Builds a token dataframe from the text OCRed by Document AI (DAI) in an asynchronous request. Rows are tokens, in the order DAI proposes to read them. Columns are location variables such as page coordinates and block bounding box numbers.

build_token_df(object, type = "sync")



either a HTTP response object from dai_sync() or the path to a JSON file from dai_async().


one of "sync" or "async" depending on the function used to process the original document.


a token data frame


The location variables are: token, start index, end index, confidence, left boundary, right boundary, top boundary, bottom boundary, page number, and block number. Start and end indices refer to character position in the string containing the full text.


if (FALSE) {
resp <- dai_sync("file.pdf")
token_df <- build_token_df(resp)

token_df <- build_token_df("pdf_output.json", type = "async")