break
print("Cascade : ", cascade)
# Author dict creation
if cascade:
print("LastPublisher dict creation...")
else:
print("Author dict creation...")
Author = util.get_authors(data_path)
# GET LAMBDAS MUS AND GRAPH
print("Getting lambdas and mus...")
Rtweet, Rrtweet, total_time = util.get_activity(data_path, cascade, Author, divide_by_time=True, retweeted=False)
print("Getting leaders and followers...")
LeadGraph, FollowGraph = util.graph_from_trace(data_path, cascade, Author)
# list of users
Lusers = list(Rtweet.keys())
Lusers.sort()
N = len(Lusers)
# ## 3. Performance evaluation
# From the Linear System solution, one realises that it is necessary to first populate the matrices $A$ and $C$, which are relevant for any solution process of the system.
# **Note** We will keep in memory Dictionaries, with Key the userid and value the list of positive matrix entries.
# ### Build matrix A in sparse format
for line in open(psi_star): # oursin
line = line.split()
user, psi = int(line[0]), float(line[1])
if user in Psi['real']:
Psi['star'][user] = psi
del user, psi
# load authors
print("Getting authors...")
Author = util.get_authors(trace_path)
# load activity
print("Getting activity...")
Lambda, Mu, total_time = util.get_activity(trace_path,
False,
Author,
divide_by_time=True,
retweeted=False)
del Mu, total_time
Lambda = {u: Lambda[u] for u in Psi['real']}
# load star graph
print("Getting follow graph (star)...")
FollowGraph = dict()
_, FollowGraph['star'] = util.graph_from_trace(trace_path, False, Author)
del _, Author
FollowGraph['star'] = {
u: {v
for v in FollowGraph['star'][u] if v in Psi['real']}
for u in Psi['real']
}