def find_matched_words_from_yahoo_ads():
query = request.form['query']
#yahooスポンサードサーチは単語ごとに区切るより一文にしたほうが広告出やすい
head = 'http://search.yahoo.co.jp/search/ss?p='
tail = '&ei=UTF-8&fr=top_ga1_sa&type=websearch&x=drt'
url = head + query + tail
y_ad_page = WebPage(url)
y_ad_page.fetch_html()
y_ad_page.fetch_ads()
naradeha_results = []
bracket_words = []
for ad in y_ad_page.ads:
ad.fetch_link_title()
naradeha_results.extend(ad.pick_characteristic_words())
bracket_words.extend(ad.pick_bracket_words())
# naradeharesults => [{'なら': {'before': ['。', 'あの', '今石洋之']}}]
# bracket_words => ['アスコルビン酸', 'メルトダウン']
stop_words = ['公式', '楽天', '当日', 'お急ぎ便', 'ココ', 'ここ', 'これ', 'コレ', 'こちら', '公式', '購入', '人気', '詳細', '送料無料', '配送無料', '価格', '激安', '無料', 'アマゾン', 'ヤフオク', '０', '１', '２', '３']
for num in range(0, 10):
stop_words.append(str(num))
results = naradeha_words_to_results(naradeha_results, stop_words)
for bracket_word in bracket_words:
is_including_stop_word = False
for stop_word in stop_words:
if stop_word in bracket_word:
is_including_stop_word = True
break
if is_including_stop_word:
continue
results.append(bracket_word)
return render_template('words.tmpl', words=results)
def yahoo_sponsored_results():
query = request.form['query']
#yahooスポンサードサーチは単語ごとに区切るより一文にしたほうが広告出やすい
head = 'http://search.yahoo.co.jp/search/ss?p='
tail = '&ei=UTF-8&fr=top_ga1_sa&type=websearch&x=drt'
url = head + query + tail
y_ad_page = WebPage(url)
y_ad_page.fetch_html()
y_ad_page.fetch_ads()
result_words = []
key_phrases_of_ads = []
Engine = SearchEngine()
for ad in y_ad_page.ads:
result_words.extend(ad.pick_nouns_and_verbs(ad.title))
result_words.extend(ad.pick_nouns_and_verbs(ad.snippet))
#key_phrases_of_ads.append(Engine.yahoo_key_phrase(ad.title))
#key_phrases_of_ads.append(Engine.yahoo_key_phrase(ad.snippet))
results = to_ranked_items(result_words)
#return ad_template.render(items=results)
return render_template('find_words_with_yahoo_ads.tmpl',
items=results)
def yahoo_sponsored_results():
query = request.forms.decode().get('query')
#yahooスポンサードサーチは単語ごとに区切るより一文にしたほうが広告出やすい
head = 'http://search.yahoo.co.jp/search/ss?p='
tail = '&ei=UTF-8&fr=top_ga1_sa&type=websearch&x=drt'
url = head + query + tail
y_ad_page = WebPage(url)
ads = y_ad_page.fetch_ads()
v_and_s = []
for ad in ads:
v_and_s.extend(ad.pick_verbs(ad.title))
v_and_s.extend(ad.pick_sahens(ad.title))
v_and_s.extend(ad.pick_verbs(ad.snippet))
v_and_s.extend(ad.pick_sahens(ad.snippet))
results = to_ranked_items(v_and_s)
return ad_template.render(items=results)
def yahoo_sponsored_results():
query = request.forms.decode().get('query')
#yahooスポンサードサーチは単語ごとに区切るより一文にしたほうが広告出やすい
head = 'http://search.yahoo.co.jp/search/ss?p='
tail = '&ei=UTF-8&fr=top_ga1_sa&type=websearch&x=drt'
url = head + query + tail
y_ad_page = WebPage(url)
ads = y_ad_page.fetch_ads()
v_and_s = []
for ad in ads:
v_and_s.extend(ad.pick_verbs(ad.title))
v_and_s.extend(ad.pick_sahens(ad.title))
v_and_s.extend(ad.pick_verbs(ad.snippet))
v_and_s.extend(ad.pick_sahens(ad.snippet))
results = to_ranked_items(v_and_s)
return ad_template.render(items=results)
def query_expansion():
query = request.forms.decode().get('query')
#yahooスポンサードサーチは単語ごとに区切るより一文にしたほうが広告出やすい
head = 'http://search.yahoo.co.jp/search/ss?p='
tail = '&ei=UTF-8&fr=top_ga1_sa&type=websearch&x=drt'
url = head + query + tail
y_ad_page = WebPage(url)
ads = y_ad_page.fetch_ads()
v_and_s = []
for ad in ads:
v_and_s.extend(ad.pick_verbs(ad.title))
v_and_s.extend(ad.pick_sahens(ad.title))
v_and_s.extend(ad.pick_verbs(ad.snippet))
v_and_s.extend(ad.pick_sahens(ad.snippet))
ranked_items = to_ranked_items(v_and_s)
ranked_items.insert(0, {'name': 'まとめ', 'count': 100})
normalized_query = normalize_query(query)
query_words = normalized_query.split(' ')
page_set = set() # set型は重複をなくすため
expanded_queries = []
for item in ranked_items:
top_5 = []
if item['name'] in query_words:
#'花粉症 対策'で検索したら'対策'がitem['name']に入っていたりする
continue
else:
expanded_query = normalized_query + ' ' + item['name']
expanded_queries.append(expanded_query)
new_pages = search(expanded_query, 1)
top_5 = new_pages[0:4]
over_19 = False
for one in top_5:
page_set.add(one)
if len(page_set) > 19:
over_19 = True
break
if over_19 is True:
break # => ranked_items内で回すのから脱出
return expand_template.render(pages=page_set, queries=expanded_queries)
def query_expansion():
query = request.forms.decode().get('query')
#yahooスポンサードサーチは単語ごとに区切るより一文にしたほうが広告出やすい
head = 'http://search.yahoo.co.jp/search/ss?p='
tail = '&ei=UTF-8&fr=top_ga1_sa&type=websearch&x=drt'
url = head + query + tail
y_ad_page = WebPage(url)
ads = y_ad_page.fetch_ads()
v_and_s = []
for ad in ads:
v_and_s.extend(ad.pick_verbs(ad.title))
v_and_s.extend(ad.pick_sahens(ad.title))
v_and_s.extend(ad.pick_verbs(ad.snippet))
v_and_s.extend(ad.pick_sahens(ad.snippet))
ranked_items = to_ranked_items(v_and_s)
ranked_items.insert(0, {'name': 'まとめ', 'count': 100})
normalized_query = normalize_query(query)
query_words = normalized_query.split(' ')
page_set = set() # set型は重複をなくすため
expanded_queries = []
for item in ranked_items:
top_5 = []
if item['name'] in query_words:
#'花粉症 対策'で検索したら'対策'がitem['name']に入っていたりする
continue
else:
expanded_query = normalized_query + ' ' + item['name']
expanded_queries.append(expanded_query)
new_pages = search(expanded_query, 1)
top_5 = new_pages[0:4]
over_19 = False
for one in top_5:
page_set.add(one)
if len(page_set) > 19:
over_19 = True
break
if over_19 is True:
break # => ranked_items内で回すのから脱出
return expand_template.render(pages=page_set, queries=expanded_queries)
