X-Git-Url: https://git.netwichtig.de/gitweb/?a=blobdiff_plain;f=data%2Frbot%2Fplugins%2Fsearch.rb;h=3cd45fe526b01b57c0615ff7daf3323b1b8ae3cf;hb=76308815c268c6f7d13a31594da8b9709d17564b;hp=aec8b1345611a2994a23f711e3573664ed7c6f81;hpb=5829ccecaf24f2adb47f8f98a778ba9e3d3cc90b;p=user%2Fhenk%2Fcode%2Fruby%2Frbot.git diff --git a/data/rbot/plugins/search.rb b/data/rbot/plugins/search.rb index aec8b134..3cd45fe5 100644 --- a/data/rbot/plugins/search.rb +++ b/data/rbot/plugins/search.rb @@ -17,14 +17,26 @@ 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 = /(.*?)<\/a>/im -GOOGLE_WAP_LINK = /(.*?)<\/a>/im -GOOGLE_CALC_RESULT = %r{.*?

]*>(.+?)} +GOOGLE_WAP_LINK = /"r">(?:]*>)?]*>(.*?)<\/a>/im +GOOGLE_CALC_RESULT = %r{]*>(.+?)Results 1<\/b> - 10<\/b> of about (.*)<\/b> for} -GOOGLE_DEF_RESULT = %r{]*>(Web definitions for .*?)
(.*?)
(.*?)\s-\s+\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") @@ -40,6 +52,9 @@ class SearchPlugin < Plugin def help(plugin, topic="") case topic + 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" @@ -48,15 +63,132 @@ 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 | 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) + 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 if what.match(/^define:/) @@ -98,27 +230,46 @@ class SearchPlugin < Plugin end single ||= (results.length==1) + pretty = [] - urls = Array.new - n = 0 - results = results[0...hits].map { |res| - n += 1 - t = res[2].ircify_html(:img => "[%{src} %{alt} %{dimensions}]").strip - u = URI.unescape(res[0] || res[1]) - urls.push(u) - "%{n}%{b}%{t}%{b}%{sep}%{u}" % { - :n => (single ? "" : "#{n}. "), - :sep => (single ? " -- " : ": "), - :b => Bold, :t => t, :u => u - } - } + 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 results.first + m.reply pretty.first return end - result_string = results.join(" | ") + result_string = pretty.join(" | ") # If we return a single, full result, change the output to a more compact representation if single @@ -172,7 +323,7 @@ class SearchPlugin < Plugin searchfor = CGI.escape(what) debug "Getting gcalc thing: #{searchfor.inspect}" - url = GOOGLE_SEARCH + searchfor + url = GOOGLE_WAP_SEARCH + searchfor begin html = @bot.httputil.get(url) @@ -182,18 +333,19 @@ class SearchPlugin < Plugin end debug "#{html.size} bytes of html recieved" + debug html - results = html.scan(GOOGLE_CALC_RESULT) - debug "results: #{results.inspect}" + candidates = html.match(GOOGLE_CALC_RESULT) + debug "candidates: #{candidates.inspect}" - if results.length != 1 + if candidates.nil? m.reply "couldn't calculate #{what}" return end + result = candidates[1] - result = results[0][0].ircify_html debug "replying with: #{result.inspect}" - m.reply "#{result}" + m.reply result.ircify_html end def gcount(m, params) @@ -249,21 +401,45 @@ class SearchPlugin < Plugin return end - gdef_link = "http://www.google.com" + CGI.unescapeHTML(results[0][0]) # could be used to extract all defs - head = results[0][1].ircify_html - text = results[0][2].ircify_html - link = results[0][3] - m.reply "#{head} -- #{link}\n#{text}" + head = results[0][0].ircify_html + text = results[0][1].ircify_html + m.reply "#{head} -- #{text}" + end - ### gdef_link could be used for something like - # html_defs = @bot.httputil.get(gdef_link) - # related_index = html_defs.index(/Related phrases:/, 0) - # defs_index = html_defs.index(/Definitions of /, related_index) + def wolfram(m, params) + what = params[:words].to_s + terms = CGI.escape what + url = WOLFRAM_API_SEARCH % { + :terms => terms, :key => WOLFRAM_API_KEY + } - # related = html_defs[related_index..defs_index] - # defs = html_defs[defs_index..-1] + begin + feed = @bot.httputil.get(url) + raise unless feed + rescue => e + m.reply "error asking WolframAlfa about #{what}" + return + end + debug feed - # m.reply defs.gsub('
','
  • ').ircify_html + 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) @@ -316,6 +492,7 @@ end plugin = SearchPlugin.new +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 @@ -323,7 +500,7 @@ 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', :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 -