quickcite 1.0.0

data/Gemfile

@@ -0,0 +1,6 @@
+source :rubygems
+
+gem "nokogiri"
+gem "highline"
+
+gemspec

data/Gemfile.lock

@@ -0,0 +1,29 @@
+PATH
+  remote: .
+  specs:
+    quickcite (1.0.0)
+      bibtex-ruby (>= 2.0)
+      json
+
+GEM
+  remote: http://rubygems.org/
+  specs:
+    bibtex-ruby (2.0.12)
+      latex-decode (>= 0.0.6)
+      multi_json (~> 1.3)
+    highline (1.6.13)
+    json (1.7.3)
+    latex-decode (0.0.12)
+      unicode (~> 0.4)
+    multi_json (1.3.6)
+    nokogiri (1.5.5)
+    unicode (0.4.2)
+
+PLATFORMS
+  ruby
+
+DEPENDENCIES
+  bundler
+  highline
+  nokogiri
+  quickcite!

data/Rakefile

@@ -0,0 +1,12 @@
+require 'bundler'
+require 'rake/testtask'
+
+Bundler::GemHelper::install_tasks
+
+Rake::TestTask.new do |t|
+  t.libs << "test"
+  t.test_files = FileList['test/test*.rb']
+  t.verbose = true
+end
+
+task :default => [:build]

data/bin/quickcite

@@ -0,0 +1,17 @@
+#!/usr/bin/env ruby
+
+require 'quickcite'
+
+flags = {}
+optparse = OptionParser.new do|opts|
+ opts.banner = "Usage: quickrite -b <bibtex file> latex1 latex2 ... "
+
+ flags[:bibtex] = nil
+ opts.on( '-b', '--bibtex FILE', 'Extract/generate references to FILE' ) do|file|
+   flags[:bibtex] = file
+ end
+end
+
+optparse.parse!
+
+QuickCite::run(ARGV, flags[:bibtex])

data/lib/quickcite.rb

@@ -0,0 +1,93 @@
+#!/usr/bin/env ruby
+
+require "rubygems"
+require "bundler/setup"
+
+require "highline/system_extensions"
+require "bibtex"
+require "optparse"
+require "readline"
+require "quickcite/citeulike"
+require "quickcite/dblp"
+
+module QuickCite
+  CITE_REGEX = /\\cite\{([^\}]+)\}/
+  SPLIT_REGEX = /([A-Z]+[a-z]+|[0-9]+|[^\\p]+)/
+  
+  # Convert a citation reference (e.g "PowerPiccolo2008") into a
+  # list of strings appropriate for submission to a search engine.
+  #
+  # By default the citation is split on lowercase->uppercase transitions.
+  def self.cite_to_query(cite)
+    cite.scan(SPLIT_REGEX).map { |w| w[0] }
+  end
+
+  def self.ask_user(cite, result_list)
+    puts "Result to use for \{#{cite}\}: "
+    puts "  (0) Skip this citation"
+    result_list.each_with_index do |r, idx|
+      puts "  (#{idx + 1}) #{r.title}"
+      puts "      #{r.authors.join(', ')}"
+    end
+    
+    c = HighLine::SystemExtensions::get_character.chr.to_i
+    if c > 0 && c <= result_list.length then
+      return result_list[c - 1]
+    end
+  end
+
+  class Main
+    include QuickCite
+    def initialize(latex_files, bibtex_file)
+      @source = DBLP.new
+      @bib = BibTeX.open(bibtex_file, :include => [:meta_content])
+        
+      latex_files.each do|f|
+        puts("Processing... #{f}")
+        process_latex(f)
+      end
+      
+      puts("Writing bibtex...")
+      outfile = open(bibtex_file, "w")
+      outfile.write(@bib.to_s)
+      outfile.close
+    end
+    
+    def update_bibtex(cite, entry)
+      e = BibTeX.parse(@source.bibtex(entry)).first
+      e.key = cite
+      @bib << e
+    end
+
+    def process_cite(cite)
+      if @bib.has_key?(cite) then
+        puts("Skipping matched reference #{cite}")
+      else
+        puts("Missing reference for #{cite}")
+        query = QuickCite.cite_to_query(cite)
+        results = @source.search(query)
+        accepted = QuickCite.ask_user(cite, results)
+        if accepted == nil
+          puts "Skipping update for reference #{cite}"
+        else
+          puts "Updating bibtex for #{cite} with result: \n#{accepted.title}"
+          update_bibtex(cite, accepted)
+        end
+      end
+    end
+
+    def process_latex(f)
+      latex = File.read(f)
+      latex.scan(CITE_REGEX) do|m|
+        m[0].split(",").map do|c|
+          process_cite(c)
+        end
+      end
+      puts(f)
+    end
+  end
+
+  def self.run(latex_files, bibtex_file)
+    Main.new(latex_files, bibtex_file)
+  end
+end

