def gen_user_buy_with_coupon(matrix_offline, matrix_online, X):
user_buy = HashSet(default=matrix_offline.default)
user_buy_with_coupon = HashSet(default=matrix_offline.default)
for i in xrange(matrix_offline.ndata):
cid_str = matrix_offline.get_cell(i, 'cid')
date_str = matrix_offline.get_cell(i, 'date')
uid_str = matrix_offline.get_cell(i, 'uid')
if date_str != 'null':
user_buy.add_one(uid_str)
if cid_str != 'null':
user_buy_with_coupon.add_one(uid_str)
for i in xrange(matrix_online.ndata):
act_str = matrix_online.get_cell(i, 'act')
cid_str = matrix_online.get_cell(i, 'cid')
uid_str = matrix_online.get_cell(i, 'uid')
if act_str == '1':
user_buy.add_one(uid_str)
if cid_str != 'null':
user_buy_with_coupon.add_one(uid_str)
user_buy_with_coupon_freq = user_buy.merge_op(
user_buy_with_coupon, lambda x, y: float(y) * 1.0 / float(x), dft=0.0)
X.join("uid",
["user_buy", "user_buy_with_coupon", "user_buy_with_coupon_freq"],
user_buy.merge(user_buy_with_coupon,
dft=0.0).merge(user_buy_with_coupon_freq,
dft=0.0), ("%s" for i in xrange(3)),
dft=0.0)
def gen_user_get_coupon(offline_source, online_source, X, month):
offline = Matrix(
np.genfromtxt(paths.ccf_path + offline_source,
delimiter=',',
dtype=str),
["uid", "mid", "cid", "dis_rate", "dist", "date_rec", "date"],
["%s" for i in xrange(7)])
online = Matrix(
np.genfromtxt(paths.ccf_path + online_source, delimiter=',',
dtype=str),
["uid", "mid", "act", "cid", "dis_rate", "date_rec", "date"],
["%s" for i in xrange(7)])
user_get_coupon = HashSet()
for i in xrange(offline.ndata):
uid_str = offline.get_cell(i, "uid")
cid_str = offline.get_cell(i, "cid")
if cid_str != 'null':
user_get_coupon.add_one(uid_str)
for i in xrange(online.ndata):
uid_str = online.get_cell(i, "uid")
cid_str = online.get_cell(i, "cid")
act_str = online.get_cell(i, "act")
if cid_str != 'null':
assert act_str != '0'
user_get_coupon.add_one(uid_str)
X.join("uid", ["user_get_coupon"], user_get_coupon, ["%s"], dft=0.0)
def divide(x, y):
assert float(x) != -9999 and float(y) != -9999
assert float(x) <= float(y), "{0} {1}".format(x, y)
if float(y) == 0:
return 0.0
else:
return float(x) * 1.0 / float(y)
X.gen_arith_feature("user_buy_with_coupon",
"user_get_coupon",
"user_use_coupon_freq",
divide,
"%s",
dft=0.0)
def gen_merchant_share(X, month):
offline = Matrix(
np.genfromtxt(paths.ccf_path + 'offline_train_{0}.csv'.format(month),
delimiter=',',
dtype=str),
["uid", "mid", "cid", "dis_rate", "dist", "date_rec", "date"],
["%s" for i in xrange(7)])
online = Matrix(
np.genfromtxt(paths.ccf_path + 'online_train_{0}.csv'.format(month),
delimiter=',',
dtype=str),
["uid", "mid", "act", "cid", "dis_rate", "date_rec", "date"],
["%s" for i in xrange(7)])
merchant_user_buy = HashSet()
merchant_user_use_coupon = HashSet()
merchant_user_buy_counter = HashSet()
merchant_user_use_coupon_counter = HashSet()
for i in xrange(offline.ndata):
if i % 100000 == 0:
print i
mid_str = offline.get_cell(i, "mid")
uid_str = offline.get_cell(i, "uid")
date_str = offline.get_cell(i, "date")
cid_str = offline.get_cell(i, "cid")
if date_str != 'null':
if not merchant_user_buy.has(mid_str):
merchant_user_buy.set(mid_str, HashSet())
if not merchant_user_buy.get(mid_str).has(uid_str):
