Abonnieren

UiPath Document Understanding

UiPath Document Understanding

Formular-KI

Forms AI is part of the Document Understanding Service and can be used for uploading and processing structured forms with standard layouts and fields.

Create Forms AI

Forms AI is the first extraction method available in Document Understanding Service. Here is more information about how to create a new project in Document Understanding Service.

Once a project is created, you need to follow the next steps for creating a document type using Forms AI within the project.

  • Open your project and click on the +New button (option 1) or select the Forms AI from the Quick Start panel (option 2). See the right side bar for more information about the main concepts.
25602560

A dialog box opens, requesting a name for your document type.

25602560

📘

Hinweis:

Fixed layout forms used with Forms AI can each have a maximum length of five pages.

Importieren von Dokumenten

Once the new Forms AI is created, a new window opens, requiring you to import data. You can import a minimum of two documents and a maximum of twenty documents, each with a maximum of five pages. Drag and drop or browse for the files to upload them.

12801280

After importing all documents, Forms AI automatically detects the fields and their values from the document. You may add or remove fields by using the Edit Field button. More details about using the Edit Field button are presented here.
Automatically extracted fields should also be checked for Content Type accuracy. for example, if a date field was automatically extracted, then the Content Type should be date. Any inaccuracies should be manually corrected.

25602560

Management bar

At the top of the page you can find the management bar. The management bar enables you to perform multiple operations: navigate between documents, delete/restore a document, search/filter documents, run AI model predictions, import, and export documents.

Die folgenden Optionen finden sich in der Verwaltungsleiste:

ItemIconDescription
NavigationnavigatenavigateNavigate between documents that match the active filter.

In between the two arrows, a counter is displayed. It illustrates the number of the current document out of the total number of documents that match the active search/filter.
Search and Search in documentsearchsearchSearch - initiate a search or filter the documents. Filter is also applied when exporting documents. You can filter by words from a document or by document names.

Search in document - initiate a text search inside the document by clicking on the searchsearch or using the shortcut Ctrl + Shift + F
Delete / Restoredeletedelete / restorerestoreDelete or restore a document. Deleted documents can be found under the deleted filter.
ImportimportimportOpen Import data dialog box.
ExportexportexportOpen Export files dialog box.
Document name and typen/aThe name of the currently active document and its type.
DownloaddocumentdocumentThe option is available in the drop-down next to the document name.

Click the icon to download a Zip file containing the original document. Besides the original document, all pages converted internally by Document Manager to .jpeg images are downloaded as well.
Permanently deletepermanently deletepermanently deleteThe option is available in the drop-down next to the document name.

Permanently deletes individual files. The .pdf and all its .jpeg images are deleted from the AI Center dataset and all the metadata is deleted from the database.

When clicking the button, a pop-up message appears asking you if you are sure you want to permanently delete the document. Click OK to continue or Cancel to revert to the previous screen.
PredictpredictpredictRun AI model predictions and display the results.

After configuring Prelabelling, the button is enabled in the management bar. Click it to prelabel the current document.

At the moment, using the Predict option with Public Endpoints prelabels only the first 10 pages of a document. This is a known issue and a fix is in the working. Using the Predict option with ML Skills in AI Center, however, does not impose such a limitation.
PublishpublishpublishPublishes the Forms AI extractor and creates the associated link, available in the project's list of extractors.
SettingssettingssettingsConfigure OCR and Prelabelling settings or access the How to... panel.
The settings button has two available options:
* Settings where you can see the OCR configuration which is automatically populated from the Project's Settings.
Sessionn/aThe name of the current session, found at the top of the page, next to the UiPath Document Understanding logo.

Let's go a little bit deeper in understanding the difference between Delete and Permanently Delete options.

  • The Delete option deletes the files but, not removing them entirely from your project. The deleted files can still be found under the deleted filter from the Search bar and restored by using the Restore option.
  • The Permanently Delete option deletes the selected files without any possibility of restoring them.
    Observe the use of both options in the below GIF:
12801280

The Settings button has two available options:

  • Settings where you can configure the OCR service
  • Anleitungen zu … mit dem Zweck eines Hilfemenüs
25472547

Spaltenfelder

Erstellen eines neuen Spaltenfelds


  1. Klicken Sie auf create_fieldcreate_field im Tabellenabschnitt oben auf der Seite, um ein neues Spaltenfeld hinzuzufügen. Das Fenster Spaltenfeld erstellen wird angezeigt.
  2. Geben Sie einen eindeutigen Namen für das Feld im Feld Eindeutigen Feldnamen eingeben ein. Das Feld akzeptiert keine Großbuchstaben. Darf nur Kleinbuchstaben, Ziffern, Unterstriche _ und Bindestriche - enthalten.
  3. Klicken Sie auf OK.

Bearbeiten eines Spaltenfelds


Click the Edit field edit_fieldedit_field button. The available options for column fields can be found in the table below.

