A format essentially defines what you want to get out of a certain type of document. It defines the structure of the output and is very visible in the Data View of the Data Entry Companion.
Format is a structured collection of Fields of interest and consists of
Text fields: Pieces of information that appear literally in the document
Tag fields: Information that does not appear literally in the document, but that can be derived. Tags come from a finite set of options.
Rows: Text and Tag fields that logically belong together
Format consists of at least one table: Format table.
There can be as many additional tables as needed for information that follows a row/table logic.
Format Table
The format table contains text and tag fields. These fields have a number of properties.
Property | Description | Default |
Active | A toggle to turn the search for this field on/off | on |
System name | An immutable uuid that serves as a reference to the underlying machine learning model | UUID assigned by the system |
Type | Indicates whether this is a text field or a tag field | text field |
Options - text field | For a text field, the number of options is virtually unlimited. You can specify a data type other than string to reduce the number of options for a text field. For example, here you can specify that the text should be an amount | string |
Options - tag field | A tag field can only take values from a finite set of values. Here, you can exhaustively list the options for that tag field | Empty list |
Formula - computed field | A formula through which allows you to add excel-like business logic to our solution | |
Mandatory | You can indicate whether or not a field is mandatory. Mandatory fields will be marked as read when not found and will block submission if not filled | false |
Multiple | You can indicate if a field of interest can appear multiple times. | false |
Condition | Some fields of interest are only relevant depending on the value of other fields of interest. For example, you may only be interested in the VAT number of an invoice if if has been confirmed that the document is a valid invoice (valid_invoice field == true) | No condition (always show) |
Display name | This is the label that will be shown in the FrontEnd to the user of the data entry companion and in the stats pane | No default, must be specified |
Technical name | This is the label that will be used when communicating to servers | Same as display name |
Other tables
You may want to maintain the row logic for tables in the extraction of information. To this end, we allow you to specify tables, which contain one or more row types.
For example: you may have a table of line items with four columns (description, unit price, quantity, line total). You would then have two types of rows for this table: line items and a total line. The total line would be of a different type as it will not have a unit price. It will have a fixed description ("total"), a quantity and an overall total.
Rows are essentially a sorted list of text and tag fields. In addition, they will have a row_type, which defines to which table they belong.
Edit Formats
Navigate to "Studio" > "Formats" to get an overview of the available formats. Click on edit format to have a view on the format. You'll see the Fields table and other tables for the format.
Edit Formats through the Studio Controls in the Data Entry Companion
Format defines the fields you are interested in for a given document type. It is reflected in the data table that you can see in the Data Entry Companion. We have also made it possible for you to edit the format directly from within the Data Entry Companion.
To do this, you need to enable the Studio Controls from within the Data Entry Companion. You will recognise the Studio controls by the three dots icon.
Edit properties of fields
You can edit the properties of a field by clicking on the three dots in front of the row in the Fields table. For other tables, you can click on the three dots in the column headers to edit the properties
Add fields
You can add Text or Tag fields in two ways:
By using the placeholder in the bottom of the Fields table
By dragging a box around an area of interest and indicating that you want to create a new field
Train
This feature is not yet available