ref2bibtex 0.0.3 → 0.1.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: d7757e8fcfbfeccc415e91c00e4f2302b0ce37c4
-  data.tar.gz: 1bf31207bbc840f4ac1cd0617eb06713d47b1340
+  metadata.gz: 5f7108348cfed966bb638c38ff3b1056bc418ed6
+  data.tar.gz: b6553afe9e23c6a65d5c78a7fdc2d493379fc6ea
 SHA512:
-  metadata.gz: 09316afcae92969335e4c421860aa5a2ac56f816039b58e581ba1afd34169b4f1b57cfc317db1e178ea2be35446955b6c46e35d59717bf02dbc5eecef1c82258
-  data.tar.gz: 4dcafb5dca149065d4c9e71921fff3f26dc0a8f004609fd25ca31ae4e11f9464b926a78bcf5e84c60f0edbff7974e19b0d3c3e2878c07030b2553faf7fd495b7
+  metadata.gz: a14d2e3ee16b6fb7e04cc5c6b7900e0734aceb7548bfd0641f644bcba7e73c364b536c264dedd0df33d221d89ba1dabd36eb345ce1ba7c430c139086a0604811
+  data.tar.gz: d20b75b7ca8d76ffacb1eacb99c037ea99d98d284ceaf86347c530016c263c81dbc887e81020da1ed91c6c8e2d6748bf589a4af7d2f78139c9ac8a6ac7e2d601

data/Gemfile.lock

@@ -1,20 +1,14 @@
 PATH
   remote: .
   specs:
-    ref2bibtex (0.0.3)
+    ref2bibtex (0.1.0)
 
 GEM
   remote: https://rubygems.org/
   specs:
-    awesome_print (1.2.0)
-    byebug (3.4.0)
-      columnize (~> 0.8)
-      debugger-linecache (~> 1.2)
-      slop (~> 3.6)
-    columnize (0.8.9)
-    debugger-linecache (1.2.0)
-    rake (10.3.2)
-    slop (3.6.0)
+    awesome_print (1.7.0)
+    byebug (9.0.6)
+    rake (12.0.0)
 
 PLATFORMS
   ruby
@@ -22,6 +16,9 @@ PLATFORMS
 DEPENDENCIES
   awesome_print
   bundler (~> 1.6)
-  byebug
-  rake
+  byebug (~> 9.0)
+  rake (~> 12.0)
   ref2bibtex!
+
+BUNDLED WITH
+   1.14.6

data/README.md

@@ -1,10 +1,10 @@
-ref2bibtex
-==========
+[![Build Status](https://travis-ci.org/SpeciesFileGroup/ref2bibtex.svg?branch=master)](https://travis-ci.org/SpeciesFileGroup/ref2bibtex)
 
-A single purpose gem wrapping Crossref's API.  Pass it a full reference string, get back bibtex.
+# ref2bibtex
 
-usage
-=====
+An (almost) single purpose gem wrapping Crossref's API.  Pass it a full reference string, get back bibtex.
+
+# usage
 
 Use citation2bibtex (aliased _get_):
 
