utility-functions.rb

# encoding: UTF-8
# utility functions for researchr
$:.push(File.dirname($0))
Bibliography_header = "h1. Bibliography\n\n
Also see bibliography by [[abib:start|author]] or by [[kbib:start|keyword]].\n\n
Publications that have their own pages are listed on top, and hyperlinked. Most of these also have clippings and many have key ideas.\n\n"
Home_path = ENV['HOME']
Script_path = File.dirname(__FILE__)

PDF_content_types = [
    "application/pdf",
    "application/x-pdf",
    "application/vnd.pdf",
    "application/text.pdf"
  ]

require 'settings' if File.exists?("#{Script_path}/settings.rb")

# comment the three next lines to use your own gems, instead of the frozen ones, if you don't have OSX 10.7
# or there are other errors with incompatible libraries etc
# Dir.glob(File.join(File.dirname($0), "vendor", "gems", "*", "lib")).each do |lib|
#   $LOAD_PATH.unshift(File.expand_path(lib))
# end

# shows notification on screen. one or two arguments, if one, just shows a message, if two, the first is the title
# notice the path to growl
def growl(title,text='')
  if text == ''
    text = title
    title = ''
  end
  `#{Script_path}/growlnotify -t "#{title}" -m "#{text}"`
end

def log(text)
  File.append("#{Script_path}/log.txt",text)
end

# lookup DOI, return BibTeX using content negotiation
def doi_to_bibtex(doi)
  require 'open-uri'
  doi = doi.downcase.remove(/doi[:>\/]/,'http://','dx.doi.org/').strip
  url = "http://dx.doi.org/#{doi}"
  return try { open(url, "Accept" => "text/bibliography; style=bibtex").read }
end

# a new bibtex filter to recapitalize names with proper unicode
require 'bibtex'
class Fix_namecase < BibTeX::Filter
  def apply(field)
    require 'namecase'

    # only capitalize if all-caps, otherwise preserve to avoid deleting things like "McCoy"
    field = NameCase(Unicode::capitalize(field)) unless field.index(/[a-z]/)

    # fix problem of initials without space between
    field.to_s.gsub(/\.([A-Za-z])/, '. \1')
  end
end

# providesa new function for bibtex entries to generate a nice looking citekey
module BibTeX
  class Entry
    def std_key
      require 'iconv'
      k = names[0]
      k = k.respond_to?(:family) ? k.family : k.to_s
      cstr = Iconv.conv('us-ascii//translit', 'utf-8', k)
      cstr << (has_field?(:year) ? year : '')
      t = title.dup.split.select {|f| f.size > 3}[0]
      cstr << t ? t : ''
      cstr = cstr.downcase.remove(/[^a-zA-Z0-9\-]/)
      return cstr
    end
  end
end

# cleanup bibtex for BibDesk, convert names, clean key etc
def cleanup_bibtex_string(cit)
  require 'latex/decode'
  cit.gsub!(/\@(.+?)\{(.+?)\,(.+?)$/m, '@\1{key,\3')
  b = BibTeX::parse(cit, :filter => :latex)
  b.parse_names
  b[0][:author].convert!(:fix_namecase)
  b[0].key = b[0].std_key
  return b.to_s
end

# a few extra file functions
class File
  class << self

    # adds File.write - analogous to File.read, writes text to filename
    def write(filename, text)
      File.open(filename,"w") {|f| f << text}
    end

    # adds File.append - analogous to File.read, writes text to filename
    def append(filename, text)
      File.open(filename,"a") {|f| f << text + "\n"}
    end

    # find the last file added in directory
    def last_added(path)
      path += "*" unless path.index("*")
      Dir[path].select {|f| test ?f, f}.sort_by {|f|  File.mtime f}.pop
    end

    def replace(path, before, after, newpath = "")
      a = File.read(path)
      a.gsub!(before, after)
      newpath = path if newpath == ""
      File.write(newpath, a)
    end
  end
end

# to make multiple replacements easier, gsubs accepts array of replacements (each replacement is array of from/to)
# takes regexp or string replacement
# for example "stian".gsubs(['s', 'x'], [/^/, "\n"])
# you can also provide a universal "to" string, and a list of "from" strings
# for example "this is my house".gsubs({:all_with => ''}, 'this', /s.y/)
# uses the last function to provide remove, which takes a list of search arguments to remove
# the example above is similar to "this is my house".remove('this', /s.y/)
# also provides remove! destructive function
#
# also adds scan2, which returns named capture groups into compressed hash
class String
  def gsubs!(*searches)
    self.replace(gsubs(*searches))
  end

