j = Judge()
# f = open("user_notes_cleaned.txt")
f = open("../../data/user_notes_cleaned.txt")
counter = 1
print "sys userno topicid qno noteno noteid docno prec"
for s in f:
s = s.strip()
system, userno, topicid, q, note_no = s.split()
## remove qustion mark
if note_no[-1] == "?" or note_no[-1] == "!":
note_no = note_no[:-1]
if note_no == "":
continue
print system, userno, topicid, q, note_no,
nid, docid, note = get_note(system, userno, topicid, note_no)
note = note.replace("\r\n", " ")
note = note.lower()
prec = j.check_annotation(topicid, docid, note)
print nid, docid, prec[1]
sys.stderr.write("\rLines processed: %d" % counter)
sys.stderr.flush()
counter += 1
j = Judge()
f = open("master_notes.txt", "a")
con = MySQLdb.connect('localhost', 'root', '!!berkeley', 'yournews-tdt4')
cur = con.cursor()
n = cur.execute("select nid, userid, docid, note from notes where userid = 'vst' order by nid desc" )
counter = 1
res = []
for row in cur:
nid, userid, docid, note = row
note = note.replace("\r\n", " ")
note = note.lower()
precision = j.check_annotation('40004', docid, note)
try:
sys.stdout.write("%s %s %s %s %s %s\n" % (nid, system, user, topic, docid, precision[1])) #, note
except:
sys.stdout.write("%s %s %s %s %s error\n" % (nid, system, user, topic, docid))
# sys.stderr.write("\r%d/%d" % (counter, n))
# sys.stderr.flush()
counter += 1
res.append(float(precision[1]))
print sum(res)/len(res)
