X-Git-Url: https://git.netwichtig.de/gitweb/?a=blobdiff_plain;f=data%2Frbot%2Fplugins%2Fsearch.rb;h=e6803e58bfe8095250ec0ea8effa390411f335ff;hb=6cf365c49ce5fbe24c0a4ff0663550390b501fea;hp=3d9e5d3d354b95dd30b77ac02ef9cf6f3893db5d;hpb=ab9f1a23b6808b05df2a51c17dd3f83bbc8a8863;p=user%2Fhenk%2Fcode%2Fruby%2Frbot.git

diff --git a/data/rbot/plugins/search.rb b/data/rbot/plugins/search.rb
index 3d9e5d3d..e6803e58 100644
--- a/data/rbot/plugins/search.rb
+++ b/data/rbot/plugins/search.rb
@@ -23,12 +23,20 @@ GOOGLE_COUNT_RESULT = %r{<font size=-1>Results <b>1<\/b> - <b>10<\/b> of about <
 GOOGLE_DEF_RESULT = %r{onebox_result">\s*(.*?)\s*<br/>\s*(.*?)<table}
 GOOGLE_TIME_RESULT = %r{alt="Clock"></td><td valign=[^>]+>(.+?)<(br|/td)>}
 
-DDG_API_SEARCH = "http://api.duckduckgo.com/?format=xml&no_html=1&no_redirect=0&q="
+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")
@@ -55,12 +63,14 @@ class SearchPlugin < Plugin
       "gdef <term(s)> => use the google define mechanism to find a definition of <term(s)>"
     when "gtime"
       "gtime <location> => use the google clock to find the current time at <location>"
+    when "wa"
+      "wa <string> => searches WolframAlpha for <string>"
     when "wp"
       "wp [<code>] <string> => search for <string> on Wikipedia. You can select a national <code> to only search the national Wikipedia"
     when "unpedia"
       "unpedia <string> => search for <string> on Uncyclopedia"
     else
-      "search <string> (or: google <string>) => search google for <string> | ddg <string> to search DuckDuckGo | wp <string> => search for <string> on Wikipedia | unpedia <string> => search for <string> on Uncyclopedia"
+      "search <string> (or: google <string>) => search google for <string> | ddg <string> to search DuckDuckGo | wp <string> => search for <string> on Wikipedia | wa <string> => search for <string> on WolframAlpha | unpedia <string> => search for <string> on Uncyclopedia"
     end
   end
 
@@ -68,6 +78,11 @@ class SearchPlugin < Plugin
     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
@@ -81,39 +96,97 @@ class SearchPlugin < Plugin
     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
+    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 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}%{abfrom}" % {
+        :bold => Bold, :heading => heading,
+        :abstract => abstract, :abfrom => abfrom
       }
-      xml.elements.each("//RelatedTopics/RelatedTopic/Text") { |element|
-        text << " #{element.text}"
+    end
+    unless heading.empty? or definition.empty?
+      m.reply "%{bold}%{heading}:%{bold} %{abstract}%{abfrom}" % {
+        :bold => Bold, :heading => heading,
+        :abstract => definition, :abfrom => deffrom
       }
-      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 +406,54 @@ class SearchPlugin < Plugin
     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_type, answer = [], []
+    xml.elements.each("//pod") { |element|
+      answer_type << element.attributes['title']
+      answer << element.elements['subpod/plaintext'].text
+    }
+    # find the first answer that isn't nil,
+    # starting on the second pod in the array
+    n = 1
+    answer[1..-1].each { |a|
+      break unless a.nil?
+      n += 1
+    }
+    if answer[n].nil?
+      m.reply "no results"
+      return
+    end
+    # strip spaces, pipes, and line breaks
+    sep = Bold + ' :: ' + Bold
+    chars = [ [/\n/, sep], [/\t/, " "], [/\s+/, " "], ["|", "-"] ]
+    chars.each { |c| answer[n].gsub!(c[0], c[1]) }
+    m.reply answer_type[n] + sep + answer[n]
+  end
+
   def wikipedia(m, params)
     lang = params[:lang]
     site = "#{lang.nil? ? '' : lang + '.'}wikipedia.org"
@@ -383,7 +504,7 @@ end
 
 plugin = SearchPlugin.new
 
-plugin.map "ddg *words", :action => 'duckduckgo'
+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
@@ -391,6 +512,7 @@ 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