Add other antiword parameters about antiword HOT 2 CLOSED

NLP pipelines use antiword R package, for example the
readtext::readtext() function. readtext() has a general parameter "...":

...: additional arguments passed through to low-level file reading
          function, such as ‘file’, ‘fread’, etc.  Useful for
          specifying an input encoding option, which is specified in
          the same was as it would be give to ‘iconv’.  See the
          Encoding section of file for details.

This could be used to pass -w 0 to antiword. Currently, a workaround to readtext(file=full_path_ext) is this:

 if(file_ext=="doc"){
    file_text<-system2(command = "antiword", 
       args = paste0(" -w 0 ", full_path_ext), stdout = TRUE )  
    } else {
    file_text<-readtext(file=full_path_ext)
    }

This workaround does not always work as help(system2) says:

Output lines of more than 8095 bytes will be split.

The problem with readtext and antiword without the w parameter is that NLP pipelines often fail with additional line breaks added by antiword.
For example, take the sentence "I spoke to Mr. Walking in the street, I saw him." If antiword splits the sentence after Mr. with a new line, most NLP pipelines will believe there are 2 sentences:

1. "I spoke to Mr."
2. "Walking in the street, I saw him."
   In the original sentence, "Mr. Walking" is a named entity recognised by NLP. In the second sentence, "walking" is the verb "to walk" of which "I" is the subject.

from antiword.

Can you send a PR?

from antiword.