text 将训练集中的SEQ归为从1开始的序号序列

Posted 2021-04-30

item_dict = {}
item_ctr = 1
train_seqs = []
train_dates = []
for s ,date in train_sess:
	seq = sess_clicks[s]
	outseq = []
	for i in seq:
		if item_dict.has_key(i):
			outseq += [item_dict[i]]
		else:
			outseq += [item_ctr]
			item_dict[i] = item_ctr
			item_ctr += 1
	if len(outseq) < 2:
		continue
	train_seqs += [outseq]
	train_dates += [date]