gentle-scholar 1.0.2 → 2.0.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: 9a16b27e1ce933a215cfc7d0e8ba1763f0c24aec
-  data.tar.gz: 778624c2dcc3146444938c40869775192ab65edc
+  metadata.gz: aa9d57fcccb8fd1fea02fe1a2651942dcbbf3940
+  data.tar.gz: 0f9d9a785ff33dba50e9418f9d892e3991e09e15
 SHA512:
-  metadata.gz: d1f88ec34776d3abadca4c53ba613151c12c0b77df0dd993aef403ec72a7aba3aa27485728e543f0cfa0a308c6c2b6674d7761394beea3543c299f37b79eb2d7
-  data.tar.gz: 462726a60bf05d720f9503390f4e3ebed7cdbc55d825631275a40ccb7aed4d2b65f8ea8aa2afd9e6434f261e765b428af1331243ed609f0d60e3aa8ddacbcc95
+  metadata.gz: 814517c4c9ed75066116a33106013645f14a0d67e964f89b92db1ae8bfebbca4b5a82822d349c97ff9081f3093dbdfd5946db2dfee90f9de0209588a8a52b8bd
+  data.tar.gz: 38af181ecc696bc79056c638ac96337806c363e3178c0fa30f36ae2f9e43161156975fedc42fa388c87ec55b94dda259fc8c61e707954b1209fed23cbf0c5842

data/.gitignore

@@ -0,0 +1,2 @@
+*.gem
+Gemfile.lock

data/.travis.yml

@@ -0,0 +1,9 @@
+language: ruby
+rvm:
+  - ruby-head
+  - 2.1.0
+  - 1.9.3
+  - jruby-head
+	- jruby-19mode
+only:
+  - master

data/Gemfile

@@ -0,0 +1,8 @@
+source 'https://rubygems.org'
+
+gem 'pry'      # allows us to use pry for bundle console
+gemspec
+
+group :test do
+  gem 'rake'
+end

data/README.md

@@ -0,0 +1,42 @@
+# gentle-scholar gem
+[![Gem Version](https://badge.fury.io/rb/gentle-scholar.svg)](http://badge.fury.io/rb/gentle-scholar)
+[![Build Status](https://travis-ci.org/soumyaray/gentle-scholar.svg?branch=v1.0.1)](https://travis-ci.org/soumyaray/gentle-scholar)
+
+Gem to extract Google Scholar publication given URL of a publication
+
+##About:
+**NOTE:** This gem is not to to search or crawl through Google Scholar.
+This also will not extract bibtex (please see the excellent
+  [gscholar](https://rubygems.org/gems/gscholar) gem for that).
+
+This gem is for publishing academics to extract their own publication
+information from Google Scholar. It can retrieve standard citation information
+(authors, journal, title, etc.). Additionally, it can retrieve number of
+citations reported by Google Scholar, link to citations page, chart of
+citations over time, and link to author's Google Scholar profile.
+
+##Usage:
+
+Given a google scholar article such as:
+
+    http://scholar.google.com.tw/citations?view_op=view_citation&hl=en&user=6WjiSOwAAAAJ&citation_for_view=6WjiSOwAAAAJ:9yKSN-GCB0IC
+
+Retrieve information by copying the author and article ID from the URL:
+
+    sec_paper = GScholarPub.new('6WjiSOwAAAAJ:u5HHmVD_uO8C')
+
+This returns:
+
+    #<GScholarPub:0x007fb0bc18df70
+      @article_url="http://pubsonline.informs.org/doi/abs/10.1287/isre.1100.0340",
+      @authors=[["Soumya", "Ray"], ["Sung", "S", "Kim"], ["James", "G", "Morris"]],
+      @chart_url="http://www.google.com/chart?chs=475x90&cht=bvs&chf=bg,s,e8f4f7&chco=1111cc&chbh=r,2.0,0.0&chxt=x,y&chxr=1,0,5,5&chd=t:100.0,80.0,100.0
+      @cites=14,
+      @cites_url="http://scholar.google.com/scholar?oi=bibs&hl=en&oe=ASCII&cites=11777343089817755068",
+      @description="The highly competitive and rapidly changing market for online services is becoming increasingly effective at locking users in through
+      @issue="1",
+      @journal="Information Systems Research",
+      @pages="197-213",
+      @title="Research Note\u0097Online Users' Switching Costs: Their Nature and Formation",
+      @volume="23"
+    >

