feat(idp-extraction-connector): update the idp connector to accept documents

[reiballa]

Description

This PR introduces support for documents as input for the outbound IDP extraction connector.

How to test locally:

In order to test this, you need to :

• configure an AWS access key and secret,
• upload a pdf document in the local document store. See https://docs.camunda.io/docs/next/apis-tools/camunda-api-rest/specifications/create-document. After uploading the document, save the reference json.
• create a bpmn diagram with the new idp-connector element template and deploy it.
• run the diagram with this as variable:

{
  "baseRequest": {
    "authentication": {
      "accessKey": "<<your access key>",
      "secretKey": "<<your access secret>>"
    },
    "configuration": {
      "region": "us-east-1"
    }
  },
  "input": {
    "extractionEngineType": "AWS_TEXTRACT",
    "document": <<the json that was returned when uploading the document>>
    },
    "converseData": {
      "modelId": "anthropic.claude-3-5-sonnet-20240620-v1:0",
      "maxTokens": 512,
      "temperature": 0.5,
      "topP": "0.9"
    },
    "taxonomyItems": [
      {
        "name": "sum",
        "prompt": "the total amount that was paid for this invoice"
      },
      {
        "name": "supplier",
        "prompt": "who provided the goods or services"
      }
    ]
  }
}

You can update the taxonomy items to something that makes more sense to the pdf document that you have uploaded.

Related issues

closes #3824

Checklist

• PR has a milestone or the no milestone label.

[mathias-vandaele]

Great work, one minor comment

[mathias-vandaele]

We usually prefer to use tooltip over description