import numpy as np

from flask import Flask

from flask_restful import Api, Resource, reqparse, abort

app = Flask(__name__)

api = Api(app)

#DB SIZE LIMITATIONs

LIMITATION = 10000

PORT = 5000

#POST arguments

person_post_args = reqparse.RequestParser()

person_post_args.add_argument("name", type=str, required=True, help="name of the person")

person_post_args.add_argument("features", action='append', required=True, help="person features vector")

#GET arguments

person_get_args = reqparse.RequestParser()

person_get_args.add_argument("features", action='append', required=True, help="person features vector")

persons_data = {}

def get_norm(vec):

norm = np.linalg.norm(vec)

return vec / norm

def get_closest(db, input_person):

#print(f"DB BEFORE REMOVE: {db}")

res = []

# tmp_db = {key:val for key, val in db.items() if val != input_person}

#print(f"DB AFTER REMOVE: {tmp_db}")

for name, feature in db.items():

feature = np.array(feature).astype(np.float)

if np.array_equal(feature,input_person):

continue

# print('KEY', name)

# print('VAL == INPUT' , feature, input_person)

try:

#normalize the vectors

norm_vec = list(map(get_norm, [feature, input_person]))

res.append({"name" : name, "distance": np.dot(norm_vec[0], norm_vec[1])})

# print(f'cheching dot product between seatch: {input_person} and db feature {feature}')

# res.append({"name": name, "distance": np.dot(input_person, feature)})

# print(f'results is : {np.dot(input_person, feature)}')

except Exception as e:

print(e)

abort(500, message = "[ERROR] calculating closest person")

top_3 = sorted(res, key=lambda d: d['distance'], reverse=True)[:3]

# print(f"top 3 are: {top_3}")

# print(db[name])

# print([(d['name'], d['distance']) for d in top_3])

return [(d['name'], db[name]) for d in top_3]

def is_not_in_db(name):

if name not in persons_data:

abort(404, message = f"name: {name} was not found in DB")

def person_exist(name):

if name in persons_data:

abort(409, message = f"Features for 'name': {name} already exists in DB")

def limitation():

if len(persons_data) > LIMITATION:

return True

class Person(Resource):

def get(self, name):

args = person_get_args.parse_args()

nparray = np.array(args['features']).astype(np.float)

res = get_closest(persons_data, nparray)

return {'SUCESS' : res}

def post(self, name):

try:

args = person_post_args.parse_args()

#check 10K limitation

if limitation():

return 'DB IS FULL', 403

name = args["name"]

# check if person already exist in DB

person_exist(name)

features = args["features"]

persons_data[name] = features

return 'SUCESS', 201

except:

return 'FAIL', 400

def delete(self, name):

is_not_in_db(name)

del persons_data[name]

# Deleted succesfully

return '', 204

api.add_resource(Person, "/person/<string:name>")

if __name__ == "__main__":

app.run(debug=True)

#app.run(debug=True, host="0.0.0.0")

# app.run(host="0.0.0.0", port=PORT)