X-Git-Url: https://git.netwichtig.de/gitweb/?a=blobdiff_plain;f=data%2Frbot%2Fplugins%2Fsearch.rb;h=a4c4a89c4aec0388fa697d80561fd40b51b79cd5;hb=983ad7a177ddb05519ff09b68d7bbf03bf9830a5;hp=c44100b1abaa106ba5367f683223b599200906ff;hpb=4911de2a2b6bc373ef7d5e73f34a14133c07b5e5;p=user%2Fhenk%2Fcode%2Fruby%2Frbot.git diff --git a/data/rbot/plugins/search.rb b/data/rbot/plugins/search.rb index c44100b1..a4c4a89c 100644 --- a/data/rbot/plugins/search.rb +++ b/data/rbot/plugins/search.rb @@ -1,65 +1,506 @@ -require 'uri' +#-- vim:sw=2:et +#++ +# +# :title: Google and Wikipedia search plugin for rbot +# +# Author:: Tom Gilbert (giblet) +# Author:: Giuseppe "Oblomov" Bilotta +# +# Copyright:: (C) 2002-2005 Tom Gilbert +# Copyright:: (C) 2006 Tom Gilbert, Giuseppe Bilotta +# Copyright:: (C) 2006-2007 Giuseppe Bilotta -Net::HTTP.version_1_2 +# TODO:: use lr=lang_ or whatever is most appropriate to let google know +# it shouldn't use the bot's location to find the preferred language +# TODO:: support localized uncyclopedias -- not easy because they have different names +# for most languages -GOOGLE_WAP_LINK = /(.*?)<\/a>/im +GOOGLE_SEARCH = "http://www.google.com/search?oe=UTF-8&q=" +GOOGLE_WAP_SEARCH = "http://www.google.com/m/search?hl=en&q=" +GOOGLE_WAP_LINK = /"r">(?:]*>)?]*>(.*?)<\/a>/im +GOOGLE_CALC_RESULT = %r{]*>(.+?)Results 1<\/b> - 10<\/b> of about (.*)<\/b> for} +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_SEARCH = "http://api.wolframalpha.com/v2/query?input=%{terms}&appid=%{key}&format=plaintext" + "&scantimeout=3.0&podtimeout=4.0&formattimeout=8.0&parsetimeout=5.0" + "&excludepodid=SeriesRepresentations:*" +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") + Config.register Config::IntegerValue.new('google.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('wikipedia.hits', + :default => 3, + :desc => "Number of hits to return from Wikipedia searches") + Config.register Config::IntegerValue.new('wikipedia.first_par', + :default => 1, + :desc => "When set to n > 0, the bot will return the first paragraph from the first n wikipedia search hits") + def help(plugin, topic="") case topic - when "search" - "search => search google for " - when "google" - "google => search google for " + when "ddg" + "Use '#{topic} ' to return a search or calculation from " + + "DuckDuckGo. Use #{topic} define to return a definition." + when "search", "google" + "#{topic} => search google for " + when "gcalc" + "gcalc => use the google calculator to find the answer to " + when "gdef" + "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 | wp => search for on Wikipedia" + "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) + 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 + # abstract is returned for definitions etc + abstract = xml.elements['//AbstractText/text()'].to_s + abfrom = "" + unless abstract.empty? + absrc = xml.elements['//AbstractSource/text()'].to_s + aburl = xml.elements['//AbstractURL/text()'].to_s + unless absrc.empty? and aburl.empty? + abfrom = " --" + abfrom << " " << absrc unless absrc.empty? + abfrom << " " << aburl unless aburl.empty? + end + end + + # but also definition (yes, you can have both, see e.g. printf) + definition = xml.elements['//Definition/text()'].to_s + deffrom = "" + unless definition.empty? + defsrc = xml.elements['//Definition/@source/text()'].to_s + defurl = xml.elements['//Definition/@url/text()'].to_s + unless defsrc.empty? and defurl.empty? + deffrom = " --" + deffrom << " " << defsrc unless defsrc.empty? + deffrom << " " << defurl unless defurl.empty? + end + end + + if heading.empty? and answer.empty? and abstract.empty? and definition.empty? + m.reply "no results" + return end + + # if we got a one-shot answer (e.g. a calculation, return it) + unless answer.empty? + m.reply answer + 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}%{abfrom}" % { + :bold => Bold, :heading => heading, + :abstract => abstract, :abfrom => abfrom + } + end + unless heading.empty? or definition.empty? + m.reply "%{bold}%{heading}:%{bold} %{abstract}%{abfrom}" % { + :bold => Bold, :heading => heading, + :abstract => definition, :abfrom => deffrom + } + 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) what = params[:words].to_s - searchfor = URI.escape what + if what.match(/^define:/) + return google_define(m, what, params) + end + + searchfor = CGI.escape what + # This method is also called by other methods to restrict searching to some sites if params[:site] site = "site:#{params[:site]}+" else site = "" end + # It is also possible to choose a filter to remove constant parts from the titles + # e.g.: "Wikipedia, the free encyclopedia" when doing Wikipedia searches + filter = params[:filter] || "" + + url = GOOGLE_WAP_SEARCH + site + searchfor - url = "http://www.google.com/wml/search?q=#{site}#{searchfor}" + hits = params[:hits] || @bot.config['google.hits'] + hits = 1 if params[:lucky] + first_pars = params[:firstpar] || @bot.config['google.first_par'] + + single = params[:lucky] || (hits == 1 and first_pars == 1) begin wml = @bot.httputil.get(url) + raise unless wml rescue => e m.reply "error googling for #{what}" return end results = wml.scan(GOOGLE_WAP_LINK) + if results.length == 0 m.reply "no results found for #{what}" return end - results = results[0...3].map { |res| - "#{res[0]}. #{Bold}#{Utils.decode_html_entities res[2].strip}#{Bold}: #{URI.unescape res[1].strip}" - }.join(" | ") - m.reply "Results for #{what}: #{results}" + single ||= (results.length==1) + pretty = [] + + begin + urls = Array.new + + debug results + results.each do |res| + t = res[1].ircify_html(:img => "[%{src} %{alt} %{dimensions}]").strip + u = res[0] + if u.sub!(%r{^http://www.google.com/aclk\?},'') + u = CGI::parse(u)['adurl'].first + debug "skipping ad for #{u}" + next + elsif u.sub!(%r{^http://www.google.com/gwt/x\?},'') + u = CGI::parse(u)['u'].first + elsif u.sub!(%r{^/url\?},'') + u = CGI::parse(u)['q'].first + end + urls.push(u) + pretty.push("%{n}%{b}%{t}%{b}%{sep}%{u}" % { + :n => (single ? "" : "#{urls.length}. "), + :sep => (single ? " -- " : ": "), + :b => Bold, :t => t, :u => u + }) + break if urls.length == hits + end + rescue => e + m.reply "failed to understand what google found for #{what}" + error e + debug wml + debug results + return + end + + if params[:lucky] + m.reply pretty.first + return + end + + result_string = pretty.join(" | ") + + # If we return a single, full result, change the output to a more compact representation + if single + m.reply "Result for %s: %s -- %s" % [what, result_string, Utils.get_first_pars(urls, first_pars)], :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_define(m, what, params) + begin + wml = @bot.httputil.get(GOOGLE_SEARCH + CGI.escape(what)) + raise unless wml + rescue => e + m.reply "error googling for #{what}" + return + end + + begin + related_index = wml.index(/Related phrases:/, 0) + raise unless related_index + defs_index = wml.index(/Definitions of /, related_index) + raise unless defs_index + defs_end = wml.index(/ e + m.reply "no results found for #{what}" + return + end + + related = wml[related_index...defs_index] + defs = wml[defs_index...defs_end] + + m.reply defs.ircify_html(:a_href => Underline), :split_at => (Underline + ' ') + + end + + def lucky(m, params) + params.merge!