@@ -34,20 +34,24 @@ If you have the doi:
 
 ```
 
-faq
-===
+If you want a score: 
+
+```ruby
+
+  Ref2bibtex.get_score('E. Ven. 1337. Fake articles. Journal Get Scores. Hm:mm') # => 23.688715
+
+```
+
+# faq
 
-What if there are multiple results?
------------------------------------
+## What if there are multiple results?
 The code is dumb, it takes the first.
 
-acknowledgements
-================
+# acknowledgements
 
-The Crossref API, and Jon Hill, University of York, for his Python version and bringing up the approach.
+The Crossref API. Jon Hill, University of York, for his Python version and bringing up the approach.
 
-license
-=======
+# license
 
 NCSA, UI flavor. [http://opensource.org/licenses/NCSA]
 

data/lib/ref2bibtex/version.rb

@@ -1,3 +1,3 @@
 module Ref2bibtex
-  VERSION = "0.0.3"
+  VERSION = "0.1.0"
 end

data/lib/ref2bibtex.rb

@@ -7,8 +7,25 @@ require 'net/http'
 
 module Ref2bibtex
 
+  # By default sorts by score
   CROSSREF_URI = URI('http://search.crossref.org/links')
 
+  DEFAULT_CUTOFF = 50
+
+  @@cutoff = DEFAULT_CUTOFF
+
+  def self.cutoff
+    @@cutoff
+  end
+
+  def self.cutoff=(value)
+    @@cutoff = value
+  end
+
+  def self.reset_cutoff
+    @@cutoff = DEFAULT_CUTOFF
+  end
+
   # Parse the response into json
   def self.parse_json(string)
     begin
@@ -26,26 +43,36 @@ module Ref2bibtex
     return false if uri.class == URI::Generic
     response = Ref2bibtex.request(uri, headers: {'Accept' => 'application/x-bibtex' }, protocol: 'GET', process_response_as: 'text') 
   end
-
-  # Pass a String citation, get a doi back
+ 
+  # Pass a String citation, get a DOI back
   def self.get_doi(citation)
-    if citation.class == String
-      citation = [citation]
-    elsif citation.class != Array
-      raise
-    end
-    
+    citation = validate_query(citation)
     response = Ref2bibtex.request(payload: citation) 
-    if response['results'][0]['match']
-      response['results'].first['doi']
-    else
-      false
-    end
+
+    return false if !response['results'][0]['match']
+    return false if response['results'][0]['score'] < @@cutoff
+
+    response['results'][0]['doi']
+  end
+
+  # Pass a String citation, get a score back
+  def self.get_score(citation)
+    citation = validate_query(citation)
+    response = Ref2bibtex.request(payload: citation) 
+    return false if !response['results'][0]['match']
+    response['results'][0]['score']
+  end
+
+  def self.validate_query(citation)
+    return [citation] if citation.kind_of?(String) && citation.length > 0
+    raise 'citation is not String or Array' if !citation.kind_of?(Array) 
+    raise 'citation in array is empty' if citation.empty? || citation.select{|a| a.length == 0}.size > 0
+    citation
   end
 
-  # Pass a citation, get a String in bibtex back
+  # Pass a citation, get a String in BibTeX back
   def self.citation2bibtex(citation)
-    get_bibtex( get_doi(citation) )
+    get_bibtex(get_doi(citation) )
   end
 
   class << self
@@ -54,47 +81,66 @@ module Ref2bibtex
 
   def self.request(url = CROSSREF_URI, payload: nil, headers: {'content-type' => 'application/json' }, protocol: 'POST', process_response_as: 'json', redirect_limit: 10)
     raise 'Infinite redirect?' if redirect_limit == 0
-    data = nil
-    if protocol == 'POST'
-      if payload.nil?
-        payload = {}
-      end
-      data = JSON.generate(payload) # Json.new(payload) # utf-8 encoding?
-    else
-      data = nil
-    end
 
-    if protocol == 'POST'
-      request = Net::HTTP::Post.new(url, initheader = headers) 
-    elsif protocol == 'GET'
-      request = Net::HTTP::Get.new(url, initheader = headers) 
-    end
+    body = request_body(protocol, payload)
+    request = new_request(protocol, url, headers)
 
-    response = Net::HTTP.start(request.uri.hostname, request.uri.port) do |http|
-      request.body = data 
+    response = Net::HTTP.start(request.uri.hostname, request.uri.port, use_ssl: request.uri.scheme == 'https') do |http|
+      request.body = body 
       http.request(request)
     end
 
     case response
-    when Net::HTTPSuccess then
+    when Net::HTTPSuccess 
       response = response
-    when Net::HTTPRedirection then
