'S3' object has no attribute 'detect_document_text' (python code)

Sounds like you're doing something like this:

s3 = boto3.client("s3")
# maybe some other code here
s3.detect_document_text(...)

But you need to create a textract client:

textract = boto3.client("textract")
# maybe some other code here
response = textract.detect_document_text(
    Document={
        "S3Object": { "Bucket": "mybucketname", "Name": "my-object-name.jpg" }
    }
)

Check out the boto3 documentation for textract, and for detect_document_text() specifically for more info, including the value of the response you'll get back.