def initailizeUserDictByURL(url, count=5):
user_dict = {}
for urlposts in pydelicious.get_urlposts(url=url)[1 : count+1]:
user = urlposts['user']
user_dict[user] = {}
return user_dict
def initializeUserDict(tag,count=5):
user_dict = {}
all_items = {}
# get the top count' popular posts
for p1 in get_popular(tag=tag)[0:count]:
# find all users who posted this
for p2 in get_urlposts(p1['href']):
user = p2['user']
user_dict.setdefault(user,{})
# get user posts
for i in range(3):
try:
posts = get_userposts(user)
break
except:
print "Failed user "+user+", retrying"
time.sleep(4)
for post in posts:
url = post['href']
user_dict[user][url] = 1.0
all_items[url] = 1
#fill in missing items with 0
[ratings.setdefault(item,0.0) for item in all_items for ratings in user_dict.values()]
return user_dict
def initializeUserDict(tag, count = 5):
usr_dict = {}
for most_popular in get_popular(tag = tag)[0 : count]:
for person in get_urlposts(most_popular['href']):
usr = person['user']
usr_dict[usr] = {}
return usr_dict
def initializeUserDict(tag, count=5):
userDict = {}
for p1 in get_popular(tag=tag)[0:count]:
for p2 in get_urlposts(p1['url']):
user = p2['user']
userDict[user] = {}
return userDict
def initialize(tag,count=5):
user_dict = {}
for p1 in get_popular(tag=tag)[0:count]:
for p2 in get_urlposts(p1['url']):
user = p2['user']
user_dict[user] = {}
return user_dict
def initializeUserDict(tag, count=5):
user_dict = {}
for p1 in tag[0:count]:
for p2 in get_urlposts(p1):
user = p2["user"]
user_dict[user] = {}
return user_dict
def initializeUserDict(tag, count=5):
user_dict = {}
for p1 in get_popular(tag=tag)[0:count]:
for p2 in get_urlposts(p1["url"]):
user = p2["user"]
user_dict[user] = {}
return user_dict
def initializeUserDict(tag, count=5):
user_dict={}
# 获取前count个最受欢迎的张
for p1 in get_popular(tag = tag)[0:count]:
for p2 in get_urlposts(p1['href']):
user=p2['user']
user_dict[user] = {}
return user_dict
def initializeUserDict(tag ,count = 5):
user_dict = {}
#get the top count popular posts
for p1 in get_popular(tag = tag)[0:count]:
for p2 in get_urlposts(p1['url']):
user = p2['user']
user_dict[user] = {}
return user_dict
def initUserDict(tag, count=5):
user_dict = {}
for p1 in py.get_popular(tag)[0:count]:
for p2 in py.get_urlposts(p1['href']):
user = p2['user']
user_dict[user] = {}
return user_dict
def initialze_user_dict(tag, count=5):
user_dict = {}
# 获取前count个最受欢迎的链接张贴记录
for p1 in get_popular(tag=tag)[0:count]:
for p2 in get_urlposts(p1['url']): # 此处与书中不同，现在返回字典的'href' 键已经改为 'url'
user = p2['user']
user_dict[user] = {}
return user_dict
def initlizeUserDict(tag, count=5):
user_dict = {}
#获取所有用户都提交过的链接
for p1 in get_popular(tag=tag)[0:count]:
for p2 in get_urlposts(p1['href']):
user = p2['user']
user_dict[user] = {}
return user_dict
def initialize_user_dict(tag,count=5):
user_dict={}
# Get the top "count" popular posts
for p1 in pydelicious.get_popular(tag=tag)[0:count]:
# Find all users who posted the same post url
for p2 in pydelicious.get_urlposts(p1['url']):
user=p2['user']
user_dict[user]={}
return user_dict
def initializeUserDict(tag,count=5):
user_dict={}
# count개의 인기 게시물을 얻음
for p1 in get_popular(tag=tag)[0:count]:
# 이 게시물을 올린 모든 사용자를 얻음
for p2 in get_urlposts(p1['href']):
user=p2['user']
user_dict[user]={}
return user_dict
def initializeUserDict(tag, count=5):
user_dict = {}
# отримати рахунок найпопулярніших лінків
for p1 in get_popular(tag=tag)[0:count]:
# знайти всіх користувачів, які зберегли цей лінк
for p2 in get_urlposts(p1['href']):
user = p2['user']
user_dict[user] = {}
return user_dict
def initializeUserDict(tag,count=5):
user_dict={}
# get the top count' popular posts
for p1 in get_popular(tag=tag)[0:count]:
# find all users who posted this
for p2 in get_urlposts(p1['url']):
user=p2['user']
user_dict[user]={}
return user_dict
def initializeUserDict(tag,count=5):
user_dict={}
#获取前count个最受欢迎的链接张贴记录
for p1 in get_popular(tag=tag)[0:count]:
#查看所有张贴该链接的用户
for p2 in get_urlposts(p1['herf']):
user=p2['user']
user_dict[user]={}
return user_dict
def initializeUserDict(tag,count=5):
user_dict={}
#get the top "count" popular posts
for p1 in get_popular(tag=tag)[0:count]: #from api
#find all users who posted this
for p2 in get_urlposts(p1['href']):
user=p2['user']
user_dict[user]={} #looking at a key:value pair?