  def gsubs(*searches)
    if searches[0].kind_of?(Hash)
      args = searches.shift
      all_replace = try { args[:all_with] }
    end
    tmp = self.dup
    searches.each do |search|
      if all_replace
        tmp.gsub!(search, all_replace)
      else
        tmp.gsub!(search[0], search[1])
      end
    end
    return tmp
  end

  def remove(*searches)
    gsubs({:all_with => ''}, *searches)
  end

  def remove!(*searches)
    self.replace(remove(*searches))
  end

  def scan2(regexp) # returns named capture groups into compressed hash, inspired by http://stackoverflow.com/a/9485453/764519
    names = regexp.names
    captures = Hash.new
    scan(regexp).collect do |match|
      nzip = names.zip(match)
      nzip.each do |m|
        captgrp = m[0].to_sym
        captures.add(captgrp, m[1])
      end
    end
    return (captures == {}) ? nil : captures
  end
end


# download a path to a location, require_type is array of acceptable content_types
def dl_file(full_url, to_here, require_type = false)
  require 'open-uri'
  writeOut = open(to_here, "wb")
  url = open(full_url)
  if require_type
    raise NameError unless require_type.index( url.content_type.strip.downcase )
  end
  writeOut.write(url.read)
  writeOut.close
end

# uses online server to check if a file is OA or not
def check_oa(fname)
  require 'open-uri'
  puts "http://reganmian.net/check-oa/#{fname}"
  result = try { open("http://reganmian.net/check-oa/#{fname}").read }
  return (try {result.strip} == "true") ? true : false
end


# writes text to clipboard, using a pipe to avoid shell mangling
# rewritten using osascript for better UTF8 support (from http://www.coderaptors.com/?action=browse&diff=1&id=Random_tips_for_Mac_OS_X)
def pbcopy(text)
  IO.popen("osascript -e 'set the clipboard to do shell script \"cat\"'","w+") {|pipe| pipe << text}
end

# gets text from clipboard
def pbpaste
  a = IO.popen("osascript -e 'the clipboard as unicode text' | tr '\r' '\n'", 'r+').read
  a.strip.force_encoding("UTF-8")
end

# runs pagename through php file from DokuWiki to generate a clean version
def clean_pagename(pname)
  # only send the ones that need it to the external php script
  if pname.downcase =~ /[^0-9a-zA-Z ]/
    # make sure we can manually close the process, otherwise we run out of processes
    ret = ''
    IO.popen("php #{Script_path}/clean_id.php '#{pname}'", 'r+') do |iop|
      iop.close_write
      ret = iop.read
    end
    return ret.strip
  else
    return pname.gsub(" ", "_").downcase
  end
end


# show GUI selector listing all wiki pages, and letting user choose one, or manually enter a new one
def wikipage_selector(title, retfull = false, additional_code = "")
  require 'find'
  require 'pashua'
  include Pashua

  config = "
  *.title = researchr
  cb.type = combobox
  cb.completion = 2
  cb.label = #{title}
  cb.default = start
  cb.width = 220
  cb.tooltip = Choose from the list or enter another name
  db.type = cancelbutton
  db.label = Cancel
  db.tooltip = Closes this window without taking action" + "\n" + additional_code + "\n"

  # insert list of all wiki pages from filesystem into Pashua config
  wpath = "#{Wiki_path}/data/pages/"
  Find.find(wpath) do |path|
    next unless File.file?(path)
    fname = path[wpath.size..-5].gsubs(["/",":"],["_", " "])
    idx = fname.index(":")
    config << "cb.option = #{capitalize_word(fname)}\n" if (path[-4..-1] == ".txt" && path[0] != '_')
  end
  pagetmp = pashua_run config

  pagetmp['cancel'] == 1 ? nil : (retfull ? pagetmp : pagetmp['cb'] )
end


# capitalize the first letter of each word
def capitalize_word(text)
  out = Array.new
  text.split(":").each do |t|
    out << t.split(/ /).each {|word| word.capitalize!}.join(" ")
  end
  out.join(":")
end


# returns nicely formatted citation for a given citekey (very slow, rather used preparsed json file)
def get_citation(citekey)
  require 'bibtex'
  require 'citeproc'

  b = BibTeX.open(Bibliography)
  b.parse_names
  item = b[citekey.to_sym]
  return CiteProc.process(item.to_citeproc, :style => :apa)
end

def utf8safe(text)
  require 'iconv'
  ic = Iconv.new('UTF-8//IGNORE', 'UTF-8')
  return ic.iconv(text + ' ')[0..-2]
end

