def get_person_data(self):
while True:
response = requests.request(
"GET", base_url + odm_people + "name=" + self.first_name +
" " + self.last_name + "&user_key=" + user_key)
if response.status_code != 401: # Workaround: sometimes '401 - Unauthorized user_key'
break
response_json = json.loads(response.text)
person_list = response_json["data"]["items"]
for person in person_list:
person = person["properties"]
f_name = HumanName(
person["first_name"] + " " + person["last_name"]
).first # CB stores titles in the first_name field *sigh*
if (self.first_name == f_name) & (self.last_name
== person["last_name"]):
if (utils.sim(self.organization, person["organization_name"],
2 / 3)
| utils.sim(self.organization, person["title"],
2 / 3)): # title = job
return person
return dict()
def __call__(self, x, x_src, cg_edge_index, batch_src):
"""
No edge_attr and no u on this one.
Arguments :
- x (node feature tensor, size [X, f_x]) : node features of the
current graph
- x_src (node feature tensor, size [X, f_x]) : node features of the
other graph
- cg_edge_index (edge index tensor, size [2, E]) : cross-graph edge
index tensor, mapping nodes of the other graph (source) to the
current graph (dest)
- batch (batch tensor, size [X]) : batch tensor mapping the nodes
of the other graph to their respective graph in the batch.
Returns :
- attentions (size [X], node size of the other graph.)
"""
src, dest = cg_edge_index
# exp-cosine-similarity vector
ecs = torch.exp(sim(x_src[src], x[dest]))
# attentions
a = ecs / (scatter_add(ecs, batch_src[src])[batch_src[src]])
vec = x[dest] - x_src[src] # to be multiplied by attentions
mu = (a * vec.T)
return scatter_add(mu,
dest).T # see if we don't need mean instead here
def create_f(a, b):
f = np.zeros((len(a), len(b)))
for j in xrange(len(b)):
f[0, j] = d * j
for i in xrange(len(a)):
f[i, 0] = d * i
for i in xrange(1, len(a)):
for j in xrange(1, len(b)):
match = f[i-1, j-1] + sim(a[i], b[j])
delete = f[i-1, j] + d
insert = f[i, j-1] + d
f[i, j] = max(match, delete, insert)
return f
def get_contact_by_name(first_name, last_name, organization_name):
''' Get contact_ids matching first name '''
rows = get(
"contacts_values", "contact_id",
"field_name = '{}' AND value = '{}'".format('first_name', first_name))
if not rows:
return tuple([None] * 4) # Contact not in database
first_name_contact_id_list = list(map(lambda x: x[0], rows))
''' Get contact_ids matching last name '''
rows = get(
"contacts_values", "contact_id",
"field_name = '{}' AND value = '{}'".format('last_name', last_name))
if not rows:
return tuple([None] * 4)
last_name_contact_id_list = list(map(lambda x: x[0], rows))
name_contact_id_list = set(first_name_contact_id_list).intersection(
last_name_contact_id_list)
''' Match the remaining contacts' organization field to get a contact-ID'''
contact_id = None
if len(name_contact_id_list) == 1:
contact_id = list(name_contact_id_list)[0]
for tmp_contact_id in name_contact_id_list:
rows = get(
"contacts_values", "value",
"contact_id = '{}' AND field_name = '{}'".format(
tmp_contact_id, 'orga_name'))
if rows:
if utils.sim(rows[0][0], organization_name, 2 / 3):
contact_id = tmp_contact_id
break
if not contact_id:
return tuple([None] * 4)
# retrieve data from contacts table
rows = get("contacts", "last_sync, sync_interval",
"contact_id = '{}'".format(contact_id))
if not rows:
return tuple([None] * 4)
contact_last_sync = rows[0][0]
contact_sync_interval = rows[0][1]
# retrieve all fields in contacts_values tables
contact_data = get_contact_values("contacts_values",
"contact_id = '{}'".format(contact_id))
return contact_id, contact_data, contact_last_sync, contact_sync_interval
def align(a, b):
f = create_f(a, b)
al_a = ''
al_b = ''
i = len(a) - 1
j = len(b) - 1
while i > 0 or j > 0:
sim_val = sim(a[i], b[j])
if i > 0 and j > 0 and f[i, j] == f[i-1, j-1] + sim_val:
al_a = a[i] + al_a
al_b = b[j] + al_b
i -= 1
j -= 1
elif i > 0 and f[i, j] == f[i-1, j] + d:
al_a = a[i] + al_a
al_b = GAP + al_b
i -= 1
else:
al_a = GAP + al_a
al_b = b[j] + al_b
j -= 1
return al_a, al_b