data/lib/quickcite/citeulike.rb

@@ -0,0 +1,29 @@
+#!/usr/bin/env ruby
+#
+# Searches and downloads bibtex entries from CiteULike (citeulike.org)
+#
+# Author:: Russell Power (power@cs.nyu.edu)
+
+require "net/http"
+require "uri"
+require "quickcite/json"
+
+module QuickCite
+  class CiteULike
+    HOST = "http://www.citeulike.org"
+    SEARCH_BASE = "/json/search/all"
+    BIBTEX_OPTIONS = "do_username_prefix=0&key_type=4&incl_amazon=1&clean_urls=1&smart_wrap=1&q="
+ 
+    def search(query)
+      http = Net::HTTP.new(HOST)
+      request = Net::HTTP::Get.new(SEARCH_BASE)
+      request.set_form_data({ "q" => query })
+      response = http.request(request)
+      js = JSONUtil.Parse(response.body)
+      return js
+    end
+    
+    def bibtex(identifier)
+    end
+  end
+end

data/lib/quickcite/dblp.rb

@@ -0,0 +1,49 @@
+#!/usr/bin/env ruby
+#
+# Searches and downloads bibtex entries from CiteULike (citeulike.org)
+#
+# Author:: Russell Power (power@cs.nyu.edu)
+
+require "net/http"
+require "nokogiri"
+require "quickcite/json"
+require "uri"
+
+module QuickCite
+  class DBLP
+    # NB -- DBLP seems to expect query arguments in a certain order - 
+    # the q= argument has to come first.
+    SEARCH_BASE = "http://www.dblp.org/search/api/?"
+    SEARCH_SUFFIX = "&c=4&f=0&format=json&h=10"
+
+    Result = Struct.new("Result", :title, :authors, :url)
+    def search(query)
+      #      json = JSONUtil.parse(File.read("test/power-piccolo.json"))
+      query_str = "q=" + URI::escape(query.join(" "))
+      uri = URI::parse(SEARCH_BASE + query_str + SEARCH_SUFFIX)
+      
+      response = Net::HTTP::get(uri)
+      json = JSONUtil.parse(response)
+
+      json.result.hits.hit.map do |h|
+        Result.new(
+        h.title["dblp:title"].text,
+        h.title["dblp:authors"]["dblp:author"].to_a,
+        h.url)
+      end
+    end
+
+    def bibtex(result)
+      #      dom = Nokogiri.Slop(open("test/power-piccolo.html"))
+      puts("RESULT #{result.url}")
+      dom = Nokogiri.Slop Net::HTTP.get(URI::parse(result.url))
+      entries = dom.html.body.pre
+      case entries
+        when Nokogiri::XML::NodeSet
+          return entries[0].to_str
+        else
+          return entries.to_str
+      end
+    end
+  end
+end

