Advanced Search

人工無脳のソース　アルゴリズムコンテスト勉強会にて

/ Published in: Python
Save to your folder(s)
アルゴリズムコンテストの勉強会で紹介したGoogle App Engineで動くpythonコードです
Djangoフレームワークを前提にしています。
あまりコードとしては参考にならないかも知れませんが、これだけでも簡単な人工無脳ができるんですね

Expand | Embed | Plain Text

Copy this code and paste it in your HTML 
  1. #!-*- coding:utf-8 -*-
  2. import sys
  3. import os
  4. import string
  5. import urllib
  6. import urllib2
  7. import math
  8. import htmlentitydefs
  9. import re
  10. from django.shortcuts import render_to_response
  11. from google.appengine.api import users
  12. from django.http import HttpResponseRedirect
  13. from BeautifulSoup import BeautifulSoup
  14. from google.appengine.api import urlfetch
  15. from base64 import b64encode
  16.  
  17.  
  18. #初期ページ
  19. def Mainpage(request):
  20.     return render_to_response('input.html',)
  21.  
  22.  
  23. #質問をもらって回答を出すまでのメソッド
  24. def getAnswer(request):
  25.     search_text = ''
  26.     from_user = 'username'
  27.     try:
  28.         search_text=request.POST['query']
  29.         search_text=search_text.encode('utf-8')
  30.     except:pass
  31.     if search_text == '':
  32.         search_text,from_user = GetresponseTwitter()
  33.         #print >>sys.stderr,'search_text=%s'%search_text
  34.         search_text = search_text.encode('utf-8')
  35.     to_ = ''
  36.     from_ = ''
  37.  
  38.     #質問文をそのまま検索にかける
  39.     if search_text!='' :
  40.         url = 'http://pcod.no-ip.org/yats/search?query=%s&rss' % (search_text)
  41.         xml_search2=urlfetch.fetch(url).content
  42.         entry = BeautifulSoup(xml_search2)('entry')
  43.         words=[]
  44.         titledict=dict()
  45.         answer = ''
  46.         for en in entry:
  47.             #print >>sys.stderr,en('summary')[0]
  48.             words=en('summary')[0](text = True)[0].split(' ')
  49.             update_before = en('updated')[0](text = True)[0]
  50.             #print >>sys.stderr,status_id
  51.             try:
  52.                 regexp = re.compile('username')
  53.                 if regexp.search(en('summary')[0](text = True)[0])== None:#tam_botの発言は無視
  54.                     from_=words[0].split('@')[1]
  55.                     to_=words[2].split('@')[1]
  56.  
  57.                     url='http://pcod.no-ip.org/yats/search?query=user:'+to_+'%20%40'+from_+'%20&lang=ja&rss'
  58.                     #print >>sys.stderr,url
  59.                     xml_search=urlfetch.fetch(url).content
  60.                     title1to2 = BeautifulSoup(xml_search)('entry')
  61.                     for x in title1to2:
  62.                         #titledict[x('updated')[0](text = True)[0]]=x('summary')[0].contents[0]
  63.                         update_after=x('updated')[0](text = True)[0]
  64.                         if  update_after > update_before:
  65.                             #print >>sys.stderr,'update_before=%s update_after=%s'%(update_before,update_after)
  66.                             tmp_answer=x('summary')[0](text = True)[0].split(' ')[3]
  67.                             regexp1 = re.compile('@')
  68.                             if regexp1.search(tmp_answer)==None and tmp_answer!='':
  69.                                 print >>sys.stderr,'answer=%s'%tmp_answer
  70.                                 answer = tmp_answer
  71.                                 break#一個でもみつかればbreak
  72.                     if tmp_answer != '':
  73.                         break
  74.             except:pass
  75.         #print >>sys.stderr,'answer=%s'%answer
  76.  
  77.  
  78.         if answer != '':#検索結果があった場合
  79.             post = '@%s %s'%(from_user,answer)
  80.         else:#検索結果がなかった場合
  81.             #############文章だけで回答が出せない場合#############
  82.             question_word=GetKeitaiso_mecapi(search_text).encode('utf-8')
  83.             q = {'query':question_word,'lang' : 'ja'}            
  84.             url = 'http://pcod.no-ip.org/yats/search?rss&' + urllib.urlencode(q)
  85.             xml=urlfetch.fetch(url).content
  86.             #print >>sys.stderr,xml
  87.             entry = BeautifulSoup(xml)('entry')
  88.             for en in entry:
  89.                 tmp_answer=en('summary')[0](text = True)[0].split(' : ')[1].rstrip()
  90.                 regexp = re.compile('tam_bot')
  91.                 if regexp.search(en('summary')[0](text = True)[0])== None:#tam_botの発言は無視
  92.                     regexp1 = re.compile('@')
  93.                     if regexp1.search(tmp_answer)==None and tmp_answer!='' and search_text.rstrip()!=tmp_answer.rstrip().encode('utf-8'):
  94.                         answer = tmp_answer
  95.                         #print >>sys.stderr,'tmp_answer=%s'%tmp_answer
  96.                         break
  97.             post = '@%s %s'%(from_user,answer)
  98.  
  99.             #########最後までわからない場合##############
  100.             if answer == '':
  101.                 post = u'@%s %s'%(from_user,unicode(search_text,'utf-8'))
  102.  
