Question

I would like to use R to extract the speaker out of scripts formatted like in the following example:

“Scene 6: Second Lord: Nay, good my lord, put him to’t; let him have his way. First Lord: If your lordship find him not a hilding, hold me no more in your respect. Second Lord: On my life, my lord, a bubble. BERTRAM: Do you think I am so far deceived in him? Second Lord: Believe it, my lord, in mine own direct knowledge, without any malice, but to speak of him as my kinsman, he’s a most notable coward, an infinite and endless liar, an hourly promise-breaker, the owner of no one good quality worthy your lordship’s entertainment.”

In this example, I would like to extract: (“Second Lord”, “First Lord”, “Second Lord”, “BERTRAM”, “Second Lord”). The rule is obvious: it is the group of words situated between the end of a sentence and a semi-column.

How can I write this in R ?

Answer 1

Maybe something like this:

library(stringr)  
body <- "Scene 6: Second Lord: Nay, good my lord, put him to't; let him have his way. First Lord: If your lordship find him not a hilding, hold me no more in your respect. Second Lord: On my life, my lord, a bubble. BERTRAM: Do you think I am so far deceived in him? Second Lord: Believe it, my lord, in mine own direct knowledge, without any malice, but to speak of him as my kinsman, he's a most notable coward, an infinite and endless liar, an hourly promise-breaker, the owner of no one good quality worthy your lordship's entertainment." 
p <- str_extract_all(body, "[:.?] [A-z ]*:")

# and get rid of extra signs
p <- str_replace_all(p[[1]], "[?:.]", "")
# strip white spaces
p <- str_trim(p)
p
"Second Lord" "First Lord"  "Second Lord" "BERTRAM"     "Second Lord"

# unique players
unique(p)
[1] "Second Lord" "First Lord"  "BERTRAM"  

Explanations of regex: (which are not perfect)

str_extract_all(body, "[:.?] [A-z ]*:") a match is started with either : or . or ? ([:.?]) followed by a whitespace. Any character and whitespace is matched until the next :.

Get position

You can use str_locate_all with the same regex:

str_locate_all(body, "[:.?] [A-z ]*:")