stack_overflow

---
title: "Apache Arrow R Questions on Stack Overflow"
format: html
---

```{r}
#| label: load-packages-and-code
#| include: false
library(httr)
library(dplyr)
library(lubridate)
library(tidyr)
library(DT)
library(rlang)
```

```{r}
#| label: functions
#| include: false
#| warning: false

get_raw_so_data <- function(){

  questions_content <- get_data(
    api_name = "questions",
    url = "https://api.stackexchange.com/2.3/questions?order=desc&sort=activity&tagged=apache-arrow%3Br&site=stackoverflow&filter=!-nt6H9OZ4WW*msaSa)YvngdWhKQ).R9VfXkayFbhnB61(g5UUJbH7f"
  )

  # get raw data
  questions <- questions_content$items %>%
    mutate(last_activity = as_datetime(last_activity_date)) %>%
    # ignore tags as joins get weird in Arrow as it's a list col
    select(-last_activity_date, -tags) %>%
    mutate(retrieved = now())

  # retrieve comments
  comments <- get_so_comments(questions$question_id)

  # get counts of comments
  reply_counts <- comments %>%
    select(question_id = post_id) %>%
    group_by(question_id) %>%
    summarise(comments = n())

  answers <- get_so_answers(questions$question_id)
  answer_counts <- answers %>%
    select(question_id) %>%
    group_by(question_id) %>%
    mutate(answers = n())

  answer_accepted <- answers %>%
    filter(is_accepted) %>%
    select(question_id, is_accepted)

  # add in raw data, reply counts, and whether an answer has been accepted
  left_join(questions, reply_counts, by = "question_id") %>%
    left_join(answer_counts, by = "question_id") %>%
    left_join(answer_accepted, by = "question_id") %>%
    replace_na(list(is_accepted = FALSE, comments = 0, answers = 0))
}

#' Given a vector of question IDs, pull the comments
#'
#' @param question_ids Vector of question IDs
get_so_comments <- function(question_ids){

  comments_content <- get_data(
    api_name = "comments",
    url = paste0(
      "https://api.stackexchange.com/2.3/questions/",
      paste(question_ids, collapse = ";"),
      "/comments?order=desc&sort=creation&site=stackoverflow&filter=!-).qJXDT0Z5I"
    )
  )

  if (nrow(comments_content$items) == 0) {
    return(NULL)
  }
  # here's the table of comments! we can count them
  comments_content$items %>%
    tidyr::unnest(owner) %>%
    select(display_name, account_id, score, post_id, comment_id, creation_date)

}

get_data <- function(api_name, url){
  # message("retrieving ", api_name, " data from Stack Exchange API")

  api_data <- httr::GET(url)

  # if (api_data$status_code == 200) {
  #   message(api_name, " data successfully retrieved")
  # } else {
  #   warning("status code ", api_data$status_code, " when querying", api_name, "API")
  # }

  out <- content(api_data, as = "text") %>%
    jsonlite::fromJSON()

  # message(nrow(out$items) %||% 0, " items retrieved")

  out

}

get_so_answers <- function(question_ids){

  answers_content <- get_data(
    api_name = "answers",
    url = paste0(
      "https://api.stackexchange.com/2.3/questions/",
      paste(question_ids, collapse = ";"),
      "/answers?order=desc&sort=activity&site=stackoverflow&filter=!9Rgp29w2U"
    )
  )

  # here's the table of comments! we can count them
  answers_content$items %>%
    tidyr::unnest(owner) %>%
    select(display_name, account_id, score, is_accepted, question_id, creation_date)
}

```

# Stack Overflow

Stack Overflow questions tagged both with 'apache-arrow' and 'r'.

```{r}
#| label: create-dt
#| echo: false
DT::datatable(
  escape = FALSE,
  get_raw_so_data() %>%
        mutate(issue = paste0("<a href='",link,"' target='_blank'>", title , "</a>")) %>%
        select(issue, last_activity, comments, answers, accepted_answer = is_accepted) %>%
        collect() %>%
        mutate(days_since_last_activity = round(as.numeric(as.duration(interval(last_activity, now())), "days"))) %>%
        select(-last_activity)  
)

```