discourse/script/import_scripts/google_groups.rb

Ignoring revisions in .git-blame-ignore-revs. Click here to bypass and see the normal blame view.

295 lines
8.0 KiB
Ruby
Raw Normal View History

#!/usr/bin/env ruby
# frozen_string_literal: true
require "bundler/inline"
gemfile(true) do
source "https://rubygems.org"
gem "webdrivers"
gem "colored2"
end
require "fileutils"
require "optparse"
require "yaml"
DEFAULT_OUTPUT_PATH = "/shared/import/data"
DEFAULT_COOKIES_TXT = "/shared/import/cookies.txt"
ABORT_AFTER_SKIPPED_TOPIC_COUNT = 10
def driver
@driver ||=
begin
chrome_args = ["disable-gpu"]
chrome_args << "headless" unless ENV["NOT_HEADLESS"] == "1"
chrome_args << "no-sandbox" if inside_container?
options = Selenium::WebDriver::Chrome::Options.new(args: chrome_args)
Selenium::WebDriver.for(:chrome, options: options)
end
end
def inside_container?
File.foreach("/proc/1/cgroup") { |line| return true if line.include?("docker") }
false
end
MAX_GET_RETRIES = 5
MAX_FIND_RETRIES = 3
def get(url)
begin
retries ||= 0
driver.get(url)
rescue Net::ReadTimeout
sleep retries
retry if (retries += 1) < MAX_GET_RETRIES
end
end
2018-10-30 20:30:14 -04:00
def extract(css, parent_element = driver)
begin
retries ||= 0
2018-10-30 20:30:14 -04:00
parent_element.find_elements(css: css).map { |element| yield(element) }
rescue Net::ReadTimeout, Selenium::WebDriver::Error::StaleElementReferenceError
sleep retries
retry if (retries += 1) < MAX_FIND_RETRIES
end
end
2018-10-30 20:30:14 -04:00
def find(css, parent_element = driver)
begin
retries ||= 0
2018-10-30 20:30:14 -04:00
parent_element.find_element(css: css)
rescue Net::ReadTimeout, Selenium::WebDriver::Error::ElementNotInteractableError
sleep retries
retry if (retries += 1) < MAX_FIND_RETRIES
end
end
def base_url
if @domain.nil?
"https://groups.google.com/forum/?_escaped_fragment_=categories"
else
"https://groups.google.com/a/#{@domain}/forum/?_escaped_fragment_=categories"
end
end
def crawl_topics
1
.step(nil, 100)
.each do |start|
url = "#{base_url}/#{@groupname}[#{start}-#{start + 99}]"
get(url)
begin
exit_with_error(<<~TEXT.red.bold) if start == 1 && find("h2").text == "Error 403"
Unable to find topics. Try running the script with the "--domain example.com"
option if you are a G Suite user and your group's URL contains a path with
your domain that looks like "/a/example.com".
TEXT
rescue Selenium::WebDriver::Error::NoSuchElementError
# Ignore this error. It simply means there wasn't an error.
end
2018-10-30 20:30:14 -04:00
topic_urls =
extract(".subject a[href*='#{@groupname}']") do |a|
a["href"].sub("/d/topic/", "/forum/?_escaped_fragment_=topic/")
end
2018-10-30 20:30:14 -04:00
break if topic_urls.size == 0
topic_urls.each do |topic_url|
crawl_topic(topic_url)
# abort if this in an incremental crawl and there were too many consecutive, skipped topics
if @finished && @skipped_topic_count > ABORT_AFTER_SKIPPED_TOPIC_COUNT
puts "Skipping all other topics, because this is an incremental crawl.".green
return # rubocop:disable Lint/NonLocalExitFromIterator
end
end
end
end
def crawl_topic(url)
skippable = @scraped_topic_urls.include?(url)
# Skip this topic if there were already too many consecutive, skipped topics.
# Otherwise we have to look if there are new messages in the topic.
if skippable && @skipped_topic_count > ABORT_AFTER_SKIPPED_TOPIC_COUNT
puts "Skipping".green << " #{url}"
return
end
puts "Scraping #{url}"
get(url)
2022-02-27 21:20:58 -05:00
messages_crawled = false
extract(".subject a[href*='#{@groupname}']") do |a|
[a["href"].sub("/d/msg/", "/forum/message/raw?msg="), a["title"].empty?]
end.each do |msg_url, might_be_deleted|
2022-02-27 21:20:58 -05:00
messages_crawled |= crawl_message(msg_url, might_be_deleted)
end
2022-02-27 21:20:58 -05:00
@skipped_topic_count = skippable && messages_crawled ? 0 : @skipped_topic_count + 1
@scraped_topic_urls << url
