Can't figure out the problem in inputs for Wildcard tagger node


I am using the workflow as shown in the figure below. I want to provide a dictionary(R source) and a document (Table Reader: contain only one document) to Wildcard tagger to the map the text within the document to the tokens in the dictionary and get the bag of words at the end that are related to dimension environment. For this workflow I have created my own node extension CSR tagger. You can see the images below. But my problem right now is - somehow I am not able to provide right inputs to the Wildcard tagger.

My dictionary looks like the following image:

This is how my node extension look like inside Wildcard tagger:

In the Recursive loop, the first input is the document to make my document run through all the tokens, and the second input is the list of tokens which should be processed iteratively. I used Row splitter to sort the tokens by the number of words, therefore only one line is passed on to the wildcard tagger. The rest of the tokens are passed back to the recursive loop.

Does anyone has an idea what is the problem here or what am I doing wrong?

Thanks in advance! Looking forward to your responses. :slight_smile:

I don’t see any document in your workflow. You probably need to first convert your string to a document (String to Document node) and then you can use various nodes which only work on documents.
You could give this a try
hope that helps

The problem is solved, in the first port of wildcard tagger the document to tag should be provided, in above example, the input from table reader should be in first port and the input from rule engine should be in second port of wildcard tagger.

1 Like

This topic was automatically closed 7 days after the last reply. New replies are no longer allowed.