OptionDescription
Field nameThe unique name for the field.
The field does not accept uppercase letters. It can only contain lowercase letters, numbers, underscore _ and dash -.
Content typeThe content type of a field:
string: appropriate for company names or addresses, as well as payment terms, or for any other field where the RPA developer prefers to build the parsing or formatting logic manually, in the RPA workflow.
number: appropriate for amounts or quantities, with intelligent parsing of the decimal/thousands separators.
date: the model parses, formats and unifies the output in a yyyy-mm-dd format.
phone: appropriate for phone numbers. Formatting removes letters and parentheses, and replaces spaces with dashes.
id-no: appropriate for alphanumeric codes, numbers of IDs, it is similar to the string content type, but includes cleaning of any characters coming before a colon :. If the id number you need to extract might contain colon : characters, please use string as content type instead to avoid data loss.
ShortcutThe shortcut key for the field. One or two keys allowed.
Split itemsSelect this checkbox if you want this field to be used as a delimiter between line items or rows in a table. Any line on which this field appears is considered to be a new line item or row in the table. Most commonly, this is used on Line Amount fields on Invoice line items.

Klicken Sie auf Speichern, um Ihre Einstellungen zu speichern.

Grouping table rows is different than in the AI Center Document Manager. Here, the rows are automatically grouped based on the state of the Split items check box on each column fields. This is only relevant for tables with rows that contain multiple lines of text. In this case you must check the Split items check box on any of the fields that have only one line for each table row. For instance, on an invoice, the line item amount would be a typical field on which you might check the Split items option. In the context of Forms AI you would do the same thing on forms.

The example below shows a two line row description for one item. For this case, the description column field doesn't have the Split items option checked, while the other two column fields have the Split items option checked.

25472547

Löschen eines Spaltenfelds

Führen Sie die folgenden Schritte aus, um ein Spaltenfeld zu löschen:

  1. Klicken Sie beim Spaltenfeld, das Sie löschen möchten, auf die Schaltfläche Feld bearbeiten edit_fieldedit_field.
  2. Klicken Sie auf die Schaltfläche Löschen.
  3. Klicken Sie auf OK.
  4. The column field and its associated labelled data are deleted.

Felder

Create a new field

  1. Click create_fieldcreate_field on the right pane in the Fields section. The Create a new regular field window is displayed.
  2. Geben Sie einen eindeutigen Namen für das Feld im Feld Eindeutigen Feldnamen eingeben ein. Das Feld akzeptiert keine Großbuchstaben. Darf nur Kleinbuchstaben, Ziffern, Unterstriche _ und Bindestriche - enthalten.
  3. Klicken Sie auf OK.

Delete all Fields


  1. Click deletedelete in the table section at the top of the page to delete all created fields. Use this function for deleting all fields, including Regular and Column fields, and all the labels on the documents in the current Document Type collection. This action cannot be undone.
  2. Click the Delete button from the Delete all fields dialog box.

Edit a field

Click the Edit field edit_fieldedit_field button. The available options for regular fields can be found in the table below.

OptionDescription
Field nameThe unique name for the field.
The field does not accept uppercase letters. It can only contain lowercase letters, numbers, underscore _ and dash -.
Content typeThe content type of a field:
string: appropriate for company names or addresses, as well as payment terms, or for any other field where the RPA developer prefers to build the parsing or formatting logic manually, in the RPA workflow.
number: appropriate for amounts or quantities, with intelligent parsing of the decimal/thousands separators.
date: the model parses, formats and unifies the output in a yyyy-mm-dd format.
phone: appropriate for phone numbers. Formatting removes letters and parentheses, and replaces spaces with dashes.
id-no: appropriate for alphanumeric codes, numbers of IDs, it is similar to the string content type, but includes cleaning of any characters coming before a colon :. If the id number you need to extract might contain colon : characters, please use string as content type instead to avoid data loss.
ShortcutThe shortcut key for the field. One or two keys allowed.
Multi lineGeneral

Klicken Sie auf Speichern, um Ihre Einstellungen zu speichern.

Löschen eines regulären Felds

Führen Sie die folgenden Schritte aus, um ein reguläres Feld zu löschen:

  1. Klicken Sie beim regulären Feld, das Sie löschen möchten, auf die Schaltfläche Feld bearbeiten edit_fieldedit_field.
  2. Klicken Sie auf die Schaltfläche Löschen.
  3. Klicken Sie auf OK.
  4. The field and its associated labeled data are deleted.

Document View and Labelling

Bei mehrseitigen Dokumenten können Sie natürlich wie bei jedem PDF-Anzeigeprogramm durch die Seiten scrollen. Zum Vergrößern oder Verkleinern halten Sie STRG gedrückt und scrollen mit der Maus.

Sie können Dokumente beschriften, indem Sie die Wortfelder auswählen und sie einem Feld durch Tastendruck zuweisen. Sie können auch mit der rechten Maustaste auf das Wortfeld klicken und die extrahierten Informationen überprüfen.

For more details on how to label documents, visit this page.

Once the model is created, you can verify it in the Extractors tab, under the created project.

19201920

Kontrollkästchen

Checkboxes that are available in Forms AI should be manually labelled for each field. Checkboxes from tables can also be labelled by using the Column Fields option. When a checkbox is labelled in Forms AI, both checked and unchecked boxes should be considered.

Here you can find more detailed information about how to label checkboxes.

You can choose to integrate your Document Understanding project into an RPA workflow by following the steps presented here.

Aktualisiert vor 3 Monaten


Formular-KI


Auf API-Referenzseiten sind Änderungsvorschläge beschränkt

Sie können nur Änderungen an dem Textkörperinhalt von Markdown, aber nicht an der API-Spezifikation vorschlagen.