(:lucky => true) + google(m, params) + end + + def gcalc(m, params) + what = params[:words].to_s + searchfor = CGI.escape(what) + + debug "Getting gcalc thing: #{searchfor.inspect}" + url = GOOGLE_WAP_SEARCH + searchfor + + begin + html = @bot.httputil.get(url) + rescue => e + m.reply "error googlecalcing #{what}" + return + end + + debug "#{html.size} bytes of html recieved" + debug html + + candidates = html.match(GOOGLE_CALC_RESULT) + debug "candidates: #{candidates.inspect}" + + if candidates.nil? + m.reply "couldn't calculate #{what}" + return + end + result = candidates[1] + + debug "replying with: #{result.inspect}" + m.reply result.ircify_html + end + + def gcount(m, params) + what = params[:words].to_s + searchfor = CGI.escape(what) + + debug "Getting gcount thing: #{searchfor.inspect}" + url = GOOGLE_SEARCH + searchfor + + begin + html = @bot.httputil.get(url) + rescue => e + m.reply "error googlecounting #{what}" + return + end + + debug "#{html.size} bytes of html recieved" + + results = html.scan(GOOGLE_COUNT_RESULT) + debug "results: #{results.inspect}" + + if results.length != 1 + m.reply "couldn't count #{what}" + return + end + + result = results[0][0].ircify_html + debug "replying with: #{result.inspect}" + m.reply "total results: #{result}" + + end + + def gdef(m, params) + what = params[:words].to_s + searchfor = CGI.escape("define " + what) + + debug "Getting gdef thing: #{searchfor.inspect}" + url = GOOGLE_WAP_SEARCH + searchfor + + begin + html = @bot.httputil.get(url) + rescue => e + m.reply "error googledefining #{what}" + return + end + + debug html + results = html.scan(GOOGLE_DEF_RESULT) + debug "results: #{results.inspect}" + + if results.length != 1 + m.reply "couldn't find a definition for #{what} on Google" + return + end + + head = results[0][0].ircify_html + text = results[0][1].ircify_html + m.reply "#{head} -- #{text}" + end + + def wolfram(m, params) + what = params[:words].to_s + terms = CGI.escape what + url = WOLFRAM_API_SEARCH % { + :terms => terms, :key => WOLFRAM_API_KEY + } + + begin + feed = @bot.httputil.get(url) + raise unless feed + rescue => e + m.reply "error asking WolframAlfa about #{what}" + return + end + debug feed + + 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" + debug "Looking up things on #{site}" + params[:site] = site + params[:filter] = / - Wikipedia.*$/ + params[:hits] = @bot.config['wikipedia.hits'] + params[:firstpar] = @bot.config['wikipedia.first_par'] + return google(m, params) + end + + def unpedia(m, params) + site = "uncyclopedia.org" + debug "Looking up things on #{site}" params[:site] = site + params[:filter] = / - Uncyclopedia.*$/ + params[:hits] = @bot.config['wikipedia.hits'] + params[:firstpar] = @bot.config['wikipedia.first_par'] return google(m, params) end + + def gtime(m, params) + where = params[:words].to_s + where.sub!(/^\s*in\s*/, '') + searchfor = CGI.escape("time in " + where) + url = GOOGLE_SEARCH + searchfor + + begin + html = @bot.httputil.get(url) + rescue => e + m.reply "Error googletiming #{where}" + return + end + + debug html + results = html.scan(GOOGLE_TIME_RESULT) + debug "results: #{results.inspect}" + + if results.length != 1 + m.reply "Couldn't find the time for #{where} on Google" + return + end + + time = results[0][0].ircify_html + m.reply "#{time}" + end end plugin = SearchPlugin.new -plugin.map "search *words", :action => 'google' -plugin.map "google *words", :action => 'google' -plugin.map "wp :lang *words", :action => 'wikipedia', :requirements => { :lang => /^\w\w\w?$/ } -plugin.map "wp *words", :action => 'wikipedia' - +plugin.map "ddg *words", :action => 'duckduckgo', :threaded => true +plugin.map "search *words", :action => 'google', :threaded => true +plugin.map "google *words", :action => 'google', :threaded => true +plugin.map "lucky *words", :action => 'lucky', :threaded => true +plugin.map "gcount *words", :action => 'gcount', :threaded => true +plugin.map "gcalc *words", :action => 'gcalc', :threaded => true +plugin.map "gdef *words", :action => 'gdef', :threaded => true +plugin.map "gtime *words", :action => 'gtime', :threaded => true +plugin.map "wa *words", :action => 'wolfram', :threaded => true +plugin.map "wp :lang *words", :action => 'wikipedia', :requirements => { :lang => /^\w\w\w?$/ }, :threaded => true +plugin.map "wp *words", :action => 'wikipedia', :threaded => true +plugin.map "unpedia *words", :action => 'unpedia', :threaded => true