Connect to Intelligent Document Processing

DocuWare Intelligent Document Processing (IDP) provides functions with artificial intelligence techniques for processing and managing documents:

  • Classification
    For example documents can be classified by document type

  • Extraction
    Data is automatically extracted from almost all document types. This includes invoices, contracts, HR documents and emails. For invoices, also line items can be recognized and extracted.

  • Extraction with handwriting
    IDP recognizes handwritten text in images or documents and extracts it into editable text.

This article describes how to connect DocuWare with Intelligent Document Processing via a DocuWare workflow.

1. Create web service connection

Go to DocuWare Configuration > Web Services and create a web service connection.

Add a name e.g. “IDP” and the URL of the IDP service https://api.natif.ai. Natif.ai is a DocuWare-owned service.

connect to IPD_1(1)

2. Add a web service activity

Go the DocuWare Workflow Manager and add an activity of type web service to the workflow with which you want to use IDP.

More information about a web service activity in DocuWare Workflow Designer

3. Import specification file with the endpoints

Select the web service you have created in step 1, here “IDP”

Download the OpenAPI specification file (.json) of the IDP service you want to use, for example extraction here
https://platform.natif.ai/api-hub/workflow/invoice_extraction/documentation.

The specification file provides the endpoints so that you can use the feature “extraction”.

connect to IPD_2

Import the OpenAPI specification file (.json) into the web service. Once imported, all available endpoints are listed in the dropdown.

Select the extraction endpoint, e.g. POST /processing/invoice_extraction

connect to IPD_3(1)

Add the desired query parameters, e.g. add:

  • include: extractions

  • wait_for: 60

4. Enter the API key

Switch to IDP to generate an API Key. Enter the URL https://platform.natif.ai/api-keys and create a new API Key.

Add a name and an expiration date and copy the secret.

Connect to IPD
Go back to Docuware Workflow Manager and enter the key in the request header of the web service activity:  

  • Authorization: ApiKey <secret>
    and

  • Replace <secret> with the secret from IDP

    connect to IPD_5

5. Configure HTTP Body  

In the HTTP Body tab of the web service activity, enable the checkbox Attach document.

Choose in original format or as PDF without annotations:

  • The In Original Format option attaches each file from the document individually.

  • The As PDF Without Annotations option generates and attaches a single PDF file containing all the files from the document

connect to IPD_6

6. Send HTTP Request

Copy any sample PDF file into the folder %AppData%\DocuWare\Workflow Designer\OpenApi\

Rename the file to “testrequest.pdf”. This file will be sent as part of the test http request.

Press Send and wait for the response. The test document is sent to the IDP service and is extracted. You should get a response from the IPD service.

connect to IPD_7

7. Assign the response data

Switch to the tab Data Assign. Select the extraction data points from the JSON response and assign them to your index fields, e.g. vendor.name.value, date.value, number.value

connect to IPD_8

Extract and process line items and table data (starting with DocuWare 7.11)

For this, select the appropriate extractions data point from the JSON response and assign it to your index table, e.g. line_item

connect to IPD_9

connect to IPD_10

Then, map each data point of the line items to the corresponding column in the index table, e.g. description.value, quantity.value, unit_price.value

connect to IPD_11

On the tab HTTP Response Status Codes change the Timeout to 60 seconds.

connect to IPD_12

Limitations

DocuWare IDP has some default limitations e.g. the service processes not more than 10 documents per second. Consider this when designing your workflow and make sure that the requests do not exceed these limits.

See limitations of DocuWare IDP.