rescue StandardError
puts "Failed to scrape topic at #{url}".red
raise if @abort_on_error
end
def crawl_message(url, might_be_deleted)
get(url)
filename = File.join(@path, "#{url[%r{#{@groupname}/(.+)}, 1].sub("/", "-")}.eml")
content = find("pre")["innerText"]
if !@first_message_checked
@first_message_checked = true
exit_with_error(<<~TEXT.red.bold) if content.match?(/From:.*\.\.\.@.*/i) && !@force_import
It looks like you do not have permissions to see email addresses. Aborting.
Use the --force option to import anyway.
TEXT
end
old_md5 = Digest::MD5.file(filename) if File.exist?(filename)
File.write(filename, content)
old_md5 ? old_md5 != Digest::MD5.file(filename) : true
rescue Selenium::WebDriver::Error::NoSuchElementError
if might_be_deleted
puts "Message might be deleted. Skipping #{url}"
else
puts "Failed to scrape message at #{url}".red
raise if @abort_on_error
end
rescue StandardError
puts "Failed to scrape message at #{url}".red
raise if @abort_on_error
end
def login
puts "Logging in..."
get("https://google.com/404")
add_cookies("myaccount.google.com", "google.com")
get("https://myaccount.google.com/?utm_source=sign_in_no_continue")
begin
wait_for_url { |url| url.start_with?("https://accounts.google.com") }
rescue Selenium::WebDriver::Error::TimeoutError
exit_with_error("Failed to login. Please check the content of your cookies.txt".red.bold)
end
end
def add_cookies(*domains)
File
.readlines(@cookies)
.each do |line|
parts = line.chomp.split("\t")
if parts.size != 7 || !domains.any? { |domain| parts[0] =~ /^\.?#{Regexp.escape(domain)}$/ }
next
end
driver.manage.add_cookie(
domain: parts[0],
httpOnly: "true".casecmp?(parts[1]),
path: parts[2],
secure: "true".casecmp?(parts[3]),
expires: parts[4] == "0" ? nil : DateTime.strptime(parts[4], "%s"),
name: parts[5],
value: parts[6],
)
end
end
def wait_for_url
wait = Selenium::WebDriver::Wait.new(timeout: 5)
wait.until { yield(driver.current_url) }
end
def exit_with_error(*messages)
STDERR.puts messages
exit 1
end
def crawl
start_time = Time.now
status_filename = File.join(@path, "status.yml")
if File.exist?(status_filename)
yaml = YAML.load_file(status_filename)
@finished = yaml[:finished]
@scraped_topic_urls = yaml[:urls]
else
@finished = false
@scraped_topic_urls = Set.new
end
@skipped_topic_count = 0
login
begin
crawl_topics
@finished = true
ensure
File.write(status_filename, { finished: @finished, urls: @scraped_topic_urls }.to_yaml)
end
elapsed = Time.now - start_time
puts "", "", "Done (%02dh %02dmin %02dsec)" % [elapsed / 3600, elapsed / 60 % 60, elapsed % 60]
end
def parse_arguments
puts ""
# default values
@force_import = false
@abort_on_error = false
@cookies = DEFAULT_COOKIES_TXT if File.exist?(DEFAULT_COOKIES_TXT)
parser =
OptionParser.new do |opts|
opts.banner = "Usage: google_groups.rb [options]"
opts.on("-g", "--groupname GROUPNAME") { |v| @groupname = v }
opts.on("-d", "--domain DOMAIN") { |v| @domain = v }
opts.on("-c", "--cookies PATH", "path to cookies.txt") { |v| @cookies = v }
opts.on("--path PATH", "output path for emails") { |v| @path = v }
opts.on("-f", "--force", "force import when user isn't allowed to see email addresses") do
@force_import = true
end
opts.on("-a", "--abort-on-error", "abort crawl on error instead of skipping message") do
@abort_on_error = true
end
opts.on("-h", "--help") do
puts opts
exit
end
end
begin
parser.parse!
rescue OptionParser::ParseError => e
exit_with_error(e.message, "", parser)
end
mandatory = %i[groupname cookies]
missing = mandatory.select { |name| instance_variable_get("@#{name}").nil? }
if missing.any?
exit_with_error("Missing arguments: #{missing.join(", ")}".red.bold, "", parser, "")
end
exit_with_error("cookies.txt not found at #{@cookies}".red.bold, "") if !File.exist?(@cookies)
@path = File.join(DEFAULT_OUTPUT_PATH, @groupname) if @path.nil?
FileUtils.mkpath(@path)
end
parse_arguments
crawl