ブラウザから保存したTogetterさんのページからツイートを抽出してタブ区切りファイルとして整形する

togetter-tweets.rb

#!/usr/bin/ruby
#
# Copyright 2023 zunda <zundan at gmail.com>
#
# Permission is granted for use, copying, modification, distribution,
# and distribution of modified versions of this work as long as the
# above copyright notice is included.
#
require 'time'

Dir.glob("togetter-*.html") do |src|
  File.read(src).scan(%r|<div class="list_box type_tweet impl_profile" data-index="(\d+)">(.*?)</div>|m).each do |i, entry|
    idx = Integer(i)
    username = entry.scan(%r|<span class="status_name">@(.*?)</span>|).flatten.first
    time = Time.at(Integer(entry.scan(%r|<a class="link" .* data-timestamp="(\d+)"|).flatten.first))

    text = entry.scan(%r|<p class="tweet">(.*?)</p>|m).flatten.first.chomp
    text.gsub!(%r|<img draggable="false" class="emoji" alt="(.*?)".*?>|, '\1')
    text.gsub!(%r|<(\w+).*?>.*?</\1>|m, "")
    text.gsub!(/&gt;/, ">")
    text.gsub!(/&lt;/, "<")
    text.gsub!(/&nbsp;|\t|　/, " ")
    text.gsub!(/\s+/, " ")

    puts [idx, time.utc.iso8601, username, text].join("\t")
  end
end