Add paginate_req_perform_continue()

[mgirlich]

Closes #298.

devtools::load_all("~/GitHub/httr2/")
#> ℹ Loading httr2
page_size <- 7

req_pokemon <- request("https://pokeapi.co/api/v2/pokemon") %>%
  req_url_query(limit = page_size) %>%
  req_verbose(header_resp = FALSE) %>% 
  req_paginate_next_url(
    next_url = function(resp) resp_body_json(resp)[["next"]],
    n_pages = function(resp) {
      total <- resp_body_json(resp)$count
      ceiling(total / page_size)
    }
  )

responses <- paginate_req_perform(req_pokemon, max_pages = 2)
#> -> GET /api/v2/pokemon?limit=7 HTTP/2
#> -> Host: pokeapi.co
#> -> User-Agent: httr2/0.2.3.9000 r-curl/5.0.2 libcurl/8.1.2
#> -> Accept: */*
#> -> Accept-Encoding: deflate, gzip
#> -> 
#> -> GET /api/v2/pokemon?offset=7&limit=7 HTTP/2
#> -> Host: pokeapi.co
#> -> User-Agent: httr2/0.2.3.9000 r-curl/5.0.2 libcurl/8.1.2
#> -> Accept: */*
#> -> Accept-Encoding: deflate, gzip
#> ->

# oh, I actually wanted 3 pages, not 2
responses_all <- paginate_req_perform_continue(max_pages = 3)
#> -> GET /api/v2/pokemon?offset=14&limit=7 HTTP/2
#> -> Host: pokeapi.co
#> -> User-Agent: httr2/0.2.3.9000 r-curl/5.0.2 libcurl/8.1.2
#> -> Accept: */*
#> -> Accept-Encoding: deflate, gzip
#> ->

^{Created on 2023-09-05 with reprex v2.0.2}

[mgirlich]

Maybe this should go into a list variable?

the$last_pagination <- list(
  request = NULL,
  responses = NULL,
  page = NULL,
  n_pages = NULL,
  max_pages = NULL
)

[mgirlich]

current is actually not the current page as I though but whether this is the current progress bar or not.

[mgirlich]

Should we also export last_pagination_page()?

[hadley]

Would it be better to copy the multi_req_perform() interface here? i.e. paginate_req_perform() should always succeed but sometimes return an error object? Or maybe in this case it's sufficient to return a shorter list than expected with a warning.

You could use the same technique as in pool_run() to handle user interrupts.

[hadley]

Superseded by #353