def get_social_encounters(social_path, encounters_path, dest_path, save_file=True):
print 'loading social graph:', social_path
social_graph = utils.load_pickle(social_path)
print 'loading encounters graph:', encounters_path
encs_graph = utils.load_pickle(encounters_path)
encs_edges_set = set(encs_graph.edges())
df_dict = {constants.USER_1: [], constants.USER_2: [], constants.ENCS_COUNT: [], constants.SOC_DIST: []}
print 'iterating through', len(encs_edges_set), 'edges.'
disconnected_edges = 0
count = 0
hund = len(encs_edges_set)/100
for source, dest in encs_edges_set:
count += 1
if count % hund == 0:
print float(count)/(hund*100)
encs_count = len(encs_graph[source][dest])
df_dict[constants.USER_1].append(source)
df_dict[constants.USER_2].append(dest)
df_dict[constants.ENCS_COUNT].append(encs_count)
try:
distance = nx.shortest_path_length(social_graph, source, dest)
df_dict[constants.SOC_DIST].append(distance)
except Exception as e:
disconnected_edges += 1
df_dict[constants.SOC_DIST].append(NO_PATH)
pass
print 'edges without a social connection:', disconnected_edges
dataframe = pd.DataFrame(df_dict)
if save_file:
dataframe.to_csv(dest_path, index=False)
return dataframe
def get_social_encounters(social_path_0, social_path_1, social_path_2, encounters_path, dest_path, save_file=True):
print 'loading social graph:', social_path_0
social_graph_0 = utils.load_pickle(social_path_0)
print 'loading social graph:', social_path_1
social_graph_1 = utils.load_pickle(social_path_1)
print 'loading social graph:', social_path_2
social_graph_2 = utils.load_pickle(social_path_2)
print 'loading encounters graph:', encounters_path
encs_graph = utils.load_pickle(encounters_path)
encs_edges_set = set(encs_graph.edges())
df_dict = {constants.USER_1: [], constants.USER_2: [], constants.ENCS_COUNT: [], constants.MODE_0_DIST: [],
constants.MODE_1_DIST: [], constants.MODE_2_DIST: []}
print 'iterating through', len(encs_edges_set), 'edges.'
disconnected_edges = 0
count = 0
hund = len(encs_edges_set)/100
for source, dest in encs_edges_set:
count += 1
if count % hund == 0:
print float(count)/(hund*100)
encs_count = len(encs_graph[source][dest])
df_dict[constants.USER_1].append(source)
df_dict[constants.USER_2].append(dest)
df_dict[constants.ENCS_COUNT].append(encs_count)
if (source in social_graph_0) and (dest in social_graph_0) and (nx.has_path(social_graph_0, source, dest)):
distance_0 = nx.shortest_path_length(social_graph_0, source, dest)
df_dict[constants.MODE_0_DIST].append(distance_0)
else:
df_dict[constants.MODE_0_DIST].append(NO_PATH)
if (source in social_graph_1) and (dest in social_graph_1) and (nx.has_path(social_graph_1, source, dest)):
distance_1 = nx.shortest_path_length(social_graph_1, source, dest)
df_dict[constants.MODE_1_DIST].append(distance_1)
else:
df_dict[constants.MODE_1_DIST].append(NO_PATH)
if (source in social_graph_2) and (dest in social_graph_2) and (nx.has_path(social_graph_2, source, dest)):
distance_2 = nx.shortest_path_length(social_graph_2, source, dest)
df_dict[constants.MODE_2_DIST].append(distance_2)
else:
df_dict[constants.MODE_2_DIST].append(NO_PATH)
print 'edges without a social connection:', disconnected_edges
dataframe = pd.DataFrame(df_dict)
if save_file:
dataframe.to_csv(dest_path, index=False)
return dataframe
def combine_maps_for_months(data_dir='/home/niquo/niquo_data',
months_paths=USE_MONTHS):
print 'combing interactin graphs for data from:', data_dir
TOWER_ENCS = 'tower_encounters'
master_graph_filename = 'MASTER_GRAPH.p'
month_master_paths = []
for dir_str in months_paths:
month_path = os.path.join(data_dir, dir_str)
print 'combing for month path:', month_path
inter_map_obj = imap.InteractionMap(month_path)
tower_data_path = os.path.join(month_path, TOWER_ENCS)
month_master = os.path.join(tower_data_path, master_graph_filename)
month_master_paths.append(month_master)
inter_map_obj.combine_all_graphs(tower_data_path)
general_imap = imap.InteractionMap(data_dir)
print 'made a general imap object for:', data_dir
dest_filename = os.path.join(data_dir, master_graph_filename)
for m_mast in month_master_paths:
print 'combing map from:', m_mast
month_graph = utils.load_pickle(m_mast)
general_imap.combine_maps(month_graph)
print 'storing data:', dest_filename
general_imap.store_data(dest_filename)
print 'complete.'
