I needed a way to sync highlights and vocabulary from my Kindle to my Anki. This would help me learn new words, but also remember the interesting parts of the books I read. I wrote up some utility functions in Ruby, it stores a local version of what you've already synced then uses AnkIWeb to sync. It uses a dictionary API to get the definition o…

kindle-sync-to-anki.rb

require 'mechanize'
require 'sqlite3'
require 'wordnet-defaultdb'
require 'wordnet'

LOCAL_DB_PATH = '<yourpath>.db'.freeze
WORDS_TABLE_NAME = 'words_v2'.freeze
HIGHLIGHTS_TABLE_NAME = 'highlights_v2'.freeze

USERNAME = '<your username>'.freeze
PASSWORD = '<your password>'.freeze

# Some parameter the API uses, not sure what it does (might be the card ID?)
MID = 1446933620314

@db = SQLite3::Database.new(LOCAL_DB_PATH)

@agent = Mechanize.new
page = @agent.get("https://ankiweb.net/account/login")

page.form_with(action: "https://ankiweb.net/account/login") do |form|
  form.username = USERNAME
  form.password = PASSWORD
end.click_button

@wordnet = WordNet::Lexicon.new

# Initialize the DB.
if @db.execute("SELECT * FROM sqlite_master WHERE name ='#{WORDS_TABLE_NAME}' and type='table'").empty?
  @db.execute(
    "CREATE TABLE #{WORDS_TABLE_NAME} (
      word VARCHAR(255) NOT NULL,
      synced INTEGER DEFAULT 0
    );"
  )
  puts "Created table '#{WORDS_TABLE_NAME}'..."
end

if @db.execute("SELECT * FROM sqlite_master WHERE name ='#{HIGHLIGHTS_TABLE_NAME}' and type='table'").empty?
  @db.execute(
    "CREATE TABLE #{HIGHLIGHTS_TABLE_NAME} (
      timestamp INTEGER,
      highlight TEXT NOT NULL,
      synced INTEGER DEFAULT 0
    );"
  )
  puts "Created table '#{HIGHLIGHTS_TABLE_NAME}'..."
end

def synced_highlight?(highlight_id)
  rows = @db.execute(
    "SELECT * FROM #{HIGHLIGHTS_TABLE_NAME} WHERE timestamp = ?",
    highlight_id
  )
  !rows.empty?
end

def synced_word?(word)
  rows = @db.execute(
    "SELECT * FROM #{WORDS_TABLE_NAME} WHERE word = ? AND synced = 1",
    word
  )
  !rows.empty?
end

def api_save_card(deck_name, front, back, tag = '')
  @agent.post(
    'https://ankiweb.net/edit/save',
    data: [[front, back], tag],
    deck: deck_name,
    mid: MID
  )
end

def api_get_back_of_card_for_word(word)
  results = @wordnet.lookup_synsets(word.downcase)
  return nil if results.nil? || results.empty?

  back_of_card = ''
  results.each do |synset|
    other_words = synset.words.map(&:to_s)
    back_of_card <<
      "#{synset.part_of_speech}: #{synset.definition} : #{other_words}\n\n"
  end

  back_of_card
end

def sync_highlight(deck_name, text, id)
  puts "Sync highlight ... '#{text[0 .. 140]}...'"
  api_save_card(deck_name, text, '', 'books')
  @db.execute(
    "INSERT INTO #{HIGHLIGHTS_TABLE_NAME} (highlight, synced, timestamp)
    VALUES (?, ?, ?)",
    [text, 1, id]
  )
end

def sync_word(deck_name, word, text)
  api_save_card(deck_name, word, text)
  @db.execute(
    "INSERT INTO #{WORDS_TABLE_NAME} (word, synced)
    VALUES (?, ?)",
    [word, 1]
  )
end

def read_words
  `sqlite3 /Volumes/Kindle/system/vocabulary/vocab.db "SELECT stem FROM WORDS"`
end

def read_highlights
  IO.read('/Volumes/Kindle/documents/My Clippings.txt')
end

def highlight_timestamp_from_line(line)
  time_segment = line.split('|').last
  time_segment.gsub!('Added on', '')
  time_segment.strip!
  Time.parse(time_segment).to_i
end

def sync_highlights
  read_highlights.split('==========')[0..-2].each do |highlight_data|
    source_title,
    meta,
    text = highlight_data.split("\r\n").select { |x| !x.empty? }
    # Don't sync bookmarks.
    next if meta =~ /Your Bookmark on/

    # Generate a UNIX timestamp for the highlight which we'll use to uniquely
    # identify it.
    highlight_timestamp = highlight_timestamp_from_line(meta)
    next if synced_highlight?(highlight_timestamp)
    sync_highlight(source_title, text, highlight_timestamp)
  end
end

def sync_words
  read_words.split("\n").each do |word|
    back = api_get_back_of_card_for_word(word)
    if back.nil? || synced_word?(word)
      puts "Skipped #{word}..."
      next
    end
    sync_word('words', word, back)
    puts "Synced #{word}..."
  end
end

puts "Start syncing words...\n"
sync_words

puts "Start syncing highlights...\n"
sync_highlights