mkearney/jeopboty

My Jeopardy Twitter bot

#jeopboty: a jeopardy twitter bot

I scraped Jeopardy clues and answers from an online archive and created a bot-like script to post them to Twitter.

The bot script can be found in the R folder. It's also pasted below.

## save current directory
od <- getwd()

## for now, go to R dir in jeopboty
setwd("~/r/jeopboty/R")

## Reading data and the [correct] Twitter token
## load rtweet
library(rtweet)

## If you haven't created and saved a personal access Twitter API
## token, do it now following these instructions (uncomment next line):
## vignette("auth", "rtweet")

## Make sure you've given the token "write" access (permission)
## which you can check through your Twitter account.
## Uncomment next line to open browser to Twitter's app page.
## browseURL("https://apps.twitter.com")

## If you already saved your token as an environment variable, you can
## fetch it using the get_tokens() function.
token <- get_tokens()

## I have a token I use exclusively for my #jeopboty statuses,
## so I've saved that in my data folder (note: token won't
## appear in my Github repository; nice try!)
token <- readRDS("../data/rtw.rds")

## Get the screen name associated with your token.
my_screen_name <- token$credentials$screen_name

## This should print out YOUR screen name:
my_screen_name

## Read jeopardy data.
jeop <- readRDS("../data/jeopardy.rds")

## Posting the CLUE status on Twitter
## Randomly select a question.
i <- sample(seq_len(NROW(jeop)), 1L)

## Compose CLUE status.
clue <- paste0("Clue: ", jeop$clue[i], " #jeopboty")

## Post CLUE status to Twitter.
post_tweet(clue, token = token)

## Keeping track of questions you've tweeted over time
## Either extract and update or create "used" data object.
if ("used" %in% names(attributes(jeop))) {
    ## Extract used data attribute
    used <- attr(jeop, "used")

    ## Determine clue [number] n post.
    n <- NROW(used[["id"]]) + 1L

    ## Update (append) used object
    used[["id"]][n] <- n
    used[["datetime"]][n] <- Sys.time()
    used[["clue"]][n] <- jeop$clue[i]
    used[["answer"]][n] <- jeop$answer[i]

} else {
    ## Create used object.
    used <- list(
        id = 1,
        datetime = Sys.time(),
        clue = jeop$clue[i],
        answer = jeop$answer[i])
}

## Add "used" data attribute to "jeop" data.
attr(jeop, "used") <- used

## Save data (dropping the row that was just posted.
saveRDS(jeop[-i, ], "../data/jeopardy.rds")

## Posting the ANSWER status on Twitter
## Get recent timeline data for your account.
tw <- get_timeline(my_screen_name)

## Find and return status ID of most recent clue.
status_id <- tw$status_id[grep("^Clue", tw$text)[1]]

## Use the status ID to create quote link.
quote <- paste0("https://twitter.com/",
                my_screen_name,
                "/status/",
                status_id)

## Compose ANSWER status.
answer <- paste0(
    "Answer: ", jeop$answer[i], " #jeopboty ", quote)

## Post ANSWER status to Twitter.
post_tweet(answer, token = token)

## on exit return wd
setwd(od)