# TODO:: support localized uncyclopedias -- not easy because they have different names
# for most languages
-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">(?:<div[^>]*>)?<a href="([^"]+)"[^>]*>(.*?)<\/a>/im
-GOOGLE_CALC_RESULT = %r{<h[1-6] class="r" [^>]*>(.+?)</h}
-GOOGLE_COUNT_RESULT = %r{<font size=-1>Results <b>1<\/b> - <b>10<\/b> of about <b>(.*)<\/b> for}
-GOOGLE_DEF_RESULT = %r{onebox_result">\s*(.*?)\s*<br/>\s*(.*?)<table}
-GOOGLE_TIME_RESULT = %r{alt="Clock"></td><td valign=[^>]+>(.+?)<(br|/td)>}
-
+GOOGLE_SEARCH = "https://www.google.com/search?hl=en&oe=UTF-8&ie=UTF-8&gbv=1&q="
+GOOGLE_WAP_SEARCH = "https://www.google.com/m/search?hl=en&ie=UTF-8&gbv=1&q="
+GOOGLE_WAP_LINK = /<a\s+href="\/url\?(q=[^"]+)"[^>]*>\s*<div[^>]*>(.*?)\s*<\/div>/im
+GOOGLE_CALC_RESULT = />Calculator<\/span>(?:<\/?[^>]+>\s*)+([^<]+)/
+GOOGLE_XPATH_DEF = "//img[@id='flex_text_audio_icon_chunk']/../../../../div[3]//text()"
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"
+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"
+WOLFRAM_API_KEY = "9RW6XR-QTL2JT7J4W"
class SearchPlugin < Plugin
Config.register Config::IntegerValue.new('duckduckgo.hits',
"gcalc <equation> => use the google calculator to find the answer to <equation>"
when "gdef"
"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"
m.reply "error duckduckgoing for #{what}"
return
end
- debug feed
xml = REXML::Document.new feed
heading = xml.elements['//Heading/text()'].to_s
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()']
- aburl = xml.elements['//AbstractURL/text()']
+ 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?
# 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}" % {
+ m.reply "%{bold}%{heading}:%{bold} %{abstract}%{abfrom}" % {
:bold => Bold, :heading => heading,
- :abstract => abstract, :absrc => absrc, :aburl => aburl
+ :abstract => abstract, :abfrom => abfrom
}
end
unless heading.empty? or definition.empty?
- m.reply "%{bold}%{heading}:%{bold} %{abstract} -- %{absrc} %{aburl}" % {
+ m.reply "%{bold}%{heading}:%{bold} %{abstract}%{abfrom}" % {
:bold => Bold, :heading => heading,
- :abstract => definition, :absrc => defsrc, :aburl => defurl
+ :abstract => definition, :abfrom => deffrom
}
end
# return zeroclick search results
# 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)}" : ""
+ fp = first_pars > 0 ? " -- #{Utils.get_first_pars(@bot, links, first_pars)}" : ""
m.reply("Result for %{what}: %{string}%{fp}" % {
:what => what, :string => result_string, :fp => fp
}, :overlong => :truncate)
return unless first_pars > 0
- Utils.get_first_pars urls, first_pars, :message => m
+ Utils.get_first_pars(@bot, urls, first_pars, :message => m)
end
def google(m, params)
return
end
- single ||= (results.length==1)
- pretty = []
+ results = results.map {|result|
+ url = CGI::parse(Utils.decode_html_entities(result[0]))['q'].first
+ title = Utils.decode_html_entities(result[1].gsub(/<\/?[^>]+>/, ''))
+ [url, title] unless url.empty? or title.empty?
+ }.reject {|item| not item}[0..hits]
- 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
+ result_string = results.map {|url, title| "#{Bold}#{title}#{NormalText}: #{url}"}
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
+ m.reply result_string.first
+ Utils.get_first_pars(@bot, [results.map {|url, title| url}.first], first_pars, :message => m)
return
end
- m.reply "Results for #{what}: #{result_string}", :split_at => /\s+\|\s+/
+ m.reply "Results for #{what}: #{result_string.join(' | ')}", :split_at => /\s+\|\s+/
return unless first_pars > 0
- Utils.get_first_pars urls, first_pars, :message => m
-
+ Utils.get_first_pars(@bot, results.map {|url, title| url}, first_pars, :message => m)
end
def google_define(m, what, params)
m.reply "couldn't calculate #{what}"
return
end
- result = candidates[1]
+ result = candidates[1].remove_nonascii
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)
url = GOOGLE_WAP_SEARCH + searchfor
begin
- html = @bot.httputil.get(url)
+ resp = @bot.httputil.get(url, resp: true)
rescue => e
m.reply "error googledefining #{what}"
return
end
- debug html
- results = html.scan(GOOGLE_DEF_RESULT)
- debug "results: #{results.inspect}"
+ results = resp.xpath(GOOGLE_XPATH_DEF).map(&:content)
- if results.length != 1
+ if results.empty?
m.reply "couldn't find a definition for #{what} on Google"
- return
+ else
+ m.reply "#{results.first} -- #{results[1..-1].join(' ')}"
end
-
- head = results[0][0].ircify_html
- text = results[0][1].ircify_html
- m.reply "#{head} -- #{text}"
end
def wolfram(m, params)
m.reply "no data available"
return
end
- answer = []
- xml.elements.each("//pod/subpod/plaintext") { |element|
- answer << element.text
+ answer_type, answer = [], []
+ xml.elements.each("//pod") { |element|
+ answer_type << element.attributes['title']
+ answer << element.elements['subpod/plaintext'].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
+ # 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)
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', :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
projects / user / henk / code / ruby / rbot.git / blobdiff