return user_dict
def initializeUserDict(tag, count=30): # 引数に tag と count をとる、count の指定がなければ 5
user_dict = {}
# popular な投稿を 0 から count 番目まで取得
for p1 in get_popular(tag=tag)[0:count]:
# このリンクを投稿したすべてのユーザを取得
for p2 in get_urlposts(p1['href']):
user = p2['user']
user_dict[user] = {}
return user_dict
def initlizeUserDict(tag, count = 5):
user_dict = {}
#获取所有用户都提交过的链接
for p1 in get_popular(tag=tag)[0:count]:
for p2 in get_urlposts(p1['href']):
user = p2['user']
user_dict[user]={}
return user_dict
def initializeUserDict(tag, count=5):
user_dict={}
for p1 in get_popular(tag=tag)[0:count]:
# get_urlposts()는 주어진 url에 대한 모든 게시글을 리턴
for p2 in get_urlposts(p1['url']):
# 리턴된 게시글들의 '유저이름'을 가져옴. '유저이름'의 value는 일단 null로 둠
user=p2['user']
user_dict[user]={}
return user_dict
def initializeUserDict(tag,count=2):
user_dict={}
# get the top count' popular posts
for p1 in get_popular(tag=tag)[0:count]:
# find all users who posted this
for p2 in get_urlposts(p1['href']):
user=p2['user']
user_dict[user]={}
return user_dict
def initializeUserDict(tag, count=5):
user_dict={}
# отримати рахунок найпопулярніших лінків
for p1 in get_popular(tag=tag)[0:count]:
# знайти всіх користувачів, які зберегли цей лінк
for p2 in get_urlposts(p1['href']):
user=p2['user']
user_dict[user]={}
return user_dict
def initialize_user_dict(tag, count=5):
user_dict = {}
# get the top count' posts
for p1 in get_tagposts(tag=tag)[0:count]:
# find all users who posted this
for p2 in get_urlposts(p1['url']):
user = p2['user']
user_dict[user] = {}
return user_dict
def initializeUserDict(tag, count=5):
user_dict = {}
#获取当前count个最受欢迎的链接张贴记录
for p1 in get_popular(tag=tag)[0:count]:
#查找所有张贴该链接的用户
for p2 in get_urlposts(p1['href']):
user = p2['user']
user_dict[user] = {}
return user_dict
def initializeUserDict(tag, count=5):
user_dict = {}
# count개의 인기 게시물을 얻음
for p1 in get_popular(tag=tag)[0:count]:
# 이 게시물을 올린 모든 사용자를 얻음
for p2 in get_urlposts(p1['href']):
user = p2['user']
user_dict[user] = {}
return user_dict
def initializeUserDict(tag, count=5):
user_dict = {}
# 获取某tag下前count个最受欢迎的url post记录
for p1 in get_popular(tag=tag)[0:count]:
# 获取所有post该url的用户
for p2 in get_urlposts(p1['href']):
user = p2['user']
user_dict[user] = {}
return user_dict
def initializeUserDict(tag, count=5):
user_dict = {}
# get top "count" of popular posts:
for p1 in get_popular(tag=tag)[0:count]:
# Find the users who poseted them:
for p2 in get_urlposts(p1['href']):
user = p2['user']
user_dict[user] = {}
return user_dict
def initializeUserDict(tag, count = 5):
user_dict = {}
# read in top "count" popular posts
for p1 in get_popular(tag = tag)[0:count]:
#find all users who posted this
for p2 in get_urlposts(p1['url']):
user = p2['user']
