fech-search 0.0.2 → 0.1.0

checksums.yaml

@@ -1,7 +1,7 @@
 ---
 SHA1:
-  metadata.gz: 90335ce5578a65937b78fef37e814bf8e44e1eeb
-  data.tar.gz: 6af0151ace9bece2b377c8599f8017a729482113
+  metadata.gz: dd2966cc82225142ba65490ea702ce5a2c43b024
+  data.tar.gz: 2ba7b2ae6ff2fcf40d325684b9e44a78ce540ced
 SHA512:
-  metadata.gz: 7a60a65bc1673108369a12296c25b6de2327adc7c120444ad5896cdd3e8077f3bc4aa977bd6a6f3d28d2922b2d4928e17dedbfb7c07e038a72d449d8b51da9f9
-  data.tar.gz: ea87f7e55dc46809b444dd158ab72a523c535a2dce4cadbba04b577f58d7ca9d4191c1923c6bf9f25d0d2a0e521720f9f8c5ef1d0ae7d5fc4c17d3dccab23e31
+  metadata.gz: 563565fb0c897400bcd8973a9e38184e63d63a7f9eed3f65a4688864636bbd8a9eae18c34716b5ac92eda6f77e525d52c9614a1a60b3fbe78ed62d03ab097c1a
+  data.tar.gz: 92a475a7072ad9ecc0f448daf9d05807da4594fc82e404eb3494582528dfc10a72d634ec37f34d4df246655c4b5adf0745828a607e01c9d991a5f1bc3c875339

data/fech-search.gemspec

@@ -16,6 +16,7 @@ Gem::Specification.new do |gem|
   gem.version       = Fech::Search::VERSION
 
   gem.add_dependency "fech", "~> 1.6.2"
+  gem.add_dependency "nokogiri", "1.6.6.2"
 
   gem.add_development_dependency "rspec", "~> 2.13.0"
 end

data/lib/fech-search/search.rb

@@ -1,4 +1,5 @@
 require 'net/http'
+require 'nokogiri'
 
 module Fech
 
@@ -11,6 +12,7 @@ module Fech
     def initialize(search_params={})
       @search_params = validate_params(make_params(search_params))
       @search_url = 'http://query.nictusa.com/cgi-bin/dcdev/forms/'
+      @search_url = 'http://docquery.fec.gov/cgi-bin/forms/'
       @response = search
     end
 
@@ -63,66 +65,21 @@ module Fech
     # results if called directly, or will yield the results one
     # by one if a block is passed.
     def results(&block)
-      if @search_params['date'] != ''
-        results_from_date_search(&block)
-      else
-        results_from_nondate_search(&block)
-      end
-    end
-
-    # Parse the results from a search that does not include a date.
-    # Will return an array of results if called directly, or will
-    # yield the results one by one if a block is passed.
-    def results_from_nondate_search(&block)
+      lines = body.split("\n")
+      parsing = false
+      committee = nil
       parsed_results = []
-      regex = /<DT>(.*?)<P/m
-      match = body.match regex
-      return [] if match.nil?
-      content = match[1]
-      committee_sections = content.split(/<DT>/)
-      committee_sections.each do |section|
-        data = parse_committee_section(section)
-        data.each do |result|
-          search_result = SearchResult.new(result)
-
-          if block_given?
-            yield search_result
-          else
-            parsed_results << search_result
-          end
+      lines.each do |line|
+        if line.match(/^<TABLE style='border:0;'>$/)
+          parsing = true
         end
-      end
-      block_given? ? nil : parsed_results
-    end
-
-    # For results of a search that does not include a date, parse
-    # the section giving information on the committee that submitted
-    # the filing.
-    # @param [String] section
-    def parse_committee_section(section)
-      data = []
-      section.gsub!(/^<BR>/, '')
-      rows = section.split(/\n/)
-      committee_data = parse_committee_row(rows.first)
-      rows[1..-1].each do |row|
-        data << committee_data.merge(parse_filing_row(row))
-      end
-      data
-    end
-
-    # Parse the results from a search that includes a date.
-    # Will return an array of results if called directly, or will
-    # yield the results one by one if a block is passed.
-    def results_from_date_search(&block)
-      parsed_results = []
-      dl = body.match(/<DL>(.*?)<BR><P/m)[0]
-      rows = dl.split('<DT>')[1..-1].map { |row| row.split("\n") }
-      rows.each do |committee, *filings|
-        committee = parse_committee_row(committee)
-        filings.each do |filing|
-          next if filing == "<BR><P"
-          data = committee.merge(parse_filing_row(filing))
-          search_result = SearchResult.new(data)
+        next unless parsing
+        if line.match(/<td colspan='8'>/)
+          committee = parse_committee_line(line)
+        end
+        if line.match(/>FEC-\d+</)
+          merged = parse_filing_line(line).merge(committee)
+          search_result = SearchResult.new(merged)
           if block_given?
             yield search_result
           else
@@ -130,55 +87,21 @@ module Fech
           end
         end
       end
-      block_given? ? nil : parsed_results
+      parsed_results
     end
 
