Export SAP, default normalization
rules for standard fields will be applied, which convert the standard fields to the format expected by SAP ERP.
For export to Validation, separate rules can be specified at the Export Validation action point.
In Validation there is an action point Display, which converts the internal language-independent format into
the format that should be displayed. At a second action point Manual Input data captured by the Validation
operator are checked and normalized. Before the data are exported to SAP ERP in Validation, again the
normalization rules of the action point Export SAP are applied.
In the project tree of the Settings dialog box, all normalization rules can be found under a Rules node, which is a
subnode of the Fields node. The Rules node has a subnode for each action point:
At the field definition nodes, a Rules subnode with the same structure is shown. But here only those rules are
shown that are attached to the respective field. The following example shows the InvoiceDate field node
where you can again see that a custom rule has been attached at the Display action point, which will replace
the standard rule:
The Add New Rule dialog box opens. In the Rule name box, type a name for the new rule. To create a copy of an
existing rule, click the respective rule in the rule list showing all predefined and all custom rules; if you accept the
default value <empty>, an empty rule will be created:
6.15.6 Rules
A rule is composed of a list of normalization actions, which are processed sequentially as long as the field state remains
OK. As soon as an action sets the field state to Error, processing continues at the next Continue on error action in the
rule; or it terminates if no such action is present.
The following table shows how the different actions affect field value or field state:
To modify the action list, right-click an item in the action column; if you just created a new sequence, use the empty first
line that will be shown. The following context menu appears:
Note: If you click a rule that is a subnode of a Predefined Rules node, its action list will be shown, but the
context menu will not be available because the default action sequences can’t be modified.
To add a character translation, click Add on the context menu, and type a single character in the From and To boxes of
the appearing new row.
With the remaining commands you can delete a selected item from the list, or change the order by moving a selected
item up or down.
In the Translation table box, type a name for the word translation table. If you choose a new name, the translation
table will be saved to a new file when you click OK later. If you enter the name of an already existing translation table, it
will be loaded, and its entries will be shown. This permits to use the same translation table in different actions. In this
case you should be aware that all actions using the same table will be affected if you modify the table later.
To modify the translation table, use the commands of the context menu to add new entries, to delete a selected entry, or
to change the order by moving a selected entry up or down.
When you click OK, the translation table will be saved to a file using the translation table name as filename.
Enter a regular expression and a replacement string. In the replacement string you can refer to parts of the found match
that correspond to parts of the regular expression that you enclosed in parentheses. With $n you refer to the n-th opening
parenthesis in the regular expression. With $0 you refer to the whole matching string without having to enclose the
complete regular expression in parentheses.
To be able to close the dialog box with OK, you have to test an example string successfully. Type a test string in the
Sample input data box and click Test. If the string matches the regular expression, the result is shown, the orange
area turns green, and the OK button becomes available.
Caution
Don't modify format specifications on the Format tab of the Region and Language dialog box of the
Windows Control Panel because this would affect checks and formatting within ICC.
Click one of the available data types Amount, Date, or Decimal in the Type list, and enter all cultures (.NET cultures
such as en-US) into the parse culture list separated by a space that should be used when trying to parse the current field
value.
At the Format parameter enter a .NET Framework format string, and at Format culture enter the culture that should
be used to format the result.
Caution
Don't modify format specifications on the Format tab of the Region and Language dialog box of the
Windows Control Panel because this would affect checks and formatting within ICC.
If your application doesn’t need to use the new extended customizing, you don’t need to migrate the application. In this
case, the application will run in legacy mode and normalization will behave as in ICC 5.2. If you open such an
application in the Customizing Client, answer with No when you are asked whether the application should be migrated.
If you want to use the new extended customizing features or if you want to write country-independent scripting code,
your application will have to be migrated. If your application uses scripting or if you modified normalization using the
Extended Customizing dialog box, you will have to check and adapt your application. The following table gives an
overview how you have to proceed depending on your application:
When an application is loaded, ICC will perform some checks and decide how to proceed. Three cases have to be
distinguished:
If the application has been created with ICC 6.0 SP1 or later, nothing has to be done.
If the application has been created with an earlier version, but it contains neither scripting nor extended
customizing (that is, no modifications were made in the Extended Customizing dialog box), automatic
migration will be performed.
If the application has been created with an earlier version, and contains scripting in Validation or extended
customizing, the application will run in legacy mode, that is, normalization will behave as in version 5.2. If such
an application is opened in the Customizing Client, you will be asked whether the application should be
migrated. Before you answer with Yes, take care to keep a copy of your application for later references.
With this in mind, check your Validation scripting and adapt it, if necessary. In Validation, field values will now be
formatted for display and after manual input controlled by extended customizing.
The columns RESERVE1 and RESERVE2 can be used to supply spelling variants for items that otherwise might not
be matched correctly.
Note: Vendor master data and purchase order data have to be available for local testing to get good recognition
results.
Tip: By default, the Open command is not available. To be able to import invoice from a SAP ERP system, click
ICC Options on the application menu, and then select the Activate Validation mode in Customizing Client
check box.