+    when Net::HTTPRedirection 
+
       url = URI(response['location'])
       request = Net::HTTP::Get.new(url, initheader = {'Accept' => 'application/x-bibtex'}) 
-      response = Net::HTTP.start(request.uri.hostname, request.uri.port) do |http|
+
+      response = Net::HTTP.start(request.uri.hostname, request.uri.port, use_ssl: request.uri.scheme == 'https') do |http|
         http.request(request) 
       end
     else
       response = response.value
     end
 
-    case process_response_as
+     process_response(response, process_response_as)
+  end
+
+  protected
+
+  def self.interpret_response(response)
+  end
+
+  def self.request_body(protocol, payload)
+    if protocol == 'POST'
+      payload = {} if payload.nil?
+      JSON.generate(payload) # Json.new(payload) # utf-8 encoding?
+    else
+      nil
+    end
+  end
+
+  def self.new_request(protocol, url, headers)
+    case protocol
+    when 'POST'
+      Net::HTTP::Post.new(url, initheader = headers) 
+    when 'GET'
+      Net::HTTP::Get.new(url, initheader = headers) 
+    else
+      raise 'invalid protocol'
+    end
+  end
+
+  def self.process_response(response, as)
+    case as
     when 'text' 
       response.body
     when 'json'
       parse_json(response.body)
     else
-      raise
+      raise 'response process type not provided'
     end
   end
 

data/ref2bibtex.gemspec

@@ -19,8 +19,8 @@ Gem::Specification.new do |spec|
   spec.require_paths = ["lib"]
 
   spec.add_development_dependency "bundler", "~> 1.6"
-  spec.add_development_dependency "rake"
-  spec.add_development_dependency "byebug"
+  spec.add_development_dependency "rake", "~> 12.0"
+  spec.add_development_dependency "byebug", "~> 9.0"
   spec.add_development_dependency "awesome_print"
 
 end

data/spec/ref2bibtex_spec.rb

@@ -2,37 +2,37 @@ require 'spec_helper'
 
 describe Ref2bibtex do
 
-  context '#request' do
+  context '.request' do
     let(:request) { Ref2bibtex.request(payload: [CITATIONS[:first]]) }
 
-    specify '#Ref2bibtex.request makes a query_ok request' do
+    specify 'makes a query_ok request' do
       expect(request['query_ok']).to eq(true)
     end
 
-    specify '#Ref2bibtex.request returns Hash' do
+    specify 'returns Hash' do
       expect(request.class).to eq(Hash)
     end
 
-    specify '#Ref2bibtex.request results is an Array' do
+    specify 'results is an Array' do
       expect(request['results'].class).to eq(Array)
     end
 
-    specify '#Ref2bibtex.request results is an Array of Hashes' do
+    specify 'results is an Array of Hashes' do
       expect(request['results'].first.class).to eq(Hash)
     end
 
-    specify '#Ref2bibtex.request results is an Array of Hashes, each one having a doi' do
+    specify 'results is an Array of Hashes, each one having a doi' do
       expect(request['results'].first['doi']).to be_truthy 
     end
   end
 
-  context '#get_doi' do
-    specify 'Ref2bibtex.get_doi() takes a full citation and returns a string' do
+  context '.get_doi' do
+    specify 'takes a full citation and returns a string' do
       expect(Ref2bibtex.get_doi(CITATIONS[:first])).to eq('http://dx.doi.org/10.3897/zookeys.20.205')
     end
 
     specify 'a citation that can not be resolved returns false' do
-      expect(Ref2bibtex.get_doi(CITATIONS[:fifth])).to eq(false)
+      expect(Ref2bibtex.get_doi(CITATIONS[:eighth])).to eq(false)
     end
 
     specify 'a badly formed DOI returns false' do
@@ -40,9 +40,9 @@ describe Ref2bibtex do
     end
   end
 
-  context '#get_bibtex' do
+  context '.get_bibtex' do  
     let(:response) { Ref2bibtex.get_bibtex('http://dx.doi.org/10.3897/zookeys.20.205')}
-    specify 'Ref2bibtex.get_bibtex() takes a full citation and returns bibtex' do
+    specify 'takes a full citation and returns bibtex' do
       expect(response).to match(/author\s=/)
       expect(response).to match(/title\s=/)
       expect(response).to match(/year\s=\s2009/)
@@ -53,8 +53,64 @@ describe Ref2bibtex do
     end
   end
 
-  specify "a citation that can not be resolved returns false" do
-    expect(Ref2bibtex.get(CITATIONS[:fifth])).to eq(false)
+  specify "a citation that can not be resolved returns false from .get" do