data/Rakefile

@@ -0,0 +1,7 @@
+require 'rake/testtask'
+
+task :default => [:spec]
+
+Rake::TestTask.new(name=:spec) do |t|
+  t.pattern = "spec/*_spec.rb"
+end

data/bin/gentle

@@ -0,0 +1,10 @@
+#!/usr/bin/env ruby
+
+require 'gentle-scholar'
+
+pub = GentleScholar::Publication.get_from_http('6WjiSOwAAAAJ:9yKSN-GCB0IC')
+
+puts 'Publication Details'
+pub.each do |field, data|
+  puts "#{field}: #{pub[field]}"
+end

data/gentle-scholar.gemspec

@@ -0,0 +1,20 @@
+$:.push File.expand_path("../lib", __FILE__)
+require 'gentle-scholar/version'
+
+Gem::Specification.new do |s|
+  s.name        =  'gentle-scholar'
+  s.version     =  GentleScholar::VERSION
+  s.add_development_dependency 'minitest'
+  s.add_development_dependency 'minitest-rg'
+  s.add_runtime_dependency 'nokogiri', '>= 1.6.2'
+  s.add_runtime_dependency 'typhoeus', '>= 0.6.8'
+  s.date        =  '2014-05-27'
+  s.summary     =  'Google Scholar infor extractor'
+  s.description =  'Extract author/paper info from Google Scholar'
+  s.authors     =  ['Soumya Ray']
+  s.email       =  'soumya.ray@gmail.com'
+  s.files       =  `git ls-files`.split("\n")
+  s.test_files  =  `git ls-files -- {test,spec,features}/*`.split("\n")
+  s.homepage    =  'https://github.com/soumyaray/gentle-scholar'
+  s.license     =  'MIT'
+end

data/lib/gentle-scholar.rb

@@ -0,0 +1 @@
+require 'gentle-scholar/publication.rb'

data/lib/gentle-scholar/publication.rb

@@ -0,0 +1,80 @@
+# Author::       Soumya Ray (mailto: soumya.ray@gmail.com)
+# License::      MIT
+require 'typhoeus'
+require 'nokogiri'
+require 'date'
+
+module GentleScholar
+  # This class loads a single publication from Google scholar and returns
+  # all its attributes, including dynamic attributes like number of citations
+  class Publication
+    GS_HOST_URL  = 'http://scholar.google.com'
+    GS_CIT_URL   = "#{GS_HOST_URL}/citations?view_op=view_citation&hl=en"
+
+    SCAN_STR = {
+      gscholar_url:  "//div[contains(@class,'g-section cit-dgb')]"\
+                                   "/div/table/tr/td/a",
+      cites:         "//div[contains(@id,'scholar_sec')]/div/a",
+      cites_url:     "//div[contains(@id,'scholar_sec')]/div/a",
+      title:         '//div[@id="title"]/a',
+      article_url:   '//div[@id="title"]/a',
+      chart_url:     '//div[contains(@class,"cit-dd")]/img'
+    }
+
+    SCAN_LAMBDAS = {
+      cites:         ->(x) { x.text[/\d+/].to_i },
+      cites_url:     ->(x) { x[0].attributes['href'].value },
+      title:         ->(x) { x.text },
+      article_url:   ->(x) { x.attr('href').value },
+      chart_url:     ->(x) { x.attr('src').value },
+      gscholar_url:  ->(x) { GS_HOST_URL + x.attr('href').value }
+    }
+
+    TABLE_ATTR = {
+      authors: 'Authors',
+      date: 'Publication date',
+      journal: 'Journal name',
+      volume: 'Volume',
+      issue: 'Issue',
+      pages: 'Pages',
+      publisher: 'Publisher',
+      description: 'Description'
+    }
+
+    TABLE_LAMBDAS = {
+      authors:  ->(x) { x.split(/,/).map { |a| a.split(' ') } },
+      date:     ->(x) { Date.strptime(x, '%Y/%m/%d') }
+    }
+
+    def self.get_from_http(scholar_pub_id)
+      auth_id, pub_id = scholar_pub_id.split(/:/)
+      url = GS_CIT_URL + '&user=' + auth_id \
+                       + '&citation_for_view=' + auth_id + ':' + pub_id
+      res = Typhoeus::Request.new(url).run
+      doc = Nokogiri::HTML(res.response_body)
+
+      extract_html_elements(doc).merge(extract_html_table(doc))
+    end
+
+    def self.extract_html_elements(doc)
+      xpath = Hash[SCAN_STR.map { |elem, path| [elem, doc.xpath(path)] }]
+      elements = SCAN_LAMBDAS.map do |key, lam|
+        if xpath[key].any? then [key, lam.call(xpath[key])] end
+      end
+
+      Hash[elements.compact]
+    end
+
+    def self.extract_html_table(doc)
+      # lambda gets text from right html column given name in left column
+      table_extract = lambda do |name|
+        doc.xpath("//div[starts-with(.,'#{name}')]")[0].children[1].text
+      end
+
+      elements = Hash[TABLE_ATTR.map { |k, v| [k, table_extract.call(v)] }]
+      elements.merge(
+        Hash[TABLE_LAMBDAS.map { |key, lam| [key, lam.call(elements[key])] }]
+      )
+    end
+  end
+end

