fix twitter scraper

This commit is contained in:
byte[] 2020-06-23 21:26:12 -04:00
parent b33dece707
commit d0e0f3dc00

View file

@ -1,5 +1,5 @@
defmodule Philomena.Scrapers.Twitter do defmodule Philomena.Scrapers.Twitter do
@gt_regex ~r|document.cookie = decodeURIComponent\("gt=(\d+);| @gt_regex ~r|gt=(\d+?);|
@url_regex ~r|\Ahttps?://(?:mobile\.)?twitter.com/([A-Za-z\d_]+)/status/([\d]+)/?| @url_regex ~r|\Ahttps?://(?:mobile\.)?twitter.com/([A-Za-z\d_]+)/status/([\d]+)/?|
@script_regex ~r|<script type="text/javascript" .*? src="(https://abs.twimg.com/responsive-web/web/main\.[\da-z]+\.js)">| @script_regex ~r|<script type="text/javascript" .*? src="(https://abs.twimg.com/responsive-web/web/main\.[\da-z]+\.js)">|
@bearer_regex ~r|"(AAAAAAAAAAAAA[^"]*)"| @bearer_regex ~r|"(AAAAAAAAAAAAA[^"]*)"|
@ -38,7 +38,7 @@ defmodule Philomena.Scrapers.Twitter do
def api_response!(url) do def api_response!(url) do
[user, status_id] = Regex.run(@url_regex, url, capture: :all_but_first) [user, status_id] = Regex.run(@url_regex, url, capture: :all_but_first)
mobile_url = "https://mobile.twitter.com/#{user}/status/#{status_id}" page_url = "https://twitter.com/#{user}/status/#{status_id}"
api_url = api_url =
"https://api.twitter.com/2/timeline/conversation/#{status_id}.json?tweet_mode=extended" "https://api.twitter.com/2/timeline/conversation/#{status_id}.json?tweet_mode=extended"
@ -46,8 +46,7 @@ defmodule Philomena.Scrapers.Twitter do
url = "https://twitter.com/#{user}/status/#{status_id}" url = "https://twitter.com/#{user}/status/#{status_id}"
{gt, bearer} = {gt, bearer} =
Philomena.Http.get!(mobile_url) Philomena.Http.get!(page_url)
|> Map.get(:body)
|> extract_guest_token_and_bearer() |> extract_guest_token_and_bearer()
Philomena.Http.get!(api_url, [{"Authorization", "Bearer #{bearer}"}, {"x-guest-token", gt}]) Philomena.Http.get!(api_url, [{"Authorization", "Bearer #{bearer}"}, {"x-guest-token", gt}])
@ -60,8 +59,9 @@ defmodule Philomena.Scrapers.Twitter do
|> Map.put("url", url) |> Map.put("url", url)
end end
defp extract_guest_token_and_bearer(page) do defp extract_guest_token_and_bearer(%Tesla.Env{body: page, headers: headers}) do
[gt] = Regex.run(@gt_regex, page, capture: :all_but_first) [{_, gt}] = Enum.filter(headers, fn {k, v} -> k == "set-cookie" and String.starts_with?(v, "gt=") end)
[gt] = Regex.run(@gt_regex, gt, capture: :all_but_first)
[script] = Regex.run(@script_regex, page, capture: :all_but_first) [script] = Regex.run(@script_regex, page, capture: :all_but_first)
%{body: body} = Philomena.Http.get!(script) %{body: body} = Philomena.Http.get!(script)