data/lib/quickcite/json.rb

@@ -0,0 +1,46 @@
+# JSON utilities.
+require "json"
+
+module QuickCite
+  module JSONUtil
+    class JSONDict
+      extend Forwardable
+      include Enumerable
+      attr_reader :dict
+      
+      def initialize(d)
+        @dict = d
+      end
+      
+      def method_missing name, *args
+        return JSONUtil._wrap(@dict[name.to_s])
+      end
+      
+      def [](key)
+        JSONUtil._wrap(@dict[key])
+      end
+      
+      def each
+        case @dict
+          when Hash
+            @dict.each_key { |k| yield self[k] }
+          when Array
+            @dict.each_index { |k| yield self[k] }
+         end
+      end
+    end
+
+    def self._wrap(value)
+      case value
+        when Hash, Array
+          JSONDict.new(value)
+        else
+          value
+        end
+    end
+ 
+    def self.parse(str)
+      d = JSONDict.new(JSON.parse(str))
+    end
+  end
+end

data/lib/quickcite/version.rb

@@ -0,0 +1,3 @@
+module QuickCite
+  VERSION = "1.0.0"
+end

data/quickcite.gemspec

@@ -0,0 +1,16 @@
+require File.expand_path("../lib/quickcite/version", __FILE__)
+
+spec = Gem::Specification.new do |s|
+  s.platform = Gem::Platform::RUBY
+  s.summary = "Efficient citation lookup and management"
+  s.name = "quickcite"
+  s.authors = ["Russell Power"]
+  s.version = QuickCite::VERSION
+  s.email = "power@cs.nyu.edu"
+  s.require_path = "lib"
+  s.files = `git ls-files`.split("\n")
+  
+  s.add_dependency "bibtex-ruby", ">=2.0"
+  s.add_dependency "json"
+  s.add_development_dependency "bundler"
+end

data/test/dean.html

@@ -0,0 +1,18 @@
+<html><head><title>DBLP BibTeX Record 'journals/cacm/DeanG10'</title><link href="http://dblp.uni-trier.de/dblp.css" rel="stylesheet" type="text/css" /></head><body>
+<table width="100%"><tr><td align="left"><a href="http://dblp.uni-trier.de/db/index.html"><img alt="dblp.uni-trier.de" src="http://dblp.uni-trier.de/db/Logo.gif" border="0" height="60" width="170"/></a></td><td align="center"><a href="http://www.dagstuhl.de/"><img alt="www.dagstuhl.de" src="http://dblp.uni-trier.de/db/lzi_logo.gif" border="0" height="56" width="251"/></a></td><td align="right"><a href="http://www.uni-trier.de"><img alt="www.uni-trier.de" src="http://dblp.uni-trier.de/db/logo_universitaet-trier.gif" border="0" height="48" width="215"/></a></td></tr></table>
+<h1>DBLP BibTeX Record '<a href="http://dblp.uni-trier.de/rec/bibtex/journals/cacm/DeanG10.xml">journals/cacm/DeanG10</a>'</h1>
+<pre>@article{<a href="http://dblp.uni-trier.de/db/about/bibtex.html">DBLP</a>:journals/cacm/DeanG10,
+  author    = {Jeffrey Dean and
+               Sanjay Ghemawat},
+  title     = {MapReduce: a flexible data processing tool},
+  journal   = {Commun. ACM},
+  volume    = {53},
+  number    = {1},
+  year      = {2010},
+  pages     = {72-77},
+  ee        = {http://doi.acm.org/10.1145/1629175.1629198},
+  bibsource = {DBLP, http://dblp.uni-trier.de}
+}
+</pre>
+<p><div class="footer"><a href="http://dblp.uni-trier.de/db/index.html">Home</a> | <a href="http://dblp.uni-trier.de/db/conf/indexa.html">Conferences</a> | <a href="http://dblp.uni-trier.de/db/journals/index.html">Journals</a> | <a href="http://dblp.uni-trier.de/db/series/index.html">Series</a> | <a href="http://dblp.uni-trier.de/db/about/faq.html">FAQ</a> &#151; Search: <a href="http://dblp.l3s.de">Faceted</a> | <a href="http://dblp.isearch-it-solutions.net/"> Free</a> | <a href="http://www.dblp.org/search/">Complete</a> | <a href="http://dblp.uni-trier.de/db/indices/a-tree/index.html">Author</a></div>
+<small>Last update 2005-12-16 CET by the <a href="http://dblp.uni-trier.de/db/about/team.html">DBLP Team</a> &#8212; <a href="http://opendefinition.org/"><img alt="This material is Open Data" border="0" src="http://assets.okfn.org/images/ok_buttons/od_80x15_blue.png" style="position:relative; top:3px" /></a> Data released under the <a href="http://opendatacommons.org/licenses/by/summary/">ODC-BY&nbsp;1.0 license</a> &#8212; See also our <a href="http://dblp.uni-trier.de/db/copyright.html">legal information page</a></small></p></body></html>