6.17.2 Validation
After opening an invoice within the Customizing Client, the Application tab will show the invoice image and the
extracted invoice data.
The Validation Client and the Customizing Client provide the same validation facilities. Thus, for more information
about the Validation user interface, see OpenText Invoice Capture Center - User’s Guide. Note that the validation view
of the Validation Client shows the original, scanned invoice image whereas the Customizing Client shows a
preprocessed image.
6.17.4 Monitoring
During application testing you may want to monitor the system. For information about monitoring, see OpenText Invoice
Capture Center - Administrator’s Guide.
Note: If an entry exists for VIM Internal Name, the field value is visible in VIM baseline screens or can be made
visible by configuration. It can be included in the business rule Missing mandatory information by
configuration.
Statements under VIM Process focus on the Business Rules or Check Rules. If there are no business rules
for a field but the field value is used elsewhere, this is mentioned. If nothing is stated under VIM Process, the
VIM baseline makes no use of this field except for showing it in the screens or checking it in business rule
Missing mandatory information.
8.1.31 PO Number
Internal Name Countries Validation Export VIM Internal Name
Visible Mandatory
InvoiceOrderNumber All Yes Yes Yes EBELN
8.1.69.1 House No
Internal Name Countries Validation Export VIM Internal Name
Visible Mandatory
RemitHouseNumber US Yes No Yes REMIT_HSNM1
8.1.69.3 City
Internal Name Countries Validation Export VIM Internal Name
Visible Mandatory
RemitCity US Yes No Yes REMIT_AD_CITY1
8.1.69.4 ZIP
Internal Name Countries Validation Export VIM Internal Name
Visible Mandatory
RemitZIP US Yes No Yes REMIT_PSTLZ
8.1.69.5 Country
Internal Name Countries Validation Export VIM Internal Name
Visible Mandatory
RemitCountry US Yes No Yes REMIT_LAND1
8.1.70.1 House No
Internal Name Countries Validation Export VIM Internal Name
Visible Mandatory
ShipHouseNumber US Yes No Yes SHIPTO_HSNM1
8.1.70.2 Street
Internal Name Countries Validation Export VIM Internal Name
Visible Mandatory
ShipStreet US Yes No Yes SHIPTO_STREET
8.1.70.3 City
Internal Name Countries Validation Export VIM Internal Name
Visible Mandatory
ShipCity US Yes No Yes SHIPTO_AD_CITY1
8.1.70.5 Country
Internal Name Countries Validation Export VIM Internal Name
Visible Mandatory
ShipCountry US Yes No Yes SHIPTO_LAND1
8.1.70.6 Region
Internal Name Countries Validation Export VIM Internal Name
Visible Mandatory
ShipRegion US Yes No Yes SHIPTO_REGIO
8.2.2 PO Number
Internal Name Countries Validation Export VIM Internal Name
Visible Mandatory
ItemOrderNumber All Yes No Yes EBELN
8.2.4 Amount
Internal Name Countries Validation Export VIM Internal Name
Visible Mandatory
ItemAmount All Yes No Yes WRBTR
8.2.5 Quantity
Internal Name Countries Validation Export VIM Internal Name
Visible Mandatory
ItemQuantity All Yes No Yes MENGE
Adaptive Classification
An adaptive classification field does some sort of classification of a document. Normally the field value is one of a set of
terms. The term itself is not contained in the text of the document but can be determined by examining the entire contents
of the document. When configuring an adaptive classification field you will enter the set of valid terms for the field. At
runtime the adaptive classification field will be presented to the user with a combo box containing the list of terms. The
Recognition Server and the Recognition Nodes have to learn how to select the correct term for an adaptive classification
field from the set of terms. They need a sample for each kind of invoice and the correct term for each sample. These
samples and the according term have to be provided by training. During training, the term is selected manually by the
user. The document and the term are then stored in a learning database. If a similar document is processed later on,
adaptive classification will then find a matching sample in the learning database, and will then use the learned term.
Adaptive Recognition
The Recognition Server and the Recognition Nodes have to learn how to extract data for an adaptive recognition field.
They need a sample for each kind of invoice, e.g. for each vendor, and the position where the field can be extracted on
this kind of invoice. These samples (=templates) and the according position information have to be provided by training.
During training, the field is captured manually using the mouse. The layout of the document and the positions of the
respective field are then stored in a learning database. If a similar document is processed later on, adaptive recognition
will then find a matching sample in the learning database, and will then extract the field at the learned position.
Application
An invoice application bundles all settings that are needed to process invoices of a certain kind, e.g. the country in which
the recipients of the invoices reside, or which fields have to be extracted by which extraction method, or which fields
should be visible in the Validation Client.
ART
Abbreviation of Adaptive Recognition Technology
Cluster
See Recognition Cluster
Cluster Node
A Cluster Node is a component of a Recognition Cluster performing job processing. The Recognition Cluster contains
one Cluster Node for each processor of the Recognition Cluster. The Load Manager performs load balancing and
distributes recognition jobs among all Cluster Nodes of the Recognition Cluster. If there is only a Recognition Server
and no Recognition Nodes, there will be as many Cluster Nodes as processors in the Recognition Server machine. If the
Recognition Cluster consists of a Recognition Server and several Recognition Nodes, there will be as many Cluster
Nodes as processors in the Recognition Server machine and in all Recognition Node machines.