user_dict[user] = {}
return user_dict
def initializeUserDict(tag,count = 5):
user_dict = {}
#获取前count个最受欢迎的链接张贴记录
for p1 in get_popular(tag=tag)[0:count]:
#查找所有张贴记录的用户
for p2 in get_urlposts(p1['url']):
user = p2['user']
user_dict[user] = {} #得到一个包含用户的空字典
return user_dict
def initializeUserDict(tag, count=5):
user_dict = {}
for p1 in get_popular(tag=tag)[0:count]:
#find all users that posted this
for p2 in get_urlposts(p1['href']):
user = p2['user']
user_dict[user] = {}
return user_dict
def initializeUserDict(tag, count=5):
user_dict={}
#get count most popular links
for p1 in get_popular(tag=tag)[0:count]:
#find all users saved this link
for p2 in get_urlposts(p1['url']):
user = p2['user']
user_dict[user] = {}
return user_dict
def initializeUserDict(tag, count=5):
"""Grabs the user ids of active del.icio.us users."""
user_dict = {}
# get top `count` popular posts...
for p1 in get_popular(tag=tag)[0:count]:
# ...and add all users who posted them
for p2 in get_urlposts(p1['href']):
user = p2['user']
user_dict[user] = {}
return user_dict
def _init_user_dict(self, count=5):
user_dict = {}
for p1 in pydelicious.get_popular(tag=self._tag)[0:count]:
for p2 in pydelicious.get_urlposts(p1['url']):
user = p2['user']
user_dict[user] = {}
if __debug__: print(user_dict)
self._usr_dict = user_dict
def initialize_user_dict(tag, count=5):
"""
Returns: length n dict of users: empty dict pairs.
"""
user_dict = {}
for post in get_popular(tag=tag)[:count]:
for url in get_urlposts(post['url']):
user = url['user']
user_dict[user] = {}
return user_dict
def _init_user_dict(self, count=5):
user_dict = {}
for p1 in pydelicious.get_popular(tag=self._tag)[0:count]:
for p2 in pydelicious.get_urlposts(p1['url']):
user = p2['user']
user_dict[user] = {}
if __debug__: print(user_dict)
self._usr_dict = user_dict
def initialiseUserDict(tag,count=5):
user_dict={}
for p1 in pydelicious.get_popular(tag=tag):
print p1
for p2 in pydelicious.get_urlposts(p1['url'])[0:count]:
print p2
user = p2['user']
if len(user):
user_dict[user]={}
return user_dict
def initializeUserDictByTag(tag, count=5):
user_dict = {}
# 获取前count个最受欢迎的链接张贴记录
for p1 in pydelicious.get_popular(tag=tag)[0:count] :
# 查找所有张贴该链接的用户
for p2 in pydelicious.get_urlposts(p1['url']) :
user = p2['user']
user_dict[user] = {}
return user_dict
def initializeUserDict(tag, count = 5):
'''Get a list of users who recently posted a popular link.'''
user_dict = {}
# get the top count popular posts
for p1 in get_popular(tag = tag)[0:count]:
# find all users who posted this
for p2 in get_urlposts(p1['href']):
user = p2['user']
user_dict.setdefault(user, {})
return user_dict
def initializeUserDict(tag, count=5):
"""Grabs the user ids of active del.icio.us users."""