  103.     #print >>sys.stderr,'post=%s'%post
  104.     Twitter(post)        
  105.     return render_to_response('outanswer.html',{'search_text':search_text,'post':post})
  106.  
  107.  
  108. #質問文章を形態素解析かけて単語を取り出す
  109. def GetKeitaiso_mecapi(question):
  110.     question_word=''
  111.     word_rank=dict()
  112.     i=0
  113.     q={'format':'xml','sentence':question,'response':'','filter':'uniq'}
  114.     url='http://mimitako.net/api/mecapi.cgi?'+ urllib.urlencode(q)
  115.     xml_question=urlfetch.fetch(url).content
  116.     #print >>sys.stderr,BeautifulSoup(xml_question)('mecabresult')
  117.     ngword=[u'何',u'誰']
  118.     for mlist in BeautifulSoup(xml_question)('mecabresult'):
  119.         for m in mlist('word'):
  120.             term=m('feature')[0].contents[0].split(',')
  121.             word=m('surface')[0]
  122.             word=tr(u'!"#$%&()=-^~\|[]{}@:`*;+,.<>/\?_。、・', u'                                   ',word)
  123.             word=string.replace(word, ' ', '')
  124.             #print >>sys.stderr,'%s %s %s %s %s %s %s'%(term[0],term[1],term[2],term[3],term[4],term[5],word)
  125.             if (term[0]==u'名詞' or term[0]==u'動詞' or term[0]==u'形容詞') and word != '' and word not in ngword:
  126.                 cntwd = int(countWord(word))
  127.                 #print >>sys.stderr,'%s %s %s'%(term[0],word,cntwd)
  128.                 if i<3 and cntwd<100000000:
  129.                     question_word = u'%s　%s'%(question_word,word)
  130.                     print >>sys.stderr,'%s %s %s'%(term[0],word,cntwd)
  131.                     i=i+1
  132.     return question_word
  133.  
  134. #Twitterへのpostメソッド
  135. def Twitter(post):
  136.     #basic auth
  137.     username='username'
  138.     password='password'
  139.     url = 'http://twitter.com/statuses/update.json'
  140.     payload = urllib.urlencode({"status": post.encode("utf-8")})
  141.     base64string =b64encode("%s:%s" % (username, password))
  142.     headers = {"Authorization": "Basic %s" % base64string}
  143.     json_result = urlfetch.fetch(url, payload=payload, method=urlfetch.POST, headers=headers).content
  144.     return 1
  145.  
  146.  
  147. #Twitterからあるユーザの直近の質問をとってくる
  148. def GetresponseTwitter():
  149.     #basic auth
  150.     username='username'
  151.     password='password'
  152.  
  153.     url = 'http://twitter.com/statuses/replies.xml'
  154.     base64string =b64encode("%s:%s" % (username, password))
  155.     headers = {"Authorization": "Basic %s" % base64string}
  156.     xml_result=urlfetch.fetch(url, payload='', method=urlfetch.POST, headers=headers).content
  157.     result_comment=htmlentity2unicode(BeautifulSoup(xml_result)('text')[0].contents[0].split(' ')[1])
  158.     from_user=BeautifulSoup(xml_result)('screen_name')[0].contents[0]
  159.     return result_comment,from_user
  160.  
  161.  
  162. #googleの検索結果数を見る
  163. def countWord(word):
  164.     q = {#'q':'\"'+question+'\"',
  165.         'q':word.encode('utf-8'),
  166.         'v' : 1.0,
  167.         'hl' : 'ja',
  168.         'rsz': 'small'
  169.     }
  170.     url=u'http://ajax.googleapis.com/ajax/services/search/web?'+ urllib.urlencode(q)
  171.     json_result=urlfetch.fetch(url).content
  172.     json_result = string.replace(json_result, 'null', '"null"')
  173.     try:
  174.         dict_result=eval('dict('+json_result+')')        
  175.         return dict_result['responseData']['cursor']['estimatedResultCount']
  176.     except:
  177.         return 0
  178.  
  179.  
  180. def tr(pattern, repl, string):
  181.     m = dict(zip(pattern, repl))
  182.     return ''.join((m.get(c,c) for c in string))
  183.  
  184.  
  185.  
  186. #数値文字参照→文字へ
  187. def htmlentity2unicode(text):
  188.     # 正規表現のコンパイル
  189.     reference_regex = re.compile(u'&(#x?[0-9a-f]+|[a-z]+);', re.IGNORECASE)
  190.     num16_regex = re.compile(u'#x\d+', re.IGNORECASE)
  191.     num10_regex = re.compile(u'#\d+', re.IGNORECASE)
  192.  
  193.     result = u''
  194.     i = 0
  195.     while True:
  196.         # 実体参照 or 文字参照を見つける
  197.         match = reference_regex.search(text, i)
  198.         if match is None:
  199.             result += text[i:]
  200.             break
  201.  
  202.         result += text[i:match.start()]
  203.         i = match.end()
  204.         name = match.group(1)
  205.  
  206.         # 実体参照
  207.         if name in htmlentitydefs.name2codepoint.keys():
  208.             result += unichr(htmlentitydefs.name2codepoint[name])
  209.         # 文字参照
  210.         elif num16_regex.match(name):
  211.             # 16進数
  212.             result += unichr(int(u'0'+name[1:], 16))
  213.         elif num10_regex.match(name):
  214.             # 10進数
  215.             result += unichr(int(name[1:]))
  216.  
  217.     return result

URL: http://www.team-lab.com/news/index.php?itemid=469

Report this snippet


Comments

RSS Icon Subscribe to comments

You need to login to post a comment.