PII Recognition

Just like Named Entity Recognition (NER), you can also perform PII (Personally Identifiable Information) detection. Infact, the default NER entities, is just a super-set of PII entities. For example, Entities.NAME and Entities.SSN are PII entities.

  from rhubarb import DocAnalysis, Entities

  da = DocAnalysis(file_path="./test_docs/employee_enrollment.pdf",
                 boto3_session=session,
                 pages=[1,3])
  resp = da.run_entity(message="Extract all the specified entities from this document.",
                       entities=[Entities.ADDRESS, Entities.SSN])

Sample output

{
    "output": [
        {
            "page": 1,
            "entities": [
                {"SSN": "376 12 1987"},
                {"ADDRESS": "8 Any Plaza, 21 Street"}
            ]
        },
        {
            "page": 3,
            "entities": [
                {"SSN": "791 36 9771"},
                {"ADDRESS": "8 Any Plaza, 21 Street"},
                {"SSN": "824 26 2211"},
                {"ADDRESS": "8 Any Plaza, 21 Street"}
            ]
        }
    ],
    "token_usage": {
        "input_tokens": 3534,
        "output_tokens": 183
    }
}