-    # For results of a search that includes a date, parse
-    # the portion of the results with information on the
-    # committee that submitted the filing.
-    # @param [String] row
-    # @return [Hash] the committee name and ID
-    def parse_committee_row(row)
-      regex = /
-              '>
-              (.*?)
-              \s-\s
-              (C\d{8})
-              /x
-      match = row.match regex
-      {:committee_name => match[1], :committee_id => match[2]}
+    # Parse a line that contains committee information
+    def parse_committee_line(line)
+      match = line.match(/<A.*?>(?<name>.*?) - (?<id>C\d{8})<\/A/)
+      {:committee_name => match['name'], :committee_id => match['id']}
     end
 
-    # Parse a result row with information on the filing itself.
-    # @param [String] row
-    # @return [Hash] the filing ID, form type, period, date filed, description
-    # and, optionally, the filing that amended this filing.
-    def parse_filing_row(row)
-      regex = /
-              FEC-(\d+)
-              \s
-              Form
-              \s
-              (F.*?)
-              \s\s-\s
-              (period\s([-\/\d]+),\s)?
-              filed
-              \s
-              ([\/\d]+)
-              \s
-              (-\s
-               (.*?)
-               ($|<BR>.*?FEC-(\d+))
-              )?
-              /x
-      match = row.match regex
-      {:filing_id => match[1],
-       :form_type => match[2],
-       :period => match[4],
-       :date_filed => match[5],
-       :description => match[7],
-       :amended_by => match[9]
-      }
+    # Parse a line that contains a filing
+    def parse_filing_line(line)
+      doc = Nokogiri::HTML(line)
+      cells = doc.css("td").map(&:text)
+      fields = [:form_type, :filing_id, :amended_by, :from, :to, :date_filed, :description]
+      Hash[fields.zip(cells)]
     end
 
   end

data/lib/fech-search/search_result.rb

@@ -11,26 +11,30 @@ module Fech
 
       @committee_name = attrs[:committee_name]
       @committee_id   = attrs[:committee_id]
-      @filing_id      = attrs[:filing_id]
+      @filing_id      = attrs[:filing_id].sub(/FEC-/, '').to_i
       @form_type      = attrs[:form_type]
-      @period         = parse_period(attrs[:period])
+      @period         = parse_period(attrs[:from], attrs[:to])
       @date_filed     = Date.strptime(attrs[:date_filed], @date_format)
       @description    = attrs[:description]
       @amended_by     = attrs[:amended_by]
     end
 
-    # Parse the string representing a filing period.
+    # Parse the strings representing a filing period.
     # @param [String] period a string representing a filing period
     # @return [Hash, nil] a hash representing the start and end
     # of a filing period.
-    def parse_period(period)
-      return if period.nil?
-      from, to = period.split('-')
+    def parse_period(from, to)
+      return unless valid_date(from.to_s) && valid_date(to.to_s)
       from = Date.strptime(from, @date_format)
       to = Date.strptime(to, @date_format)
       {:from => from, :to => to}
     end
 
+    # Check whether a date string is valid
+    def valid_date(s)
+      s.match(/\d\d?\/\d\d?\/\d{4}/)
+    end
+
     # The Fech filing object for this search result
     # @return [Fech::Filing]
     def filing(opts={})

data/lib/fech-search/version.rb

@@ -1,5 +1,5 @@
 module Fech
   class Search
-    VERSION = "0.0.2"
+    VERSION = "0.1.0"
   end
 end

data/spec/search_result_spec.rb

@@ -15,10 +15,10 @@ describe Fech::SearchResult do
 
     it "should have valid dates for period" do
       from = @committee_result.period[:from]
-      from.class.should == Date
+      expect(from.class).to eq(Date)
       to = @committee_result.period[:to]
-      to.class.should == Date
-      to.should >= from
+      expect(to.class).to eq(Date)
+      expect(to).to be >= from
     end
 
   end
@@ -26,18 +26,18 @@ describe Fech::SearchResult do
   context "filing" do
 
     it "should make filing accessible" do
-      @date_result.filing.class.should == Fech::Filing
-      @committee_result.filing.class.should == Fech::Filing
+      expect(@date_result.filing.class).to eq(Fech::Filing)
+      expect(@committee_result.filing.class).to eq(Fech::Filing)
     end
 
     it "should create filing object for correct filing" do
       filing = @committee_result.filing
-      filing.filing_id.should == @committee_result.filing_id
+      expect(filing.filing_id).to eq(@committee_result.filing_id)
     end
 
     it "should pass arguments to Fech::Filing#new" do
       filing = @date_result.filing(:download_dir => "/tmp")
-      filing.download_dir.should == "/tmp"
+      expect(filing.download_dir).to eq("/tmp")
     end
 
