def read_specific_question_from_post(id):
con = mdb.connect('localhost', 'root', '123456', 'answerbot')
cur = con.cursor()
sql = "SELECT * FROM posts WHERE Id=" + str(id)
try:
cur.execute(sql)
results = cur.fetchall()
for row in results:
# id,title,body,tag
q_tmp = SO_Que()
except Exception as e:
print e
cur.close()
con.close()
return q_tmp
def read_questions_from_repo(num):
sql = 'SELECT * FROM answerbot.repo WHERE PostTypeId=1 limit 0,' + str(num)
SO_datalist = []
con = mdb.connect('localhost', 'root', '123456', 'answerbot')
cur = con.cursor()
try:
cur.execute(sql)
results = cur.fetchall()
count = 0
for row in results:
count += 1
# id,type,title,title_NO_SW,title_NO_SW_Stem,tag
q_tmp = SO_Que(row[0], row[1], row[2], row[3], row[4], row[5])
SO_datalist.append(q_tmp)
except Exception as e:
print e
cur.close()
con.close()
return SO_datalist
def read_all_questions_from_repo():
sql = 'SELECT * FROM repo_qs'
SO_datalist = []
con = mdb.connect('localhost', 'root', '123456', 'answerbot')
cur = con.cursor()
try:
cur.execute(sql)
results = cur.fetchall()
count = 0
for row in results:
count += 1
# id,title,body,tag
q_tmp = SO_Que(row[0], row[1], row[2], row[3])
SO_datalist.append(q_tmp)
except Exception as e:
print e
cur.close()
con.close()
return SO_datalist
def read_all_questions_from_post():
sql = 'SELECT * FROM posts where Tags like \'%<java>%\' and AnswerCount > 0'
SO_datalist = []
con = mdb.connect('localhost', 'root', '123456', 'answerbot')
cur = con.cursor()
try:
cur.execute(sql)
results = cur.fetchall()
count = 0
for row in results:
count += 1
# id,type,title,title_NO_SW,title_NO_SW_Stem,tag
q_tmp = SO_Que(row[0], row[1], row[2], row[3], row[4], row[5])
SO_datalist.append(q_tmp)
except Exception as e:
print e
cur.close()
con.close()
return SO_datalist
def read_questions_from_java():
sql = 'SELECT * FROM answerbot.java_qs;'
con = mdb.connect('localhost', 'root', '123456', 'answerbot')
cur = con.cursor()
qlist = []
cur.execute(sql)
results = cur.fetchall()
count = 0
for row in results:
count += 1
# id,title,body,tag
q_tmp = SO_Que(row[0], row[11], row[6], row[12])
q_tmp = preprocessing_for_que(q_tmp)
qlist.append(q_tmp)
if len(qlist) % 10000 == 0:
print 'Load %s questions...' % len(qlist)
cur.close()
con.close()
return qlist
def read_q_list_from_java(id_list):
qlist = []
con = mdb.connect('localhost', 'root', '123456', 'answerbot')
cur = con.cursor()
count = 1
for qid in id_list:
sql = "SELECT * FROM java_qs WHERE Id = %s" % qid[0]
try:
cur.execute(sql)
results = cur.fetchall()
row = results[0]
# id,title,body,tags
qlist.append(SO_Que(row[0], row[11], row[6], row[12]))
count += 1
if count % 1000 == 0:
print 'reading ' + str(count) + ' question from Table java_qs'
except Exception as e:
print e
cur.close()
con.close()
return qlist