data/test/power-piccolo.html

@@ -0,0 +1,31 @@
+<html><head><title>DBLP BibTeX Record 'conf/osdi/PowerL10'</title><link href="http://dblp.uni-trier.de/dblp.css" rel="stylesheet" type="text/css" /></head><body>
+<table width="100%"><tr><td align="left"><a href="http://dblp.uni-trier.de/db/index.html"><img alt="dblp.uni-trier.de" src="http://dblp.uni-trier.de/db/Logo.gif" border="0" height="60" width="170"/></a></td><td align="center"><a href="http://www.dagstuhl.de/"><img alt="www.dagstuhl.de" src="http://dblp.uni-trier.de/db/lzi_logo.gif" border="0" height="56" width="251"/></a></td><td align="right"><a href="http://www.uni-trier.de"><img alt="www.uni-trier.de" src="http://dblp.uni-trier.de/db/logo_universitaet-trier.gif" border="0" height="48" width="215"/></a></td></tr></table>
+<h1>DBLP BibTeX Record '<a href="http://dblp.uni-trier.de/rec/bibtex/conf/osdi/PowerL10.xml">conf/osdi/PowerL10</a>'</h1>
+<pre>@inproceedings{<a href="http://dblp.uni-trier.de/db/about/bibtex.html">DBLP</a>:conf/osdi/PowerL10,
+  author    = {Russell Power and
+               Jinyang Li},
+  title     = {Piccolo: Building Fast, Distributed Programs with Partitioned
+               Tables},
+  booktitle = {OSDI},
+  year      = {2010},
+  pages     = {293-306},
+  ee        = {http://www.usenix.org/events/osdi10/tech/full_papers/Power.pdf},
+  crossref  = {DBLP:conf/osdi/2010},
+  bibsource = {DBLP, http://dblp.uni-trier.de}
+}
+</pre>
+<pre>@proceedings{<a href="http://dblp.uni-trier.de/db/about/bibtex.html">DBLP</a>:conf/osdi/2010,
+  editor    = {Remzi H. Arpaci-Dusseau and
+               Brad Chen},
+  title     = {9th USENIX Symposium on Operating Systems Design and Implementation,
+               OSDI 2010,  October 4-6, 2010, Vancouver, BC, Canada, Proceedings},
+  booktitle = {OSDI},
+  publisher = {USENIX Association},
+  year      = {2010},
+  isbn      = {978-1-931971-79-9},
+  ee        = {http://www.usenix.org/event/osdi10/tech/full_papers/osdi10_proceedings.pdf},
+  bibsource = {DBLP, http://dblp.uni-trier.de}
+}
+</pre>
+<p><div class="footer"><a href="http://dblp.uni-trier.de/db/index.html">Home</a> | <a href="http://dblp.uni-trier.de/db/conf/indexa.html">Conferences</a> | <a href="http://dblp.uni-trier.de/db/journals/index.html">Journals</a> | <a href="http://dblp.uni-trier.de/db/series/index.html">Series</a> | <a href="http://dblp.uni-trier.de/db/about/faq.html">FAQ</a> &#151; Search: <a href="http://dblp.l3s.de">Faceted</a> | <a href="http://dblp.isearch-it-solutions.net/"> Free</a> | <a href="http://www.dblp.org/search/">Complete</a> | <a href="http://dblp.uni-trier.de/db/indices/a-tree/index.html">Author</a></div>
+<small>Last update 2009-06-09 CET by the <a href="http://dblp.uni-trier.de/db/about/team.html">DBLP Team</a> &#8212; <a href="http://opendefinition.org/"><img alt="This material is Open Data" border="0" src="http://assets.okfn.org/images/ok_buttons/od_80x15_blue.png" style="position:relative; top:3px" /></a> Data released under the <a href="http://opendatacommons.org/licenses/by/summary/">ODC-BY&nbsp;1.0 license</a> &#8212; See also our <a href="http://dblp.uni-trier.de/db/copyright.html">legal information page</a></small></p></body></html>

