Keywords for the Vendor / Business Unit and Invoice Header Data groups of fields
FlexiCapture for Invoices uses keywords to search for fields on an image: if a keyword is detected, the program looks for a field in its vicinity. Common examples of keywords include field titles and captions. For example, suppose an invoice contains the textInvoice Date: <value>. In this case, Invoice Date is the keyword used to find the InvoiceDate field.
Keywords can be specified using Located elements. Each Located element can be attributed to only one field in the Document Definition, but each field can have more than one Located element, each of which describes a separate model of the relationship between the keyword and the field on the invoice.
- Keywords are specific to languages and countries. When a Document Definition is applied, the program takes keywords from the vendor’s country, the business unit’s country, and the languages of these countries.
- Normalization is applied to keywords:
- Tabs, spaces, line-break characters, and diacritical marks are ignored.
- Letter case is ignored (the program makes no distinction between lowercase and uppercase letters). For example, for keyword purposes, Tax point is the same as taxpoint.
- The program allows for 2 or 3 recognition errors in keywords (depending on the element).
- Sometimes a keyword appears in more than one language, for example Quantity/Anzahl. We recommend adding the whole keyword (Quantity/Anzahl) to both languages.
Vendor / Business Unit
| Located Elements | Description | Specified in the properties of |
|---|---|---|
| Bank Account | Keywords for the Bank Account field | Country (the Keywords tab) |
| Bank Code | Keywords for the Bank Code field | Country (the Keywords tab) |
| IBAN | Keywords for the IBAN field | Country (the Keywords tab) |
| National VATID | Keywords for the National VATID field | Country (the Keywords tab) |
| Total | Keywords for the Total field | Country (the Keywords tab) |
| Total tax | Keywords for the Total tax field | Country (the Keywords tab) |
| VATID | Keywords for the VATID field | Country (the Keywords tab) |
Invoice Header Data
| Located Elements | Description | Specified in the properties of |
|---|---|---|
| PurchaserNameLabels | Different possible names of the PurchaserName field in the language. | Language |
| PurchaserNameFalseFieldPrefixes | Different possible names of fields that may be incorrectly identified as the PurchaserName field. We recommend specifying these only if you see specific cases. | Language |
| CreditNoteKeywords | Words in the language that may be used to identify the document as a Credit Note. | Language |
| InvoiceIdentifiers | Words that indicate that the document is an invoice. | Language |
| OrderNumberLabels | Possible names of the OrderNumber field in the language. | Language |
| InvoiceNumberExcludePreffixes | Words that are placed before numbers and may be omitted, such as “No.” | Language |
| InvoiceNumberExcludeSuffixes | Words that are placed after numbers and may be omitted. We recommend specifying these only if you see specific examples. | Language |
| InvoiceNumberWithDateLabels | Text that may precede an invoice’s number if that name is written on the same line as the date and they are separated by a slash or other character. Examples: Invoice Number / date 23061336 / 07.07.2013 Rechnungsnr./ -datum 23061336 / 07.07.2013 | Language |
| InvoiceNumberHighConfidenceLabels InvoiceNumberLowConfidenceLabels | The InvoiceNumberHighConfidenceLabels list contains fragments of text that unambiguously identify the field, such as “Invoice number” and “Document number”. More ambiguous fragments such as “No.” and “Bill” are listed in InvoiceNumberLowConfidenceLabels. | Language |
| DueDateLabels | Possible names of the DueDate field in the language. | Language |
| DeliveryDateLabels | Possible names of the DeliveryDate field in the language. | Language |
| InvoiceDateLabelsNearCity | Any fragment of text that may be used to separate the name of a city and the date if they are on the same line, such as a comma. | Language |
| InvoiceDateLabelsNearInvoiceNumber | Any fragment of text that may be used to separate the invoice number and the date if they are on the same line. | Language |
| InvoiceDateHighConfidenceLabels InvoiceDateLowConfidenceLabels | The InvoiceDateHighConfidenceLabels list contains fragments of text that unambiguously identify the field, such as “Invoice date” and “Document date”. More ambiguous fragments such as “Tax Date” and “Tax Point” are listed in InvoiceDateLowConfidenceLabels. | Language |
Keywords for the Amounts group of fields
FlexiCapture is capable of determining when a word is part of a word combination. For example, if you add the keywords total and total netto, and the image contains total netto, it is identified as total netto and not total.- The program allows for up to 3 recognition errors in keywords.
