mine_shaft 0.1.0

data/spec/mine_shaft/html_table_spec.rb

@@ -0,0 +1,60 @@
+require 'spec_helper'
+require 'mine_shaft'
+include MineShaft
+
+describe HTMLTable do
+  let(:raw_table) {Nokogiri::HTML(File.read(fixture('projects_table.html')))}
+  let(:table) {HTMLTable.new(raw_table)}
+
+  describe "#deserialize" do
+    it "returns a Hash with the same number of elements as there are content rows (non-headings)" do
+      #raise table.deserialize.inspect
+      table.deserialize.should have(2).items
+    end
+
+    it "downcases & symbolizes the headings and uses them as the Hash keys" do
+      table.deserialize.first.keys.should have(2).items
+      table.deserialize.first.keys.should include(:name)
+      table.deserialize.first.keys.should include(:owner)
+    end
+
+    it "maps the associated value for each column to the heading hash key" do
+      table.deserialize.first[:name].should == "Sunbeam"
+      table.deserialize.first[:owner].should == "John Doe"
+    end
+  end
+
+  describe ".content_rows" do
+    it "returns a array with the same number of rows as the table (minus the heading)" do
+      table.content_rows.should have(2).items
+    end
+
+    it "does not include the content from the header" do
+      table.content_rows.flatten.should_not include("Name")
+      table.content_rows.flatten.should_not include("Owner")
+    end
+
+    it "does include the <td> content for non-header rows" do
+      table.content_rows.should include(["Sunbeam", "John Doe"])
+      table.content_rows.should include(["Truby", "Jane Doe"])
+    end
+  end
+
+  describe ".td_elements" do
+    it "returns an array of all the <td> elements in the table" do
+      table.should have(6).td_elements
+    end
+
+    it "returns the textual content in the <td> elements" do
+      table.td_elements.should include("Sunbeam")
+      table.td_elements.should include("Name")
+      table.td_elements.should include("John Doe")
+    end
+  end
+
+  describe ".headings" do
+    it "assumes the first row of <td>'s in the table is the header" do
+      table.headings.should == %w(Name Owner)
+    end
+  end
+end

data/spec/mine_shaft/shaft_spec.rb

@@ -0,0 +1,56 @@
+require 'spec_helper'
+require "mine_shaft"
+
+include MineShaft
+
+describe Shaft do
+  let(:base_uri) {'http://redmine.example.com'}
+  let(:login_page) {'/login'}
+  let(:shaft) {Shaft.new('username', 'password', base_uri)}
+
+  before(:each) do
+    agent = UserAgent.new('username', 'password', base_uri)
+    agent.stub!(:log_in).and_return(true)
+    UserAgent.stub(:new).with('username', 'password', base_uri).and_return(agent)
+  end
+
+  describe ".grab" do
+    let(:wiki_page) {'/projects/test/wiki/Wiki'}
+
+    context "when the specified table id does not exist on the page" do
+      before(:each) do
+        FakeWeb.register_uri(:get,
+                             "#{base_uri}#{wiki_page}",
+                             :body => fixture('wiki_page_with_no_table_id.html'),
+                             :content_type => 'text/html')
+      end
+
+      it "raises a InvalidPage exception" do
+        expect {
+          shaft.grab('my-table-id', wiki_page)
+        }.should raise_error(InvalidPage)
+      end
+    end
+
+    context "when the specified table id does not exist on the page" do
+      before(:each) do
+        FakeWeb.register_uri(:get,
+                             "#{base_uri}#{wiki_page}",
+                             :body => fixture('wiki_page_with_projects_table.html'),
+                             :content_type => 'text/html')
+      end
+
+      it "returns an array with the same number of elements as are in the table (minus header)" do
+        shaft.grab('projects', wiki_page).should have(2).items
+      end
+
+      it "returns elements with key/value pairs of heading-name/row-value for each row" do
+        results = shaft.grab('projects', wiki_page)
+        results.first[:name].should == "Sunbeam"
+        results.first[:owner].should == "John Doe"
+        results.last[:name].should == "Truby"
+        results.last[:owner].should == "Jane Doe"
+      end
+    end
+  end
+end