Customizing
Invoice Capture Center provides various settings to adapt the system to customer needs. These customizing settings are
bundled in invoice applications. To customize Invoice Capture Center for customer needs you have to create invoice
applications.
Customizing Client
The Customizing Client is a Windows program that is used on the development system to customize Invoice Capture
Center, i.e. to create and configure applications. It is also used to test applications, to export them to a SAP ERP system,
and to import them on a test or production System. In addition, it is used to monitor and administer an Invoice Capture
Center system.
Document Reader
In an Invoice Capture Center system the actual data extraction from an invoice image and super-ordinate tasks, like
splitting of pages or export of results, are performed by a component named Document Reader.
Hot Spot
The Recognition Server uses Hot Spots to fetch recognition jobs, download data, and learning data from its SAP ERP
system. A Hot Spot becomes active regularly after expiring of a specified interval or at a specified time on specified days
and then tries to fetch new data. Each application uses Hot Spots of two different types. A SAP Extraction Link is used
to fetch new recognition jobs from SAP ERP. A SAP Download Link is used to fetch new data, such as vendor master
data and purchase order data.
ICC
Abbreviation of Invoice Capture Center
Invoice Application
See Application
Link
Links are used to download recognition jobs and download data from SAP ERP. Links are implemented as so-called Hot
Spots. The Hot Spots used for the Links connect to SAP ERP regularly at specified times and fetch available jobs or
data.
Load Manager
Invoice Capture Center uses load balancing in order to distribute recognition jobs on several processors of the
Recognition Server and on additional Recognition Nodes. Load balancing is performed by the Load Manager. During
configuration of the Invoice Capture Center system, you configure the Load Manager on the Recognition Server and
create computer nodes for all additional Recognition Nodes and Runtime Nodes for all processors on each Recognition
Node.
Master
On a Recognition Cluster, the Load Manager performs load balancing over all processors of the Server and the
Recognition Nodes. The Load Manager on the Server acts as Master. It transfers recognition jobs that are fetched from
SAP ERP regularly to free Cluster Nodes for processing.
Recipient Data
Invoice Capture Center can derive the company code and target SAP ERP system automatically using recipient address
data found on the invoice. During application configuration, a database containing recipient data together with the
corresponding company codes and SAP ERP system names has to be supplied if automatic company code detection shall
be used.
Recognition
Recognition comprises the extraction of essential invoice data, such as date, amount, order number etc., from the scanned
invoices and the intelligent verification of these data. Recognition is performed by the Recognition Cluster.
Recognition Cluster
The Recognition Cluster is used for automatic data extraction. On smaller systems with low throughput it consists of the
Recognition Server - typically a computer with several processors - that also hosts the Customizing Client. On large
systems with high throughput, further Recognition Nodes can be added. The Load Manager performs load balancing and
distributes recognition jobs among all processors of the Server and the Recognition Nodes.
Recognition Node
For huge systems the Recognition Server can be supported by additional machines, installed as Recognition Nodes. The
Load Manager on the Recognition Server fetches jobs from SAP ERP using a SAP Extraction Link and distributes the
jobs among all processors of the Server and the different Recognition Nodes. By adding Recognition Nodes, the
processing power of the whole Recognition Cluster can be adapted to the required throughput.
Recognition Server
Document processing is performed on a Recognition Cluster consisting of a Recognition Server and possibly additional
Recognition Nodes. The Recognition Server comprises the Customizing Client that is used to customize, configure, and
monitor the Invoice Capture Center system. It also comprises the databases containing vendor master data, purchase
order data, and learning data. It also hosts the Hot Spots that are used to fetch jobs and download data from the SAP ERP
system regularly. The Load Manager of the Recognition Server is the master of the recognition cluster. It contains the
cluster configuration and distributes recognition jobs on the processors of the Server and additional Recognition Nodes.
Slave
On a Recognition Cluster, the Load Manager performs load balancing over all processors of the Server and the
Recognition Nodes. The Load Manager acts as a Slave on each Recognition Node. When a Runtime Node terminates a
job, it waits until the Load Manager on the Server that acts as Master supplies it with a new job.
Server
See Recognition Server
Standalone System
For development and testing of applications, all components of Invoice Capture Center can be installed on a standalone
system, i.e. on a single computer.
Template-Based Recognition
See Adaptive Classification
Training
To be able to extract adaptive recognition fields or adaptive classification fields automatically, some training has to be
performed first. Training can be performed manually by the user on the Customizing Client.
Validation Client
The Validation Client is a Windows program that is used by a clerk to check and correct data that have been extracted
automatically and to capture additional fields that could not be read automatically. The corrected and completed data are
then passed on to Vendor Invoice Management.
VIM
Abbreviation of Vendor Invoice Management
S V
Scripting 111, 148 Validation 3