return True
def create_encs_df_select_friends(first_call_csv,
root_path,
dest_filename=constants.PAIRS_CSV):
encs_path = os.path.join(root_path, constants.ENCS_DICT)
print 'loading all encounter pairs from:', encs_path
encs_dict = utils.load_pickle(encs_path)
friend_df = split_users_first_call_csv(first_call_csv)
friends_set = set([(user1, user2) for user1, user2 in friend_df[
[constants.USER_1, constants.USER_2]].values])
intersection_pairs = friends_set.intersection(set(encs_dict.keys()))
relevant_encs = {k: encs_dict[k] for k in intersection_pairs}
mode_0_path = os.path.join(root_path, constants.MODE_0_GRAPH)
print 'loading graph mode 0:', mode_0_path
mode_0_graph = utils.load_pickle(mode_0_path).to_undirected()
mode_1_path = os.path.join(root_path, constants.MODE_1_GRAPH)
print 'loading graph mode 1:', mode_1_path
mode_1_graph = utils.load_pickle(mode_1_path).to_undirected()
mode_2_path = os.path.join(root_path, constants.MODE_2_GRAPH)
print 'loading graph mode 2:', mode_2_path
mode_2_graph = utils.load_pickle(mode_2_path).to_undirected()
friend_df[constants.ENCS_COUNT] = friend_df.apply(
lambda row: apply_encs(relevant_encs, row), axis=1)
friend_df[constants.MODE_0_DIST] = friend_df.apply(
lambda row: apply_distance(mode_0_graph, row), axis=1)
friend_df[constants.MODE_1_DIST] = friend_df.apply(
lambda row: apply_distance(mode_1_graph, row), axis=1)
friend_df[constants.MODE_2_DIST] = friend_df.apply(
lambda row: apply_distance(mode_2_graph, row), axis=1)
friend_df = friend_df[friend_df[constants.ENCS_COUNT] >= 0]
dest_path = os.path.join(root_path, dest_filename)
print 'storing dataframe at ', dest_path
friend_df.to_csv(dest_path, index=False)
return friend_df
def get_encounters_for_pairs(root_path, dest_path):
encs_dict = {}
root = root_path.split('/')[-1]
encs_path = os.path.join(root_path, TOWER_ENCS_DIR)
print '**************************'
print 'entering root path:', root
for day_dir in os.listdir(encs_path):
day_path = os.path.join(encs_path, day_dir)
print '________________'
print 'opening day path:', day_path
for tower_file in os.listdir(day_path):
print 'opening tower file:', tower_file
tower_path = os.path.join(day_path, tower_file)
tower_graph = utils.load_pickle(tower_path)
tower_edges = set(tower_graph.edges())
for user_1, user_2 in tower_edges:
count = len(tower_graph[user_1][user_2])
add_encs_count(encs_dict, user_1, user_2, count)
del tower_graph
with open(dest_path, 'wb') as outfile:
cPickle.dump(encs_dict, outfile)
return encs_dict