def check_request(request):
"""
Validates that our request is well formatted
Returns:
- assertion value: True if request is ok, False otherwise
- error message: empty if request is ok, False otherwise
"""
if "observation_id" not in request.keys():
error = "Field 'observation_id' missing from request: {}".format(request)
return False, error
elif 'data' not in request.keys():
error = "Field `data` missing from request: {}".format(request)
return False, error
else:
return True, ""
def Predict():
request = extract_message()
logger.debug("Request: %s", request)
features_dict = {}
for k in request.keys():
features_dict[k] = get_data_from_json(request[k])
predictions = user_model.predict(features_dict)
logger.debug("Predictions: %s", predictions)
# If predictions is an numpy array or we used the default data then return as numpy array
predictions = np.array(predictions)
if len(predictions.shape) > 1:
class_names = get_class_names(user_model, predictions.shape[1])
else:
class_names = []
data = array_to_rest_datadef(predictions, class_names,
request.get("data", {}))
response = {"data": data, "meta": {}}
tags = get_custom_tags(user_model)
if tags:
response["meta"]["tags"] = tags
metrics = get_custom_metrics(user_model)
if metrics:
response["meta"]["metrics"] = metrics
return jsonify(response)
def validate_required_inputs(request):
required = ['text', 'model_name']
missing = missing = [
field for field in required if field not in request.keys()
]
if missing:
err = {field: f"the {field} field is required" for field in missing}
return err, 422
def populate_list(bucket):
for entry in bucket["by_top_hit"]["hits"]["hits"]:
count = len(request.keys())
for key in request.keys():
if (key == "report_date" and request[key] != "ALL KEYS"):
try:
days = compare_two_dates(entry['_source'][key][0:10],
request[key])
except:
try:
t_index = entry['_source'][key].index('T')
report_date = entry['_source'][key][0:t_index]
days = compare_two_dates(report_date, request[key])
except:
report_date = entry['_source']['time'][0:10]
days = compare_two_dates(report_date, request[key])
if (days >= 0 and days <= interval):
count -= 1
else:
break
elif (request[key] == "ALL KEYS"):
count -= 1
elif (entry['_source'].get(key) == None):
break
elif (entry['_source'][key].lower() == request[key].lower()):
count -= 1
elif (re.match(request[key].lower(),
entry['_source'][key].lower()) != None):
count -= 1
if (count == 0 and visited.get(entry["_source"]["sr_id"]) == None
and entry["_index"] == "version_string_sda"):
visited[entry["_source"]["sr_id"]] = True
lst.append(entry["_source"])
def parse_signed_request(request):
if not isinstance(request, dict) or \
set(request.keys()) != {'message', 'signature'} or \
not isinstance(request['message'], dict) or \
not isinstance(request['signature'], str):
return None, 'Unexpected formatting'
signature = request['signature']
message_dict = request['message']
message_str = json.dumps(message_dict,
separators=(',', ':'),
sort_keys=True).encode('ascii').decode()
response = host.call('verifymessage', secAuthorizerAddress, signature,
message_str)
if not response:
return None, 'Invalid signature'
return message_dict, ''
def convert_input(request):
"""
convert input given through form into a model readable form. can be done through calling preprocess after converting to something like a dataframe
additional arguements for further customization?
"""
keys = [str(k) for k in request.keys()]
print(request)
feature_names = [k.split('-')[0] for k in keys]
print(feature_names)
feature_order = [int(k.split('-')[1]) for k in keys]
feature_values = [request[keys[i]] for i in range(len(feature_names))]
print(feature_order)
print(feature_values)
print('------------------------------------------------------------------------------------------------------')
print(request)
df = pd.DataFrame(columns=feature_names)
df.loc[0] = feature_values
print(df.head())
query = 'select file_name from preprocess where model_id = ?'
cur = conn.execute(query,(str(keys[0]).split('-')[2],))
file_name = cur.fetchone()
spec = importlib.util.spec_from_file_location("module.name", 'preprocess/' + file_name[0])
preprocess = importlib.util.module_from_spec(spec)
spec.loader.exec_module(preprocess)
# preprocess = importlib.import_module('preprocess/' + file_name[0] + ".py")
df = preprocess.preprocess(df)
# df.to_csv("test.csv")
return prediction(str(keys[0]).split('-')[2],df)
def create():
"""
This function creates a new pathogen_model in the pathogen_models structure
based on the passed in pathogen_model data
:return: 201 on success, 406 on pathogen_model exists
"""
# if the item comes in as a single, let's [
if type(request.form['name']) is str:
new_model, schema = insert_one_model(form=request.form, files=request.files)
else:
for i in range(len(request.form['name'])):
new_model, schema = insert_one_model(form={k: request.form[k][i] for k in request.keys()},
files={k: request.form[k][i] for k in request.files.keys()})
db.session.commit()
# Serialize and return the newly created pathogen_model in the response
data = schema.dump(new_model).data
return data, 201