merchant_user_buy.get(mid_str).add_one(uid_str)
merchant_user_buy_counter.add_one(mid_str)
if cid_str != 'null':
if not merchant_user_use_coupon.has(mid_str):
merchant_user_use_coupon.set(mid_str, HashSet())
if not merchant_user_use_coupon.get(mid_str).has(uid_str):
merchant_user_use_coupon.get(mid_str).add_one(uid_str)
merchant_user_use_coupon_counter.add_one(mid_str)
for i in xrange(online.ndata):
if i % 100000 == 0:
print i
mid_str = online.get_cell(i, "mid")
uid_str = online.get_cell(i, "uid")
act_str = online.get_cell(i, "act")
cid_str = online.get_cell(i, "cid")
if act_str == '1':
if not merchant_user_buy.has(mid_str):
merchant_user_buy.set(mid_str, HashSet())
if not merchant_user_buy.get(mid_str).has(uid_str):
merchant_user_buy.get(mid_str).add_one(uid_str)
merchant_user_buy_counter.add_one(mid_str)
if cid_str != 'null':
if not merchant_user_use_coupon.has(mid_str):
merchant_user_use_coupon.set(mid_str, HashSet())
if not merchant_user_use_coupon.get(mid_str).has(uid_str):
merchant_user_use_coupon.get(mid_str).add_one(uid_str)
merchant_user_use_coupon_counter.add_one(mid_str)
X.join("mid", ["merchant_user_buy"],
merchant_user_buy_counter, ["%s"],
dft=0.0)
X.join("mid", ["merchant_user_use_coupon"],
merchant_user_use_coupon_counter, ["%s"],
dft=0.0)
X.check_point(month)
def gen_merchant_buy(X, month):
offline = Matrix(
np.genfromtxt(paths.ccf_path + 'offline_train_{0}.csv'.format(month),
delimiter=',',
dtype=str),
["uid", "mid", "cid", "dis_rate", "dist", "date_rec", "date"],
["%s" for i in xrange(7)])
online = Matrix(
np.genfromtxt(paths.ccf_path + 'online_train_{0}.csv'.format(month),
delimiter=',',
dtype=str),
["uid", "mid", "act", "cid", "dis_rate", "date_rec", "date"],
["%s" for i in xrange(7)])
merchant_buy = HashSet()
merchant_buy_with_coupon = HashSet()
merchant_distribute_coupon = HashSet()
for i in xrange(offline.ndata):
if i % 100000 == 0:
print i
mid_str = offline.get_cell(i, "mid")
date_str = offline.get_cell(i, "date")
cid_str = offline.get_cell(i, "cid")
if date_str != 'null':
merchant_buy.add_one(mid_str)
if cid_str != 'null':
merchant_buy_with_coupon.add_one(mid_str)
if cid_str != 'null':
merchant_distribute_coupon.add_one(mid_str)
for i in xrange(online.ndata):
if i % 100000 == 0:
print i
mid_str = online.get_cell(i, "mid")
act_str = online.get_cell(i, "act")
cid_str = online.get_cell(i, "cid")
if act_str == '1':
merchant_buy.add_one(mid_str)
if cid_str != 'null':
merchant_buy_with_coupon.add_one(mid_str)
if cid_str != 'null':
assert act_str != '0'
merchant_distribute_coupon.add_one(mid_str)
X.join("mid", ["merchant_buy"], merchant_buy, ["%s"], 0.0)
X.join("mid", ["merchant_buy_with_coupon"], merchant_buy_with_coupon,
["%s"], 0.0)
X.join("mid", ["merchant_distribute_coupon"], merchant_distribute_coupon,
["%s"], 0.0)
def divide(x, y):
assert float(x) <= float(y)
if float(y) == 0:
return 0.0
else:
return float(x) * 1.0 / float(y)
X.gen_arith_feature("merchant_buy_with_coupon",
"merchant_buy",
"merchant_buy_with_coupon_ratio",
divide,
"%s",
dft=0.0)
X.gen_arith_feature("merchant_buy_with_coupon",
"merchant_distribute_coupon",
"merchant_coupon_ratio",
divide,
"%s",
dft=0.0)
X.check_point(month)