data/lib/gentle-scholar/version.rb

@@ -0,0 +1,3 @@
+module GentleScholar
+  VERSION = '2.0.0'
+end

data/spec/minitest_helper.rb

@@ -0,0 +1,3 @@
+require './lib/gentle-scholar.rb'
+
+SEC_PAPER = GentleScholar::Publication.get_from_http('6WjiSOwAAAAJ:u5HHmVD_uO8C')

data/spec/publication_spec.rb

@@ -0,0 +1,76 @@
+require 'minitest/autorun'
+require 'minitest/rg'
+require './spec/minitest_helper.rb'
+
+describe 'Publication', 'A single publication listing' do
+
+  # let (:@sec_paper) do
+  #   GScholarPub.new('6WjiSOwAAAAJ:u5HHmVD_uO8C')
+  # end
+
+  describe 'when it is a paper' do
+    #
+    # before do
+    #   @sec_paper = GScholarPub.new('6WjiSOwAAAAJ:u5HHmVD_uO8C')
+    # end
+
+    before do
+      @sec_paper = SEC_PAPER
+    end
+
+    it 'has the right title' do
+      @sec_paper[:title].must_equal 'Security Assurance: How Online Service '\
+          'Providers Can Influence Security Control Perceptions and Gain Trust'
+    end
+
+    it 'has some number of cites' do
+      @sec_paper[:cites].must_be :>, 0
+    end
+
+    it 'has some url for listing cites' do
+      @sec_paper[:cites_url].must_match /http:\/\/.*/
+    end
+
+    it 'has some url for citations chart' do
+      @sec_paper[:chart_url].must_match /http:\/\/.*/
+    end
+
+    it 'has some url for the pulished article' do
+      @sec_paper[:article_url].must_match /http:\/\/.*/
+    end
+
+    it 'has the right author(s) (as nested array)' do
+      @sec_paper[:authors].must_equal [['Soumya', 'Ray'], ['Terence', 'Ow'], ['Sung', 'S', 'Kim']]
+    end
+
+    it 'has a publication date' do
+      @sec_paper[:date].must_be_instance_of Date
+    end
+
+    it 'has the right journal\'s name' do
+      @sec_paper[:journal].must_equal 'Decision Sciences'
+    end
+
+    it 'has the right volume number (as string)' do
+      @sec_paper[:volume].must_equal '42'
+    end
+
+    it 'has the right issue number (as string)' do
+      @sec_paper[:issue].must_equal '2'
+    end
+
+    it 'has the right page numbers' do
+      @sec_paper[:pages].must_equal '391-412'
+    end
+
+    it 'has the right publisher' do
+      @sec_paper[:publisher].must_equal 'Blackwell Publishing Inc'
+    end
+
+    it 'has some url for the main citations apage' do
+      @sec_paper[:gscholar_url].must_match /citations/
+    end
+
+  end
+
+end

