def write_graph(G, user_names, media_prov, is_conservative):
res = {'nodes': [], 'links': []}
media = list(MEDIA.keys())[::-1]
for n in G.nodes():
user_name = user_names[n]
if media_prov[user_name] == '':
media_index = 0
else:
media_index = media.index(media_prov[user_name]) + 1
if user_name.lower() in NEOCONSERVATIVES:
neoconservative = True
else:
neoconservative = False
res['nodes'].append({
'id': user_name,
'media': media_prov[user_name],
'media_index': media_index,
'neoconservative': neoconservative,
'conservative': is_conservative[user_name]
})
for e in G.edges():
res['links'].append({
'source': user_names[e[0]],
'target': user_names[e[1]],
'value': G[e[0]][e[1]]['weight']
})
writej(res, 'website/outgroup.json')
async def get_user_by_id(user_id, client):
file_path = PREFIX + user_id + '_user'
obj = loadj(file_path)
if obj is not None:
return obj
res = await client.api.users.lookup.get(user_id=user_id)
writej(res[0], file_path)
return res[0]
async def get_friends(user_id, screen_name, client):
friends_ids = client.api.friends.ids.get.iterator.with_cursor(
user_id=user_id, count=3000)
friends = []
try:
async for data in friends_ids:
friends.extend([str(f) for f in data.ids])
file_path = PREFIX + str(user_id) + '_' + screen_name + '_friends'
writej(friends, file_path, overwrite=False)
except Exception as e:
print(e)
return friends
def gen_keywords():
if os.path.exists(KEYWORDS_PATH):
keywords = load_keywords()
else:
keywords = {}
# We're keeping a cache of already parsed links and also
# are storing their matched canonical link to ensure that we can
# add a user to the set of an already parsed link.
if os.path.exists(PARSED_LINKS_PATH):
parsed_links = loadj(PARSED_LINKS_PATH)
else:
parsed_links = {}
graph_names = [n['id'] for n in loadj('website/outgroup.json')['nodes']]
user_ids = [USER_IDS[name] for name in graph_names]
user_links = get_user_links(user_ids)
all_links = []
for user_name, links in user_links.items():
for l in links:
# Don't process the link if we already did.
if l not in parsed_links:
all_links.append((user_name, l))
parsed_links[l] = ''
elif parsed_links[l] != '' and parsed_links[l] in keywords:
keywords[parsed_links[l]]['users'].add(user_name)
print(f'{len(all_links)} to parse...')
p = Pool(NUM_WORKERS)
kw_tuples = p.starmap(get_keywords_pmap, all_links)
for user_name, c_link, l, kws, p_time, title in kw_tuples:
if c_link is not None:
parsed_links[l] = c_link
keywords[c_link]['kws'] = kws
keywords[c_link]['time'] = p_time
if 'users' not in keywords[c_link]:
keywords[c_link]['users'] = set()
keywords[c_link]['users'].add(user_name)
keywords[c_link]['title'] = title
else:
parsed_links[l] = ''
# Make the keywords dict serializable.
for c_link in keywords:
keywords[c_link]['users'] = list(keywords[c_link]['users'])
keywords[c_link]['kws'] = list(keywords[c_link]['kws'])
writej(keywords, KEYWORDS_PATH)
writej(parsed_links, PARSED_LINKS_PATH)
def write_user_names():
user_names = {}
files = [f for f in os.listdir(PREFIX) if f.find('_user')>0]
for f_name in files:
try:
obj = loadj(PREFIX+f_name)
user_names[obj['id_str']] = obj['screen_name']
except Exception as e:
print(f_name, e)
writej(user_names, PREFIX + 'user_names')
# Write the reverse dict as well.
user_ids = {}
for user_id, name in user_names.items():
user_ids[name] = user_id
writej(user_ids, PREFIX + 'user_ids')
async def write_followers(user_id, client):
file_path = PREFIX + user_id + '_followers'
if os.path.exists(file_path):
return
print(f'Getting followers for {user_id}...')
followers_ids = client.api.followers.ids.get.iterator.with_cursor(
id=user_id, count=MAX_SEED_FOLLOWER_COUNT)
try:
followers = []
async for data in followers_ids:
followers.extend(data.ids)
writej(followers, file_path, overwrite=False)
except Exception as e:
print(e)
print(f'Done getting followers for {user_id}.')
return
async def get_user_by_name(screen_name, client):
res = await client.api.users.lookup.get(screen_name=screen_name)
user = res[0]
file_path = PREFIX + user['id_str'] + '_user'
writej(user, file_path, overwrite=False)
return user
def flush_cache(app):
writej(app['link_cache'], PREFIX + 'link_cache.json')