user_dict = {}
# get top `count` popular posts...
for p1 in get_popular(tag=tag)[0:count]:
# ...and add all users who posted them
for p2 in get_urlposts(p1['href']):
user = p2['user']
user_dict[user] = {}
return user_dict
def initUserDict(tag, count = 5):
user_dict = {}
# get the top count popular posts
for p1 in pydelicious.get_popular(tag=tag)[0:count]:
# find all users who posted this
# print pydelicious.get_urlposts(p1['url'])
for p2 in pydelicious.get_urlposts(p1['url']):
user = p2['user']
user_dict[user] = {}
return user_dict
def initUserDict(tag, n):
user_dict={}
#Fåtop count' populära posts
for p1 in get_popular(tag)[0:n]:
#Hitta alla användare som postade den
for p2 in get_urlposts(p1['href']):
user=p2['user']
user_dict[user]={}
return user_dict
def initializeUserDict(tag,count=5):
user_dict={}
# get popular blog posts
for p1 in get_popular(tag=tag)[0:count]:
for p2 in get_urlposts(p1['href']):
user=p2['user']
if user:
user_dict[user]={}
return user_dict
def initializeUserDict(tag, count=5):
user_dict={}
# popularな投稿をcount番目まで取得
for p1 in get_popular(tag=tag)[0:count]:
# このリンクを投稿したすべてのユーザを取得
for p2 in get_urlposts(p1['href']):
user = p2['user']
user_dict[user]={}
return user_dict
def initializeUserDict(tag, count=5):
suer_dict = {}
a = get_popular(tag=tag)[0:count]
print a
for p1 in a:
# Find all user who posted it
for p2 in get_urlposts(p1['href']):
user = p2['user']
user_dict[user] = {}
return user_dict
def initialize_object_dict(self, tag, count=5):
"""
Initialize object (user) dict with users from
the popular posts for tag
"""
for popular in get_popular(tag=tag)[0:count]:
for posts in get_urlposts(popular['url']):
object = posts['user']
if self.verbose: print object
self.object_dict[object] = dict()
def initializeUserDict(tag, count=5):
user_dict = {}
# get the top count' popular posts
for p1 in get_popular(tag=tag)[0:count]:
for p2 in get_urlposts(p1['url']):
user = p2['user']
#print(p1['url'] + ' ' + user)
user_dict[user] = {}
return user_dict
def initializeUserDict(tag, count=5):
user_dict = {}
for p1 in get_popular(tag=tag)[0:count]:
print(p1)
if (not p1["href"]):
break
for p2 in get_urlposts(p1["href"]):
user = p2['user']
user_dict[user] = {}
return user_dict
def initializeUserDict(tag='programming', count=1):
user_dict = {}
# get the top count' popular posts
for p1 in get_popular(tag=tag)[0:count]:
# find all users who posted this
print "working on count", count
pst = get_urlposts(p1['url'])
print "pst len", len(pst)
for p2 in pst:
user = p2['user']
user_dict[user] = {}
print "user=" + user
return user_dict
def initializeUserDict(tag, count=5):
user_dict = {}
try:
# get the top count popular posts
for p1 in get_popular(tag=tag)[0:count]:
# Find all users who posted this
for p2 in get_urlposts(p1['href']):
user = p2['user']
user_dict[user] = {}
#
except Exception:
print('function')
return user_dict
def getItemsTags(user_dict,n=5):
urltag_dict = {}
all_tags = {}
count = 0
url_limit_per_user = 5
break_flag = 0
for user in user_dict:
if(count==n):
break_flag = 1
break
count += 1
url_count = 0
for url in user_dict[user]:
#print "Url["+url+"]"
if(url_count>url_limit_per_user):
break;
url_count += 1
retry_count = 0
for i in range(3):
try:
posts=get_urlposts(url)
break
except:
print "Failed url "+url+", retrying"
time.sleep(4)
retry_count +=1
if(retry_count>=3):
continue
for post in posts[0:10]:
#print post['tags']
tag_count = 0
for tag in post['tags'].split():
#print "working with tag %s" % tag
all_tags[tag]=1
urltag_dict.setdefault(url,{})
urltag_dict[url][tag]=1.0
tag_count+=1
#print "\\ttags %d" %tag_count
#break
print "Collecting tags finished"
# Fill in missing tags with 0
for tags_ratings in urltag_dict.values():
for tag in all_tags:
if tag not in tags_ratings:
tags_ratings[tag]=0.0
return urltag_dict
def initializeUserDict(tag, count=5):
user_dict = {}
#获取前count个最受欢迎的张贴记录
for p1 in get_popular(tag=tag)[0:count]:
print "Got a post !"