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: gentle-scholar
 version: !ruby/object:Gem::Version
-  version: 1.0.2
+  version: 2.0.0
 platform: ruby
 authors:
 - Soumya Ray
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2014-05-25 00:00:00.000000000 Z
+date: 2014-05-27 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: minitest
@@ -72,7 +72,18 @@ executables: []
 extensions: []
 extra_rdoc_files: []
 files:
-- lib/publication.rb
+- ".gitignore"
+- ".travis.yml"
+- Gemfile
+- README.md
+- Rakefile
+- bin/gentle
+- gentle-scholar.gemspec
+- lib/gentle-scholar.rb
+- lib/gentle-scholar/publication.rb
+- lib/gentle-scholar/version.rb
+- spec/minitest_helper.rb
+- spec/publication_spec.rb
 homepage: https://github.com/soumyaray/gentle-scholar
 licenses:
 - MIT
@@ -97,4 +108,6 @@ rubygems_version: 2.1.11
 signing_key: 
 specification_version: 4
 summary: Google Scholar infor extractor
-test_files: []
+test_files:
+- spec/minitest_helper.rb
+- spec/publication_spec.rb

data/lib/publication.rb

@@ -1,78 +0,0 @@
-# Author::       Soumya Ray (mailto: soumya.ray@gmail.com)
-# License::      MIT
-
-require 'typhoeus'
-require 'nokogiri'
-require 'date'
-
-# This class loads a single publication from Google scholar and returns
-# all its attributes, including dynamic attributes like number of citations
-class GScholarPub
-  GSCHOLAR_HOST_URL = 'http://scholar.google.com'
-  GSCHOLAR_CIT_URL =
-      'http://scholar.google.com/citations?view_op=view_citation&hl=en'
-
-  attr_reader :title, :cites, :cites_url, :chart_url, :article_url
-  attr_reader :authors, :date, :journal, :volume, :issue, :pages, :publisher
-  attr_reader :description, :gscholar_url
-  # TODO: @doc only for development, testing modes
-  attr_reader :doc
-
-  SCAN_STR = {
-    gscholar_url:
-      "//div[contains(@class,'g-section cit-dgb')]/div/table/tr/td/a",
-    cites: "//div[contains(@id,'scholar_sec')]/div/a",
-    title: '//div[@id="title"]/a',
-    article_url: '//div[@id="title"]/a',
-    chart_url: '//div[contains(@class,"cit-dd")]/img'
-  }
-
-  TABLE_ATTR = {
-    authors: 'Authors',
-    date: 'Publication date',
-    journal: 'Journal name',
-    volume: 'Volume',
-    issue: 'Issue',
-    pages: 'Pages',
-    publisher: 'Publisher',
-    description: 'Description'
-  }
-
-  def initialize(scholar_pub_id)
-    auth_id, pub_id = scholar_pub_id.split(/:/)
-    url = GSCHOLAR_CIT_URL + '&user=' + auth_id \
-                           + '&citation_for_view=' + auth_id + ':' + pub_id
-    res = Typhoeus::Request.new(url).run
-    @doc = Nokogiri::HTML(res.response_body)
-
-    extract_html_elements
-    extract_html_table
-  end
-
-  def extract_html_elements
-    @cites = @doc.xpath(SCAN_STR[:cites]).text[/\d+/].to_i
-    @cites_url = @doc.xpath(SCAN_STR[:cites])[0].attributes['href'].value
-
-    @title = @doc.xpath(SCAN_STR[:title]).text
-    @article_url = @doc.xpath(SCAN_STR[:article_url]).attr('href').value
-
-    @chart_url = @doc.xpath(SCAN_STR[:chart_url]).attr('src').value
-
-    @gscholar_url = GSCHOLAR_HOST_URL + @doc.xpath(
-                                  SCAN_STR[:gscholar_url]).attr('href').value
-  end
-
-  def extract_html_table
-    # lambda gets text from right html column given name in left column
-    table_pick = lambda do |name|
-      @doc.xpath("//div[starts-with(.,'#{name}')]")[0].children[1].text
-    end
-
-    TABLE_ATTR.each do |k, v|
-      instance_variable_set("@#{k}", table_pick.call(v))
-    end
-
-    @authors = @authors.split(/,/).map { |a| a.split(' ') }
-    @date = Date.strptime(@date, '%Y/%m/%d')
-  end
-end