def get_news_content(): arrList = [] try: connect = get_connect() cursor = connect.cursor() print("connection") sql = "SELECT content FROM toutiao_news" cursor.execute(sql) result = cursor.fetchall() for row in result: news = str(row[0]) dr = re.compile(r'<[^>]+>', re.S) article_content = dr.sub('', news) print(article_content) tr4s = TextRank4Sentence() tr4s.analyze(text=article_content, lower=True, source='all_filters') print('摘要:') for item in tr4s.get_key_sentences(num=1): print(item.index, item.weight, item.sentence) print('*' * 10) tr4w = TextRank4Keyword() tr4w.analyze(text=article_content, lower=True, window=2) print('关键字:') key_word = [] for item in tr4w.get_keywords(10, word_min_len=1): print(item.word, item.weight) key_word.append(item.word) arrList.append(" ".join(key_word)) except Exception as e: print(e) finally: return arrList
def update_content(news): try: connect = get_connect() cursor = connect.cursor() sql = "UPDATE news_api SET content = %s WHERE id = %s" cursor.execute(sql, (news.content, news.id)) connect.commit() except Exception as e: print(e)
def update_content(self,news): try: connect = get_connect() cursor = connect.cursor() sql = "UPDATE toutiao_news SET content = %s WHERE id = %s " cursor.execute(sql, (news.content, news.id)) connect.commit() except Exception as e: e
def select_url(self): try: connect = get_connect() cursor = connect.cursor() print("connection") sql = "SELECT id,source_url FROM toutiao_news where id >(select min(t.id) from (select id from toutiao_news where content is null order by id) t) order by id desc " cursor.execute(sql) result = cursor.fetchall() for row in result: # print(row[0]) news = News() news.id = row[0] news.source_url = row[1] self.news_list.append(news) except Exception as e: print(e)
def get_news_content(): arrList = [] try: connect = get_connect() cursor = connect.cursor() print("connection") sql = "SELECT content FROM toutiao_news" cursor.execute(sql) result = cursor.fetchall() for row in result: news = str(row[0]) dr = re.compile(r'<[^>]+>', re.S) article_content = dr.sub('', news) arrList.append(article_content) except Exception as e: print(e) finally: return arrList
def get_news_content(): tags = [] value = [] try: connect = get_connect() cursor = connect.cursor() print("connection") sql = "select * from (SELECT tag,count(1) count FROM toutiao_news group by tag) c where c.count>42 and tag!='news'" cursor.execute(sql) result = cursor.fetchall() for row in result: tags.append(row[0]) value.append(row[1]) except Exception as e: print(e) finally: return tags, value
def select_url(): arrList = [] try: connect = get_connect() cursor = connect.cursor() print("connection") sql = "SELECT id,source_url FROM news_api" cursor.execute(sql) result = cursor.fetchall() for row in result: # print(row[0]) news = News() news.id = row[0] news.source_url = row[1] arrList.append(news) except Exception as e: print(e) finally: return arrList