data/test/power-piccolo.json

@@ -0,0 +1,91 @@
+{
+"result":{
+"query":"power* piccolo*",
+"status":
+{
+"@code":12,"text":"OK"
+},
+"time":{
+"@unit":"msecs",
+"text":4.96
+},
+"completions":{
+"@total":1,
+"@computed":1,
+"@sent":1,
+"c":{
+"@sc":202,
+"@dc":3,
+"@oc":3,
+"@id":11535429,
+"text":"piccolo"
+}
+},
+"hits":{
+"@total":3,
+"@computed":3,
+"@sent":3,
+"@first":0,
+"hit":[
+{
+"@score":102,
+"@id":1287205,
+"title":{
+"dblp:authors":{
+"dblp:author":[
+"Geev Mokryani",
+"Pierluigi Siano",
+"Antonio Piccolo",
+"Vito Calderaro",
+"Carlo Cecati"]},
+"dblp:title":{
+"@ee":"http://doi.ieeecomputersociety.org/10.1109/FUZZY.2011.6007591",
+"text":"A novel fuzzy system for wind turbines reactive power control. "},
+"dblp:venue":{
+"@url":"db/conf/fuzzIEEE/fuzzIEEE2011.html#MokryaniSPCC11",
+"text":"FUZZ-IEEE 2011:231-235"},
+"dblp:year":2011,
+"dblp:type":"inproceedings"},
+"url":"http://www.dblp.org/rec/bibtex/conf/fuzzIEEE/MokryaniSPCC11"
+},
+{
+"@score":102,
+"@id":1565069,
+"title":{
+"dblp:authors":{
+"dblp:author":[
+"Russell Power",
+"Jinyang Li"]},
+"dblp:title":{
+"@ee":"http://www.usenix.org/events/osdi10/tech/full_papers/Power.pdf",
+"text":"Piccolo: Building Fast, Distributed Programs with Partitioned Tables. "},
+"dblp:venue":{
+"@url":"db/conf/osdi/osdi2010.html#PowerL10",
+"text":"OSDI 2010:293-306"},
+"dblp:year":2010,
+"dblp:type":"inproceedings"},
+"url":"http://www.dblp.org/rec/bibtex/conf/osdi/PowerL10"
+},
+{
+"@score":102,
+"@id":2042841,
+"title":{
+"dblp:authors":{
+"dblp:author":[
+"Giuseppe Bianchi",
+"Nicola Blefari-Melazzi",
+"Francesca Lo Piccolo"]},
+"dblp:title":{
+"@ee":"http://dx.doi.org/10.1109/ISCC.2007.4381640",
+"text":"Impact of Chosen Error Criteria in RSS-based Localization: Power vs Distance vs Relative Distance Error Minimization. "},
+"dblp:venue":{
+"@url":"db/conf/iscc/iscc2007.html#BianchiBP07",
+"text":"ISCC 2007"},
+"dblp:year":2007,
+"dblp:type":"inproceedings"},
+"url":"http://www.dblp.org/rec/bibtex/conf/iscc/BianchiBP07"
+}
+]
+}
+}
+}