data/spec/mine_shaft/user_agent_spec.rb

@@ -0,0 +1,62 @@
+require 'spec_helper'
+require 'mine_shaft'
+include MineShaft
+
+describe UserAgent do
+  let(:base_uri) {'http://redmine.example.com'}
+  let(:login_page) {'/login'}
+  let(:agent) {UserAgent.new('username', 'password', base_uri)}
+
+  describe ".log_in" do
+    before(:each) do
+      FakeWeb.register_uri(:get, "#{base_uri}#{login_page}", response)
+    end
+
+    context "when it receives a 404 response code" do
+      let(:response) {{:body => "<html></html>", :content_type => 'text/html', :status => [404, "Not Found"]}}
+
+      it "raises an InvalidPage error" do
+        expect {agent.log_in}.should raise_error(InvalidPage)
+      end
+    end
+
+    context "when it does not find the login form" do
+      let(:response) {{:body => "<html></html>", :content_type => 'text/html'}}
+
+      it "raises a InvalidPage exception" do
+        expect {agent.log_in}.should raise_error(InvalidPage)
+      end
+    end
+
+    context "when it finds the login form" do
+      before(:each) do
+        login_response = {:body => fixture('login.html'), :content_type => 'text/html'}
+        FakeWeb.register_uri(:get, "#{base_uri}#{login_page}", login_response)
+      end
+
+      context "and a successful login occurs" do
+        let(:response) {{:body => fixture('home_page.html'), :content_type => 'text/html'}}
+
+        before(:each) do
+          FakeWeb.register_uri(:post, "#{base_uri}#{login_page}", response)
+        end
+
+        it "returns true" do
+          agent.log_in.should be_true
+        end
+      end
+
+      context "and the login fails (likely due to username/password)" do
+        let(:response) {{:body => fixture('failed_login.html'), :content_type => 'text/html'}}
+
+        before(:each) do
+          FakeWeb.register_uri(:post, "#{base_uri}#{login_page}", response)
+        end
+
+        it "raises a FailedLogin exception" do
+          expect {agent.log_in}.should raise_error(FailedLogin)
+        end
+      end
+    end
+  end
+end

data/spec/mine_shaft/web_page_spec.rb

@@ -0,0 +1,25 @@
+require 'spec_helper'
+require 'mine_shaft'
+include MineShaft
+
+describe WebPage do
+  describe ".find_table" do
+    let(:page) {WebPage.new(parsed_page)}
+
+    context "when the specified table id does not exist on the page" do
+      let(:parsed_page) {Nokogiri::HTML(File.read(fixture('wiki_page_with_no_table_id.html')))}
+
+      it "raises a InvalidPage exception" do
+        expect {page.find_table('my-table-id')}.should raise_error(InvalidPage)
+      end
+    end
+
+    context "when the specified table exists on the page" do
+      let(:parsed_page) {Nokogiri::HTML(File.read(fixture('wiki_page_with_projects_table.html')))}
+
+      it "returns an HTMLTable" do
+        page.find_table('projects').should be_instance_of(HTMLTable)
+      end
+    end
+  end
+end

data/spec/spec.opts

@@ -0,0 +1,4 @@
+--colour
+--format
+profile
+--diff

data/spec/spec_helper.rb

@@ -0,0 +1,8 @@
+require 'lib/mine_shaft'
+require 'fakeweb'
+
+Dir.glob('spec/support/**/*.rb').each {|f| require f}
+
+RSpec.configure do |config|
+  config.include MineShaftHelpers
+end

data/spec/support/mine_shaft_helpers.rb

@@ -0,0 +1,9 @@
+module MineShaftHelpers
+  def fixture(filename)
+    filepath = File.expand_path("spec/fixtures/#{filename}")
+    raise "Fixture file not found" unless File.exists?(filepath)
+    filepath
+  end
+end
+
+include MineShaftHelpers

