M = len(ori)
V = len(ori[0])
assert M % d == 0
assert V % d == 0
m = M / d
v = V / d
GAMMA = 0.02
LAMBDA = 0.1
STEP = 0.9
W = MutableDict(d)
H = MutableDict(d)
ori_b = dpark.broadcast(ori)
def sgd(i_j):
(i, j) = i_j
Wi = W.get(i)
if Wi is None:
Wi = numpy.random.rand(m, k)
W.put(i, Wi)
Hj = H.get(j)
if Hj is None:
Hj = numpy.random.rand(v, k)
H.put(j, Hj)
ori = ori_b.value
IBIAS_PATH='/nfs/wuhong/offline_use/ibias_0/'
RATING_PATH='/nfs/wuhong/fm_data/user_music_factor_model/user_track_rating_for_training/'
ITEM_FACTOR_PATH='/nfs/wuhong/offline_use/H_0/'
NEW_RATING_PATH='/nfs/wuhong/offline_use/rating_new/'
NEW_ITEM_FACTOR_PATH='/nfs/wuhong/offline_use/H_new/'
dpark = DparkContext()
f_global = file(MU_PATH)
line = ''
for l in f_global:
line = l
mu = float(line.strip().split('\t')[1])
f_global.close()
mu = dpark.broadcast(mu)
def local_mapper(line):
iid, v, _ = line.strip().split('\t')
return (iid, float(v))
ibias = {}
ibias = dpark.textFile(glob.glob(IBIAS_PATH)).map(
local_mapper
).collectAsMap()
ibias = dpark.broadcast(ibias)
def local_mapper2(line):
uid, iid, aid, v = line.strip().split('\t')
return '%s,%s,%s\n' % (uid , iid, float(v) - mu - ibias[iid])
.filter(lambda x:x)\
.map(lambda l: general_map.value.parse(l, spec))\
.filter(lambda x:x)\
.filter(lambda line: (not is_spider(line) and (line['uid'] or line['bid'])))\
.filter(lambda l: l['bid'] not in fraud.value and l['uid'] not in fraud.value)
spec = set(['url', 'uid', 'bid', 'unit_id', 'ad_id', 'status_code', 'user_agent', 'region', 'page_tags', 'hour', 'group'])
features = common_gen(spec)
features = features.map(feature_extract)\
.filter(lambda x:x)\
.cache()
user_list = set(features.map(lambda x: x[0]).filter(lambda x: x<>'None').collect())
user_list_b = dp.broadcast(user_list)
user_feature = dp.makeRDD([])
def _parse_list(line):
uid, features = line.split('\t')
features = [x.split(':') for x in features.split('|')]
features = [(x[0], float(x[1])) for x in features]
features = sorted(features, key=lambda x: x[1], reverse=True)
return (uid, features)
for name in ['book_cluster', 'movie_cluster', 'group_cluster', 'text_cluster']:
fn = '/home2/alg/user_profile/%s/%s' % (current_date, name)
if not os.path.exists(fn):
continue
rdd = dp.textFile(fn, splitSize=16<<20)\