- This limitation can lead to errors. For example, the program may mix up the words brutto and netto. To avoid these errors, if you add a value (such as total netto) to AmountTotalNettoLabels, add the corresponding value (such as total brutto) to AmountTotalLabels.
| Located Elements | Description | Specified in the properties of |
|---|---|---|
| AmountTotalHighConfidenceLabels AmountTotalLowConfidenceLabels | The captions of the Total field. Place captions that occur only with the total amount on the invoice in the HighConfidence group. Captions that may occur with other amounts should be placed in the LowConfidence group. | Language |
| AmountTotalNettoLabels | The keyword that may be to the left of or above the Total Netto field. | Language |
| AmountTotalTaxLabels | The keyword that may be to the left of or above the Total Tax field. | Language, Country (the Keywords tab) |
| ReversedChargeKeywords | Words that can be used to indicate “Reversed Charge” in the language. | Language |
| Tax Rates | Keywords for tax rates. | Country (the Tax Rates tab) |
| Currency | Keywords or characters that indicate currencies. | Country (the Currency tab) |
It is usually a bad idea to add the same word to Total Netto and Total.
Keywords for the Line Items group of fields
Keywords are used to find table titles that contain invoice fields, and to find specific columns in tables. Words that are frequently used in the titles of columns that correspond to Located Elements work best in this capacity. If the same word can be found in the titles of different columns, we recommend adding it to the Located Elements of these columns. In this case, the program can tell the columns apart by the contents of their cells. If the program still fails to tell the columns apart, or if the word is used frequently in one column and infrequently in others, leave the word in the Located Element of the column where it is used most frequently and remove it from the Located Elements of the other columns.- The program allows for up to 3 recognition errors in keywords.
- Keywords cannot take up more than one line.
- The contents of a column must be located directly beneath the column title.
- Keywords on images must be surrounded by spaces, commas, or periods. So Quantity/Anzahl is one keyword, and Quantity / Anzahl is two. If a keyword may be written in more than one way, add all of the different ways it can be written, for example Quantity/Anzahl, Quantity, and Anzahl.
| Located Elements | Description | Specified in the properties of |
|---|---|---|
| LineItemsArticleNumberLabels | Keyword for the Article Number column | Language |
| LineItemsArticleNumberBULabels | Keyword for the Article Number BU column | Language |
| LineItemsCurrencyLabels | Keyword for the Currency column | Language |
| LineItemsDeliveryDateLabels | Keyword for the Delivery Date column | Language |
| LineItemsDescriptionLabels | Keyword for the Description column | Language |
| LineItemsDiscountAmountLabels | Keyword for the Discount Amount column | Language |
| LineItemsDiscountPercentageLabels | Keyword for the Discount Percentage column | Language |
| LineItemsMaterialNumberBULabels | Keyword for the Material Number BU column | Language |
| LineItemsMaterialNumberLabels | Keyword for the Material Number column | Language |
| LineItemsOrderDateLabels | Keyword for the Order Date column | Language |
| LineItemsOrderNumberLabels | Keyword for the Order Number column | Language |
| LineItemsPositionLabels | Keyword for the Position column | Language |
| LineItemsQuantityLabels | Keyword for the Quantity column | Language |
| LineItemsQuantityOrderedLabels | Keyword for the Quantity Ordered column | Language |
| LineItemsQuantityUndeliveredLabels | Keyword for the Quantity Undelivered column | Language |
| LineItemsSubtotalVariants | Reserved | Language |
| LineItemsTotalPriceBruttoLabels | Keyword for the Total Price Brutto column | Language |
| LineItemsTotalPriceNettoLabels | Keyword for the Total Price Netto column | Language |
| LineItemsUnitsOfMeasureLabels | Keyword for the Units of Measure column | Language |
| LineItemsUnitsOfMeasureVariants | Possible field values in the Unit Of Measure column | Language |
| LineItemsUnitPriceDenominatorLabels | Captions in the column header used as a multiplier for the Unit Price column when computing the equality for an invoice item: Unit Price * Denominator * Quantity = Total Netto.For example, if UnitPrice is the price per unit and Quantity indicates the number of packs, the Boolean column UnitPriceDenominator corresponds to a column that indicates the number of units in one pack. | Language |
| LineItemsUnitPriceLabels | Keyword for the UnitPrice column | Language |
| LineItemsVATAmountLabels | Keyword for the VAT Amount column | Language |
| LineItemsVATCodeLabels | Keyword for the VAT Code column | Language |
| LineItemsVATPercentageLabels | Keyword for the VAT Percentage column | Language |
The quality of keyword detection depends on the quality of full-text recognition.