metadata

@@ -0,0 +1,169 @@
+--- !ruby/object:Gem::Specification 
+name: mine_shaft
+version: !ruby/object:Gem::Version 
+  hash: 27
+  prerelease: false
+  segments: 
+  - 0
+  - 1
+  - 0
+  version: 0.1.0
+platform: ruby
+authors: 
+- Tom Kersten
+autorequire: 
+bindir: bin
+cert_chain: []
+
+date: 2010-11-27 00:00:00 -06:00
+default_executable: 
+dependencies: 
+- !ruby/object:Gem::Dependency 
+  name: mechanize
+  prerelease: false
+  requirement: &id001 !ruby/object:Gem::Requirement 
+    none: false
+    requirements: 
+    - - ~>
+      - !ruby/object:Gem::Version 
+        hash: 23
+        segments: 
+        - 1
+        - 0
+        - 0
+        version: 1.0.0
+  type: :runtime
+  version_requirements: *id001
+- !ruby/object:Gem::Dependency 
+  name: rake
+  prerelease: false
+  requirement: &id002 !ruby/object:Gem::Requirement 
+    none: false
+    requirements: 
+    - - ">="
+      - !ruby/object:Gem::Version 
+        hash: 3
+        segments: 
+        - 0
+        version: "0"
+  type: :development
+  version_requirements: *id002
+- !ruby/object:Gem::Dependency 
+  name: rspec
+  prerelease: false
+  requirement: &id003 !ruby/object:Gem::Requirement 
+    none: false
+    requirements: 
+    - - ~>
+      - !ruby/object:Gem::Version 
+        hash: 11
+        segments: 
+        - 2
+        - 1
+        - 0
+        version: 2.1.0
+  type: :development
+  version_requirements: *id003
+- !ruby/object:Gem::Dependency 
+  name: fakeweb
+  prerelease: false
+  requirement: &id004 !ruby/object:Gem::Requirement 
+    none: false
+    requirements: 
+    - - ~>
+      - !ruby/object:Gem::Version 
+        hash: 27
+        segments: 
+        - 1
+        - 3
+        - 0
+        version: 1.3.0
+  type: :development
+  version_requirements: *id004
+description: Provides an easy way to authenticate and grab content from a Redmine project. A specific use-case was to deserialize HTML tables from the wiki into an Array of Hashes with key/value pairs of "table heading" -> "column value" in order to simplify collection of a seed data from domain experts.
+email: 
+- tkersten@gnresound.com
+executables: []
+
+extensions: []
+
+extra_rdoc_files: []
+
+files: 
+- spec/mine_shaft/html_table_spec.rb
+- spec/mine_shaft/shaft_spec.rb
+- spec/mine_shaft/user_agent_spec.rb
+- spec/mine_shaft/web_page_spec.rb
+- spec/spec_helper.rb
+- spec/support/mine_shaft_helpers.rb
+- lib/mine_shaft/errors.rb
+- lib/mine_shaft/html_table.rb
+- lib/mine_shaft/login_page.rb
+- lib/mine_shaft/shaft.rb
+- lib/mine_shaft/user_agent.rb
+- lib/mine_shaft/version.rb
+- lib/mine_shaft/web_page.rb
+- lib/mine_shaft.rb
+- Rakefile
+- README.textile
+- CHANGELOG
+- LICENSE
+- spec/fixtures/failed_login.html
+- spec/fixtures/home_page.html
+- spec/fixtures/login.html
+- spec/fixtures/multiple_tables.html
+- spec/fixtures/projects_table.html
+- spec/fixtures/wiki_page_with_multiple_tables.html
+- spec/fixtures/wiki_page_with_no_table_id.html
+- spec/fixtures/wiki_page_with_projects_table.html
+- spec/spec.opts
+has_rdoc: true
+homepage: http://github.com/gn-research/mine_shaft
+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: mine_shaft
+rubygems_version: 1.3.7
+signing_key: 
+specification_version: 3
+summary: "Scrape & transform content from areas of a Redmine project which do not yet provide a REST API (ex: the wiki)"
+test_files: 
+- spec/fixtures/failed_login.html
+- spec/fixtures/home_page.html
+- spec/fixtures/login.html
+- spec/fixtures/multiple_tables.html
+- spec/fixtures/projects_table.html
+- spec/fixtures/wiki_page_with_multiple_tables.html
+- spec/fixtures/wiki_page_with_no_table_id.html
+- spec/fixtures/wiki_page_with_projects_table.html
+- spec/mine_shaft/html_table_spec.rb
+- spec/mine_shaft/shaft_spec.rb
+- spec/mine_shaft/user_agent_spec.rb
+- spec/mine_shaft/web_page_spec.rb
+- spec/spec.opts
+- spec/spec_helper.rb
+- spec/support/mine_shaft_helpers.rb