citesight 0.0.4 → 0.1.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: 60bfb7a13222d0a5c5e1bdcc55d4533b45c817db
-  data.tar.gz: 250740d51c6a41f5733d1014b62ef0070b96623e
+  metadata.gz: a923e35cd2ee4dfb162449aba2f5c1fb50358efe
+  data.tar.gz: 606aeba079bda3f0ebb5248d22bbfa28b301e5df
 SHA512:
-  metadata.gz: 8d5c728f567de3370d12fcc07750d4cafe09f5c09ba6502f5560a2a386fee92e1d2059d4464a058c11ee58f9d2a7f25ed311637d2688ca401c098ea18c546e2d
-  data.tar.gz: 6c80cbc99f513a66925bb1668e2323fa7912ac4fff04d576d1a41d4c842f773a6388286652c87d69b469b388dba103aba777af502c32838372426dfe63235c99
+  metadata.gz: da9e7c7f486d9f04ba1a7e8e046adb90fb53dacb868aceca082a1ec22fa67548c99c6b71c79e1b8749a23d6a79be80845b60147d4d724805802c05937769a0dc
+  data.tar.gz: b63e4978e3e02a876cf8ee09c9a28e0f8049eb68f8b80872abe6902f30df1f734c839397c5563fe2cb5d07bde9383e7256f2a91ae138bb67598e08a52d5b480f

data/lib/citesight.rb

@@ -6,36 +6,50 @@ class PaperCitations
     new(contents).unique_cites
   end
 
+  def self.index_of_cite(contents, cite)
+    new(contents).index_of_cite(cite)
+  end
+
   def initialize(contents)
     @contents = contents
   end
 
   def unique_cites
-    # clean citations of slashes, commas, semi-colons, possesives
     clean_cites = @contents.scan(cite_match).map do |c|
-      c[0].gsub(/[\(\),;]|([\'\u2019]s)/, '').gsub(/[\'\u2019]\s/, ' ')
+      remove_punctuation(c[0])
     end
 
-    # create hash of citations (key) with counts (value)
-    Hash[clean_cites.group_by { |c| c }.map { |k, v| [k, v.count] }]
+    Hash[clean_cites.group_by { |c| c }.map { |cit, num| [cit, num.count] }]
+  end
+
+  ##
+  # Returns array of citation locations (index) in manuscript
+  #
+  #     contents = File.read('./spec/testfiles/test.txt')
+  #     PaperCitations.index_cite(contents, 'Peters et al. 2007')
+  #     # => [219, 500]
+  def index_of_cite(cite)
+    cite_parts = cite.split
+    author_s = cite_parts.take(cite_parts.size-1).join(' ')
+    year_s = cite_parts.last
+    @contents.enum_for(:scan, /(#{author_s}#{possessive}?#{year(year_s)})/
+                      ).map { Regexp.last_match.begin(0) }
   end
 
   private
 
-  def prefix
-    '(([dD]e|[vV]an[ ]?[dD]er)[ ]?)'
-  end
+  def prefix() '(([dD]e|[vV]an[ ]?[dD]er)[ ]?)' end
+  def author() "(#{prefix}?[A-Z][[:alpha:]\'\u2019\-]+)" end
+  def other_authors() "([ ]and[ ]#{author} | ([ ]et[ ]al.){1})" end
+  def possessive() "([\'\u2019]s|[\'\u2019])" end
+  def year_literal() "[1-2][0-9]{3}[a-z]?" end
+  def year(yr) "([ ][\(]?#{yr}[,\)\;])" end
 
-  def name
-    "(#{prefix}?[A-Z][[:alpha:]\'\u2019\-]+)"      # name: caps, accents, 's
+  def cite_match
+    /( #{author}{1}#{other_authors}?#{possessive}?#{year(year_literal)} )/x
   end
 
-  def cite_match
-    /(
-      #{name}{1}                                    # first author
-      ([ ]and[ ]#{name} | ([ ]et[ ]al.){1})?        # remaining authors
-      ([\'\u2019]s|s[\'\u2019])?                            # possessive form
-      ([ ][\(]?[1-2][0-9]{3}[a-z]?[,\)\;])          # year
-    )/x
+  def remove_punctuation(cite)
+    cite.gsub(/[\(\),;]|([\'\u2019]s)/, '').gsub(/[\'\u2019]\s/, ' ')
   end
 end

metadata

@@ -1,7 +1,7 @@
 --- !ruby/object:Gem::Specification
 name: citesight
 version: !ruby/object:Gem::Version
-  version: 0.0.4
+  version: 0.1.0
 platform: ruby
 authors:
 - Soumya Ray