data/test/test_quickcite.rb

@@ -0,0 +1,45 @@
+#!/usr/bin/env ruby
+
+require "quickcite"
+require "quickcite/json"
+
+require "test/unit"
+
+class TestJson < Test::Unit::TestCase
+  def test_simple()
+    v = QuickCite::JSONUtil.parse(File.read('test/power-piccolo.json'))
+    v.result.hits.hit.map { |h| puts(h.url) }
+  end
+end
+
+class TestMain < Test::Unit::TestCase
+  def test_simple()
+  texfile = open("/tmp/test.tex", "w")
+  texfile.write <<-END
+  This is a test latex file.
+I like to cite \\cite{PowerPiccolo}
+And \\cite{MapreduceDean}
+END
+  texfile.close
+
+  bibfile = open("/tmp/test.bib", "w")
+  bibfile.write("")
+  bibfile.close
+  
+  QuickCite::run(["/tmp/test.tex"], "/tmp/test.bib")
+  
+  puts("Result: ")
+  puts(File.read("/tmp/test.bib"))
+  end
+end 
+
+class TestCiteToQuery < Test::Unit::TestCase
+  def test_simple()
+    assert_equal(
+      ["Russell", "Power", "Piccolo"],
+        QuickCite::cite_to_query("RussellPowerPiccolo"))
+    assert_equal(
+      ["Piccolo", "2010"],
+        QuickCite::cite_to_query("Piccolo2010"))
+  end
+end

metadata

@@ -0,0 +1,120 @@
+--- !ruby/object:Gem::Specification 
+name: quickcite
+version: !ruby/object:Gem::Version 
+  hash: 23
+  prerelease: 
+  segments: 
+  - 1
+  - 0
+  - 0
+  version: 1.0.0
+platform: ruby
+authors: 
+- Russell Power
+autorequire: 
+bindir: bin
+cert_chain: []
+
+date: 2012-07-18 00:00:00 Z
+dependencies: 
+- !ruby/object:Gem::Dependency 
+  name: bibtex-ruby
+  prerelease: false
+  requirement: &id001 !ruby/object:Gem::Requirement 
+    none: false
+    requirements: 
+    - - ">="
+      - !ruby/object:Gem::Version 
+        hash: 3
+        segments: 
+        - 2
+        - 0
+        version: "2.0"
+  type: :runtime
+  version_requirements: *id001
+- !ruby/object:Gem::Dependency 
+  name: json
+  prerelease: false
+  requirement: &id002 !ruby/object:Gem::Requirement 
+    none: false
+    requirements: 
+    - - ">="
+      - !ruby/object:Gem::Version 
+        hash: 3
+        segments: 
+        - 0
+        version: "0"
+  type: :runtime
+  version_requirements: *id002
+- !ruby/object:Gem::Dependency 
+  name: bundler
+  prerelease: false
+  requirement: &id003 !ruby/object:Gem::Requirement 
+    none: false
+    requirements: 
+    - - ">="
+      - !ruby/object:Gem::Version 
+        hash: 3
+        segments: 
+        - 0
+        version: "0"
+  type: :development
+  version_requirements: *id003
+description: 
+email: power@cs.nyu.edu
+executables: []
+
+extensions: []
+
+extra_rdoc_files: []
+
+files: 
+- Gemfile
+- Gemfile.lock
+- Rakefile
+- bin/quickcite
+- lib/quickcite.rb
+- lib/quickcite/citeulike.rb
+- lib/quickcite/dblp.rb
+- lib/quickcite/json.rb
+- lib/quickcite/version.rb
+- quickcite.gemspec
+- test/dean.html
+- test/power-piccolo.html
+- test/power-piccolo.json
+- test/test_quickcite.rb
+homepage: 
+licenses: []
+
+post_install_message: 
+rdoc_options: []
+
+require_paths: 
+- lib
+required_ruby_version: !ruby/object:Gem::Requirement 
+  none: false
+  requirements: 
+  - - ">="
+    - !ruby/object:Gem::Version 
+      hash: 3
+      segments: 
+      - 0
+      version: "0"
+required_rubygems_version: !ruby/object:Gem::Requirement 
+  none: false
+  requirements: 
+  - - ">="
+    - !ruby/object:Gem::Version 
+      hash: 3
+      segments: 
+      - 0
+      version: "0"
+requirements: []
+
+rubyforge_project: 
+rubygems_version: 1.8.15
+signing_key: 
+specification_version: 3
+summary: Efficient citation lookup and management
+test_files: []
+