+    expect(Ref2bibtex.get(CITATIONS[:eighth])).to eq(false)
   end
 
+  context 'score' do
+    specify 'can be returned with .get_score' do
+      expect(Ref2bibtex.get_score(CITATIONS[:first])).to be > 0
+    end
+
+    context 'interpretation' do
+      before(:all) {
+        @scores =  CITATIONS.keys.inject({}) { |hsh, c|
+          hsh.merge!(
+            c => Ref2bibtex.get_score(CITATIONS[c])
+          )
+        }
+      }
+
+      specify 'mangled text is worse than good text' do
+        expect( @scores[:second] > @scores[:seventh] ).to be_truthy
+      end
+
+      context 'default @@cutoff is reasonable and slightly conservative' do
+        let(:good_citations) { [:first, :second, :third, :fourth, :sixth] }
+        let(:bad_citations) { CITATIONS.keys - good_citations }
+
+        specify 'for good citations' do
+          good_citations.each do |c|
+            expect(@scores[c]).to be > Ref2bibtex.cutoff
+          end
+        end
+
+        specify 'for bad citations' do
+          bad_citations.each do |c|
+            if @scores[c]
+              expect(@scores[c]).to be < Ref2bibtex.cutoff
+            end
+          end
+        end
+      end
+    end
+
+    context 'cutoff' do
+      after(:all) {
+        Ref2bibtex.reset_cutoff
+      }
+
+      specify 'can be used to reject good matches' do
+        Ref2bibtex.cutoff = 1000
+        expect(Ref2bibtex.get_doi(CITATIONS[:first])).to eq(false)
+      end 
+
+      specify 'can be used to accept bad matches' do
+        Ref2bibtex.cutoff = 1 
+        expect(Ref2bibtex.get_doi(CITATIONS[:fifth])).to be_truthy
+      end 
+
+    end
+
+  end
 end 

data/spec/support/citations.yml

@@ -8,4 +8,11 @@
 :fourth:
   "Otte, D. 1994. Crickets (Grylloidea). Orthoptera Species File, The Orthopterists' Society and The Academy of Natural Sciences of Philadelphia, Philadelphia 1:87"
 :fifth:
-  "Defaut. 2006. Révision préliminaire des Oedipoda ouest-paléarctiques (Caelifera, Acrididae, Oedipodinae). Matériaux Orthoptériques et Entomocénotiques 11:23-48"
+  "Iam, F.K. 1337. Let's go Crossref, give me a DOI for this fake reference in my unit test, I dare you. Journal of Examples for Unit tests. 13:37"
+:sixth:
+  "Simon, E. (1879b) Les Arachnides de France. Tome 7. Contenant les ordres des Chernetes, Scorpiones et Opiliones. Librairie Encyclopédique de Roret, Paris, pp. 1–332, pl. 17–24. "
+:seventh: 
+  'Jaost & K.P. Shuw. 2007. Phyloguny of Ensifere (Hesapoda: Prthoptera) usine threa rubosomal toci, wuth implocations far teh evolushun off acoustik commmunication. Tolecular Fylogenetics & Evolotion (Mul. Phylagenet. Evil.) 39:509-528'
+:eighth:
+  'Only two'
+

metadata

@@ -1,14 +1,14 @@
 --- !ruby/object:Gem::Specification
 name: ref2bibtex
 version: !ruby/object:Gem::Version
-  version: 0.0.3
+  version: 0.1.0
 platform: ruby
 authors:
 - Matt Yoder
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2014-11-06 00:00:00.000000000 Z
+date: 2017-06-07 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: bundler
@@ -28,30 +28,30 @@ dependencies:
   name: rake
   requirement: !ruby/object:Gem::Requirement
     requirements:
-    - - ">="
+    - - "~>"
       - !ruby/object:Gem::Version
-        version: '0'
+        version: '12.0'
   type: :development
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
-    - - ">="
+    - - "~>"
       - !ruby/object:Gem::Version
-        version: '0'
+        version: '12.0'
 - !ruby/object:Gem::Dependency
   name: byebug
   requirement: !ruby/object:Gem::Requirement
     requirements:
-    - - ">="
+    - - "~>"
       - !ruby/object:Gem::Version
-        version: '0'
+        version: '9.0'
   type: :development
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
-    - - ">="
+    - - "~>"
       - !ruby/object:Gem::Version
-        version: '0'
+        version: '9.0'
 - !ruby/object:Gem::Dependency
   name: awesome_print
   requirement: !ruby/object:Gem::Requirement
@@ -106,7 +106,7 @@ required_rubygems_version: !ruby/object:Gem::Requirement
       version: '0'
 requirements: []
 rubyforge_project: 
-rubygems_version: 2.2.2
+rubygems_version: 2.6.12
 signing_key: 
 specification_version: 4
 summary: Pass a full citation, get the bibtex back, that's all.