# wrapper around DokuWiki dwpage tool, inserts page into dokuwiki
def dwpage(page, text, msg = "Automatically added text")
  tmp = Time.now.to_i.to_s
  File.write("/tmp/researcher-#{tmp}.tmp", text)
  `/wiki/bin/dwpage.php -m '#{msg}' commit "/tmp/researcher-#{tmp}.tmp" '#{page}'`
end

# properly format full name, extracted from bibtex
def nice_name(name)
  return "#{name.first} #{name.last}".remove(/[\{\}]/)
end

# properly format list of names for citation
def namify(names)
  return names[0] if names.size == 1
  return names[0] + " et al." if names.size > 3
  names[0..-2].join(", ") + " & " + names[-1].to_s
end

# entire bibliography pre-parsed read in from json
def json_bib()
  require 'json'
  return JSON.parse(File.read(Wiki_path+"/lib/plugins/dokuresearchr/json.tmp"))
end

# given a start of a filename, and an end, looks if there are already any files existing with the filename (pre)01(post)
# increments number with one and returns. used to generate filenames like picture01.png picture02.png etc
def filename_in_series(pre,post)
  existingfile =  File.last_added("#{pre}*#{post}")
  if existingfile
    c = existingfile.scan(/(..)#{post}/)[0][0].to_i
    c += 1
  else
    c = 1
  end

  pagenum = c.to_s
  pagenum = "0" + pagenum if pagenum.size == 1
  return "#{pre}#{pagenum}#{post}", pagenum
end

# enables you to do
#   a = Hash.new
#   a.add(:peter,1)
# without checking if a[:peter] has been initialized yet
# works differently for integers (incrementing number) and other objects (adding a new object to array)
class Object
  def add_safe(var,val)
    if val.class == Fixnum
      if self[var].nil?
        self[var] = val
      else
        self[var] = self[var] + val
      end
    else
      if self[var].nil?
        self[var] = [val]
      else
        self[var] = self[var] + [val]
      end
    end
  end
end

class Hash
  def add_safe(var,val)
    super
  end
  alias :add :add_safe  # we've already used this in the code
end

class Array
  def add_safe(var,val)
    super
  end
end

# calculate SHA-2 hash for a given file
def hashsum(filename)
  require 'digest/sha2'
  hashfunc = Digest::SHA2.new
  File.open(filename, "r") do |io|
    counter = 0
    while (!io.eof)
      readBuf = io.readpartial(1024)
      #       putc '.' if ((counter+=1) % 3 == 0)
      hashfunc.update(readBuf)
    end
  end
  return hashfunc.hexdigest
end

# displays and error message and exits (could optionally log, not implemented right now)
# mainly to enable one-liners instead of if...end
def fail(message)
  growl "Failure!", message
  exit
end

# returns either the value of the block, or nil, allowing things to fail gracefully. easily
# combinable with fail unless
def try(default = nil, &block)
  if defined?(DEBUG)
    yield block
  else
    begin
      yield block
    rescue
      return default
    end
  end
end

# adds a citekey to json right away, must be in BibDesk
def add_to_jsonbib(citekey)
  require 'json'
  require 'citeproc'
  require 'bibtex'

  find = try {BibDesk.search({:for => citekey}) }
  exit unless find && find != []

  bib = find[0].BibTeX_string.get.to_s
  item = BibTeX.parse(bib, {:filter => :latex})[0]
  ax = []
  item.author.each do |a|
    ax << a.last.remove(/[\{\}]/)
  end

  cit = CiteProc.process item.to_citeproc, :style => :apa
  year = try("n.d.") { item.year.to_s }
  year = $1 if year == "n.d." and cit.match(/\((....)\)/)

  json = JSON.parse(File.read(JSON_path))
  json[item.key.to_s] = [namify(ax), year, cit, item.title]
  File.write(JSON_path, JSON.fast_generate(json) )
end


##################################################################
# Scrobblr functions

def send_to_server(path, payload)
  require 'net/http'
  require 'json'

  req = Net::HTTP::Post.new(path, {'Content-Type' => 'application/json'})
  req.body = payload
  response = Net::HTTP.new(Scrobble_server_host, Scrobble_server_port).start { |http| http.request(req) }
end

def submit_citation(bibtex)
  payload = { "bibtex" => bibtex,
    "token"  => Scrobble_token }.to_json
  send_to_server("/citations", payload)
end

def submit_wikipage(citation)
  payload = { "ref_link" => { "url" => Server_path + "/ref:" + citation },
    "citekey"  => citation,
    "token"    => Scrobble_token }.to_json
  send_to_server("/ref_links", payload)
end

def scrobble(citation)
  require 'json'
  payload = { "citekey"  => citation,
    "token"    => Scrobble_token }.to_json
  send_to_server("/scrobbles", payload)
  submit_wikipage(citation)
end