#查找所有张贴该记录的用户
#原来的API已经改了，不用href，而是url
for i in range(3):
try:
for p2 in get_urlposts(p1['url']):
user = p2['user']
user_dict[user] = {}
break
except:
print "Failed....retrying"
time.sleep(4)
return user_dict
def fillItem(user_dict):
all_item = {}
for user in user_dict:
for i in range(3):
try:
posts = get_urlposts(user)
break
except:
print "Failed user " + user + ", retrying"
time.sleep(4)
for post in posts:
url = post['href']
user_dict[user][url] = 1.0
all_item[url] = 1
for ratings in user_dict.values:
for item in all_item:
if item not in ratings:
ratings[item] = 0.0
def fillItems(user_dict):
all_items = {}
#查找所有用户都提交过的链接
for user in user_dict:
for i in range(3):
try:
posts = get_urlposts(user)
break
except:
print "failed user " + user + ", retrying"
time.sleep(4)
for post in posts:
url = post['href']
user_dict[user][url] = 1.0
all_items[url] = 1
#用0填充缺失项
for ratings in user_dict.values():
for item in all_items:
if item not in ratings: ratings[item] = 0.0
def fill_items(user_dict):
all_items = {}
# 查找所有用户都提交过的链接
for user in user_dict:
for i in range(3):
try:
posts = get_urlposts(user)
break
except:
print 'Failed user' + user + ', retrying'
time.sleep(4)
for post in posts:
url = post['url']
user_dict[user][url] = 1
all_items[url] = 1
# 用0填充缺失的内容
for ratings in user_dict.values():
for item in all_items:
if item not in ratings:
ratings[item] = 0
def fillitems(user_dict):
all_items = {}
#Find links posted by all users
for user in user_dict:
for i in range(3):
try:
posts = get_urlposts(user)
break
except:
print "Failed user " + user + ", retrying"
time.sleep(4)
for post in posts:
url = post['url']
user_dict[user][url] = 1.0
all_items[url] = 1
#Fill in missing items with 0
for ratings in user_dict.values():
for item in all_items:
if item not in ratings:
ratings[item] = 0.0
def fillItems(user_dict):
all_items = {}
# Finding links posted by the users in user_dict:
for user in user_dict:
# Make only three attempts to get data:
for i in range(3):
try:
posts = get_urlposts(user)
break
except:
print "Failed user" + user + ", trying again"
time.sleep(4)
for post in posts:
url = post['fref']
user_dict[user][url] = 1.0
all_items[url] = 1
# Fill in missing items with 0:
for ratings in user_dict.values():
for item in all_items:
if item not in ratings:
ratings[item] = 0.0
def initUserDict(tag):
user_dict = {}
popular_links = [
#u'http://coursera.org/',
u'http://ibm.com/',
]
for url in popular_links:
# The get_urlposts API returns last 30 people who post the link
while (True):
try:
posts = get_urlposts(url)
print 'fetch {0} users from {1}.'.format(len(posts), url)
break
except:
print 'Error on {0}, retrying'.format(url)
time.sleep(4)
for post in posts:
user = post['user']
if user.strip() != '':
user_dict[user] = {}
print url
return user_dict
def initializeUserDict(tag, count=5, cache_dict=True):
#get_popular doesn't work any more. Getting by tag
#return [url_post['user'] for item in get_tagposts(tag=tag)[0:count] for url_post in get_urlposts(item['url'])]
#from hashlib import md5
#print md5('http://usingtechnologybetter.com/chrome-extensions-for-teachers/').hexdigest()
user_dict_cache_file = "user_dict.p"
if cache_dict and exists(user_dict_cache_file):
print "Reading the user_dict from cache file", user_dict_cache_file
cache_file = open(user_dict_cache_file, "rb")
user_dict = load(cache_file)
cache_file.close()
else:
user_dict = dict([(url_info['user'], {})
for item in get_tagposts(tag=tag)[0:count]
for url_info in get_urlposts(item['url'])
if url_info['user'] != ''])
if cache_dict:
print "Writing the user_dict to cache file", user_dict_cache_file
cache_file = open(user_dict_cache_file, "wb")
dump(user_dict, cache_file)
cache_file.close()
return user_dict
