X-Git-Url: https://git.netwichtig.de/gitweb/?a=blobdiff_plain;f=data%2Frbot%2Fplugins%2Fsearch.rb;h=fdd9722abdded51860f008b805004cbfef23f586;hb=259e147cdfdda54db4d2a2ee3162a0de9b0e1f73;hp=515cd654ade7bba4f99326661d711f1fc9f94f72;hpb=f873da03d1565aa50d103ea4c585b28f99839058;p=user%2Fhenk%2Fcode%2Fruby%2Frbot.git diff --git a/data/rbot/plugins/search.rb b/data/rbot/plugins/search.rb index 515cd654..fdd9722a 100644 --- a/data/rbot/plugins/search.rb +++ b/data/rbot/plugins/search.rb @@ -23,10 +23,16 @@ GOOGLE_COUNT_RESULT = %r{Results 1<\/b> - 10<\/b> of about < GOOGLE_DEF_RESULT = %r{onebox_result">\s*(.*?)\s*
\s*(.*?)]+>(.+?)<(br|/td)>} +DDG_API_SEARCH = "http://api.duckduckgo.com/?format=xml&no_html=1&skip_disambig=1&no_redirect=0&q=" +WOLFRAM_API_KEY = "4EU37Y-TX9WJG3JH3" + class SearchPlugin < Plugin Config.register Config::IntegerValue.new('duckduckgo.hits', :default => 3, :validate => Proc.new{|v| v > 0}, :desc => "Number of hits to return from searches") + Config.register Config::IntegerValue.new('duckduckgo.first_par', + :default => 0, + :desc => "When set to n > 0, the bot will return the first paragraph from the first n search hits") Config.register Config::IntegerValue.new('google.hits', :default => 3, :desc => "Number of hits to return from Google searches") @@ -53,67 +59,117 @@ class SearchPlugin < Plugin "gdef => use the google define mechanism to find a definition of " when "gtime" "gtime => use the google clock to find the current time at " + when "wa" + "wa => searches WolframAlpha for " when "wp" "wp [] => search for on Wikipedia. You can select a national to only search the national Wikipedia" when "unpedia" "unpedia => search for on Uncyclopedia" else - "search (or: google ) => search google for | ddg to search DuckDuckGo | wp => search for on Wikipedia | unpedia => search for on Uncyclopedia" + "search (or: google ) => search google for | ddg to search DuckDuckGo | wp => search for on Wikipedia | wa => search for on WolframAlpha | unpedia => search for on Uncyclopedia" end end def duckduckgo(m, params) - terms = params[:words].to_s - # DuckDuckGo is picky about white spaces - # in the url, so we can't use CGI.escape. - terms.gsub!(/\%/, '%25') if terms.include? "%" - terms.gsub!(/\+|\s\+\s|\s\+|\+\s/, ' %2B ') if terms.include? "+" - terms.gsub!(/\-|\s\-\s|\s\-|\-\s/, ' %2D ') if terms.include? "-" - terms.gsub!(/\!/, '%21') if terms.include? "!" - terms.gsub!(/\%/, ' %') if terms.include? "%" - feed = Net::HTTP.get 'api.duckduckgo.com', - "/?q=#{terms}&format=xml&skip_disambig=1&no_html=1&no_redirect=0" - if feed.nil? or feed.empty? - m.reply "error connecting" + what = params[:words].to_s + terms = CGI.escape what + url = DDG_API_SEARCH + terms + + hits = @bot.config['duckduckgo.hits'] + first_pars = params[:firstpar] || @bot.config['duckduckgo.first_par'] + single = params[:lucky] || (hits == 1 and first_pars == 1) + + begin + feed = @bot.httputil.get(url) + raise unless feed + rescue => e + m.reply "error duckduckgoing for #{what}" return end + debug feed + xml = REXML::Document.new feed heading = xml.elements['//Heading/text()'].to_s # answer is returned for calculations answer = xml.elements['//Answer/text()'].to_s - if heading.empty? and answer.empty? + # abstract is returned for definitions etc + abstract = xml.elements['//AbstractText/text()'].to_s + unless abstract.empty? + absrc = xml.elements['//AbstractSource/text()'] + aburl = xml.elements['//AbstractURL/text()'] + end + # but also definition (yes, you can have both, see e.g. printf) + definition = xml.elements['//Definition/text()'].to_s + unless definition.empty? + defsrc = xml.elements['//Definition/@source/text()'].to_s + defurl = xml.elements['//Definition/@url/text()'].to_s + end + + if heading.empty? and answer.empty? and abstract.empty? and definition.empty? m.reply "no results" return end - if terms =~ /^define/ - if heading.empty? - m.reply "no definition found" - return - end - # Format and return a different string if it is a definition search. - definition = xml.elements['//AbstractText/text()'].to_s - source = " -- #{xml.elements['//AbstractURL/text()']}" - m.reply Bold + heading + ": " + Bold + definition + source - elsif heading.empty? - # return a calculation + + # if we got a one-shot answer (e.g. a calculation, return it) + unless answer.empty? m.reply answer - else - # else, return a zeroclick search - links, text = [], [] - hits = @bot.config['duckduckgo.hits'] - xml.elements.each("//RelatedTopics/RelatedTopic/FirstURL") { |element| - links << element.text + return + end + + # otherwise, return the abstract, followed by as many hits as found + unless heading.empty? or abstract.empty? + m.reply "%{bold}%{heading}:%{bold} %{abstract} -- %{absrc} %{aburl}" % { + :bold => Bold, :heading => heading, + :abstract => abstract, :absrc => absrc, :aburl => aburl } - xml.elements.each("//RelatedTopics/RelatedTopic/Text") { |element| - text << " #{element.text}" + end + unless heading.empty? or definition.empty? + m.reply "%{bold}%{heading}:%{bold} %{abstract} -- %{absrc} %{aburl}" % { + :bold => Bold, :heading => heading, + :abstract => definition, :absrc => defsrc, :aburl => defurl } - num = 0 - m.reply Bold + heading + ": " + Bold - until num >= hits - m.reply links[num] + text[num] - num += 1 - end end + # return zeroclick search results + links, texts = [], [] + xml.elements.each("//Results/Result/FirstURL") { |element| + links << element.text + break if links.size == hits + } + return if links.empty? + + xml.elements.each("//Results/Result/Text") { |element| + texts << " #{element.text}" + break if links.size == hits + } + # TODO see treatment of `single` in google search + + single ||= (links.length == 1) + pretty = [] + links.each_with_index do |u, i| + t = texts[i] + pretty.push("%{n}%{b}%{t}%{b}%{sep}%{u}" % { + :n => (single ? "" : "#{i}. "), + :sep => (single ? " -- " : ": "), + :b => Bold, :t => t, :u => u + }) + end + + result_string = pretty.join(" | ") + + # If we return a single, full result, change the output to a more compact representation + if single + fp = first_pars > 0 ? " -- #{Utils.get_first_pars(links, first_pars)}" : "" + m.reply("Result for %{what}: %{string}%{fp}" % { + :what => what, :string => result_string, :fp => fp + }, :overlong => :truncate) + return + end + + m.reply "Results for #{what}: #{result_string}", :split_at => /\s+\|\s+/ + + return unless first_pars > 0 + + Utils.get_first_pars urls, first_pars, :message => m end def google(m, params) @@ -333,6 +389,36 @@ class SearchPlugin < Plugin m.reply "#{head} -- #{text}" end + def wolfram(m, params) + terms = CGI.escape(params[:words].to_s) + feed = Net::HTTP.get 'api.wolframalpha.com', + "/v2/query?input=#{terms}&appid=#{WOLFRAM_API_KEY}&format=plaintext" + "&scantimeout=3.0&podtimeout=4.0&formattimeout=8.0&parsetimeout=5.0" + "&excludepodid=SeriesRepresentations:*" + if feed.nil? or feed.empty? + m.reply "error connecting" + return + end + xml = REXML::Document.new feed + if xml.elements['/queryresult'].attributes['error'] == "true" + m.reply xml.elements['/queryresult/error/text()'].to_s + return + end + unless xml.elements['/queryresult'].attributes['success'] == "true" + m.reply "no data available" + return + end + answer = [] + xml.elements.each("//pod/subpod/plaintext") { |element| + answer << element.text + } + # strip spaces and line breaks + answer[1].gsub!(/\n/, Bold + ' :: ' + Bold ) + answer[1].gsub!(/\t/, ' ') + answer[1].gsub!(/\s+/, ' ') + m.reply answer[1].to_s + end + def wikipedia(m, params) lang = params[:lang] site = "#{lang.nil? ? '' : lang + '.'}wikipedia.org" @@ -391,6 +477,7 @@ plugin.map "gcount *words", :action => 'gcount', :threaded => true plugin.map "gcalc *words", :action => 'gcalc', :threaded => true plugin.map "gdef *words", :action => 'gdef', :threaded => true plugin.map "gtime *words", :action => 'gtime', :threaded => true +plugin.map "wa *words", :action => 'wolfram' plugin.map "wp :lang *words", :action => 'wikipedia', :requirements => { :lang => /^\w\w\w?$/ }, :threaded => true plugin.map "wp *words", :action => 'wikipedia', :threaded => true plugin.map "unpedia *words", :action => 'unpedia', :threaded => true