   end

data/spec/search_spec.rb

@@ -17,11 +17,11 @@ describe Fech::Search do
     end
 
     it "should not raise an error if :committee_id is given by itself" do
-      expect { Fech::Search.new(:committee_id => "C00431171") }.to_not raise_error(ArgumentError)
+      expect { Fech::Search.new(:committee_id => "C00431171") }.to_not raise_error
     end
 
     it "should not raise an error if :form_type is given with another parameter" do
-      expect { Fech::Search.new(:form_type => "F3", :date => Date.new(2013, 5, 29)) }.to_not raise_error(ArgumentError)
+      expect { Fech::Search.new(:form_type => "F3", :date => Date.new(2013, 5, 29)) }.to_not raise_error
     end
 
   end
@@ -30,20 +30,20 @@ describe Fech::Search do
 
     it "should return an array of results from committee search" do
       search = Fech::Search.new(:committee_id => "C00431171")
-      search.results.class.should == Array
-      search.results[0].class.should == Fech::SearchResult
+      expect(search.results.class).to eq(Array)
+      expect(search.results[0].class).to eq(Fech::SearchResult)
     end
 
     it "should return an array of results from date search" do
       search = Fech::Search.new(:date => Date.new(2013, 5, 29))
-      search.results.class.should == Array
-      search.results[0].class.should == Fech::SearchResult
+      expect(search.results.class).to eq(Array)
+      expect(search.results[0].class).to eq(Fech::SearchResult)
     end
 
     it "should return an array of results from date and report_type search" do
       search = Fech::Search.new(:date => Date.new(2013, 5, 29), :report_type => "M4")
-      search.results.class.should == Array
-      search.results[0].class.should == Fech::SearchResult
+      expect(search.results.class).to eq(Array)
+      expect(search.results[0].class).to eq(Fech::SearchResult)
     end
 
   end

metadata

@@ -1,41 +1,55 @@
 --- !ruby/object:Gem::Specification
 name: fech-search
 version: !ruby/object:Gem::Version
-  version: 0.0.2
+  version: 0.1.0
 platform: ruby
 authors:
 - Aaron Bycoffe
 autorequire: 
 bindir: bin
 cert_chain: []
-date: 2014-01-22 00:00:00.000000000 Z
+date: 2015-04-16 00:00:00.000000000 Z
 dependencies:
 - !ruby/object:Gem::Dependency
   name: fech
   requirement: !ruby/object:Gem::Requirement
     requirements:
-    - - ~>
+    - - "~>"
       - !ruby/object:Gem::Version
         version: 1.6.2
   type: :runtime
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
-    - - ~>
+    - - "~>"
       - !ruby/object:Gem::Version
         version: 1.6.2
+- !ruby/object:Gem::Dependency
+  name: nokogiri
+  requirement: !ruby/object:Gem::Requirement
+    requirements:
+    - - '='
+      - !ruby/object:Gem::Version
+        version: 1.6.6.2
+  type: :runtime
+  prerelease: false
+  version_requirements: !ruby/object:Gem::Requirement
+    requirements:
+    - - '='
+      - !ruby/object:Gem::Version
+        version: 1.6.6.2
 - !ruby/object:Gem::Dependency
   name: rspec
   requirement: !ruby/object:Gem::Requirement
     requirements:
-    - - ~>
+    - - "~>"
       - !ruby/object:Gem::Version
         version: 2.13.0
   type: :development
   prerelease: false
   version_requirements: !ruby/object:Gem::Requirement
     requirements:
-    - - ~>
+    - - "~>"
       - !ruby/object:Gem::Version
         version: 2.13.0
 description: A Fech plugin for searching electronic FEC filings
@@ -45,7 +59,7 @@ executables: []
 extensions: []
 extra_rdoc_files: []
 files:
-- .gitignore
+- ".gitignore"
 - Gemfile
 - Gemfile.lock
 - LICENSE
@@ -67,17 +81,17 @@ require_paths:
 - lib
 required_ruby_version: !ruby/object:Gem::Requirement
   requirements:
-  - - ! '>='
+  - - ">="
     - !ruby/object:Gem::Version
       version: '0'
 required_rubygems_version: !ruby/object:Gem::Requirement
   requirements:
-  - - ! '>='
+  - - ">="
     - !ruby/object:Gem::Version
       version: '0'
 requirements: []
 rubyforge_project: 
-rubygems_version: 2.0.3
+rubygems_version: 2.2.2
 signing_key: 
 specification_version: 4
 summary: Fech-Search provides an interface for searching for electronic filings submitted
@@ -85,4 +99,3 @@ summary: Fech-Search provides an interface for searching for electronic filings
 test_files:
 - spec/search_result_spec.rb
 - spec/search_spec.rb
-has_rdoc: