2

As part of a Text Analytics use case, I need to read files from Ariba into Python and then use my already ready model to classify them. However, I am not able to find an API that connects Ariba with Python, anywhere online.

Sanidhya Garodia
  • 45
  • 5

1 Answers1

0

There are a couple of ways to achieve this and the variable here are what documents are we talking about.

Ariba Integration Toolkit [AITK]

  • You need to configure Integration Toolkit in order to extract data from Ariba.
  • To achieve this, find a customer realm and check if AITK is enabled. If yes, you will need to contact the owner of the custeomer realm to know the 'Shared Secret'.
  • By all means refrain from resetting the AITK password without contacting the administrator. This could break you existing integrations if not done right.
  • Well, after you know the 'Shared Secret', extract a dump of documents that you are most interested in. E.g: Contracts and Purchase Orders [POs].
  • Until this point you will not make use your Python language as AITK itself comes with several wrapper code snippets for both UNIX and Windows platforms.
  • After extracting the data dump or file attachment dump you can perform text analysis.
  • I will talk about two usecase here -
  • Use case 1: Getting PO data and analyzing the PO Title, line items details and comments etc. using your Python Text Analytics. You have structured text coming from Ariba already.
  • Use case 2: You extract a contract attachment which is a PDF document. In this case you will use Python OCR to extract text from such PDF document and then use text analytics.

Web Services

  • Next up, you may use the OOTB WebServices to export data from Ariba.
  • Kindly note that WebServices don't have the same level of support to export attachments from Ariba as AITK. Thus AITK scores better in this area.
  • To call Web Services, your client application can be a Python application with API libraries on the top of it.
  • Remember this is SOAP Protocol

APIs

  • Ariba opened up APIs only in the last couple of years
  • All the fancy stuff you would expect from APIs is already there and being constantly improved.
  • The process to get up and running is to request API, get your Customer Admin to approve it and then Ariba company reps to approve it.
  • Due to the lifecycle of approvals inolved before you can start building your client applications, it is better to assess your needs and find the best-fit API. There are a lot of them :)
  • Remember this is REST API
  • You can call these APIs using cURL and of course, your Python programs with HTTP request libraries.
Lalith Kumar Nimmala
  • 11
  • 3