def test_nested(self): """Tests for nested Params.""" a = Params() a.define("x", 0, "") self.assertEqual(str(a), "{\n x: 0\n}") b = Params() a.define("y", b, "") b.define("z", "z", "") self.assertEqual(str(a), '{\n x: 0\n y: {\n z: "z"\n }\n}') self.assertRaises(AttributeError, lambda: a.get("w")) self.assertEqual(a.get("x"), 0) self.assertEqual(type(a.get("y")), Params) self.assertRaises(AttributeError, lambda: a.get("y.w")) self.assertEqual(a.get("y.z"), "z") self.assertRaises(AttributeError, lambda: a.get("y.z.w"))
def __init__(self, environment, general_nn=None, inherit_nn=None): self.params = Params() if self.params.seed: seed(self.params.seed) self.creatures = [] self.entity_grid = np.zeros( (self.params.grid_size, self.params.grid_size, 4)) # 4 is for has_entity, id, strength, energy self.environment = environment if self.params.general_nn: # T self.random_policy = True else: self.random_policy = False self.general_nn = general_nn self.inherit_nn = inherit_nn self.exploration_rate = self.params.exploration_rate for _ in range(self.params.starting_creatures): self.spawn_creature() if self.params.verbose: self.logs = Logs(self.environment) self.logs.log_run() self.logs.log_header() self.batch_counter = self.general_nn.align_counter # starts 1 self.logs_random = not self.params.logs_random and self.random_policy
class TestRegionMapper(unittest.TestCase): params = Params() params.retina_enabled = True params.retina_kwidth = 15 params.s1_kwidth = 11 params.s1_scaling = 2 params.c1_kwidth = 5 params.c1_scaling = 2 params.s2_kwidth = 7 params.c2_kwidth = 3 params.c2_scaling = 2 rm = RegionMapper(params) def testRetinaToImage(self): self.assertEqual(slice(0, 15), self.rm.MapRetinaToImage(slice(0, 1))) def testS1ToImage(self): self.assertEqual(slice(0, 25), self.rm.MapS1ToImage(slice(0, 1))) def testC1ToImage(self): self.assertEqual(slice(0, 33), self.rm.MapC1ToImage(slice(0, 1))) def testS2ToImage(self): self.assertEqual(slice(0, 57), self.rm.MapS2ToImage(slice(0, 1))) def testC2ToImage(self): self.assertEqual(slice(0, 73), self.rm.MapC2ToImage(slice(0, 1)))
def test_matchzoo(): params = Params() config_file = 'config/qalocal.ini' # define dataset in the config params.parse_config(config_file) params.network_type = "anmm.ANMM" reader = qa.setup(params) qdnn = models.setup(params) model = qdnn.getModel() model.compile(loss = params.loss, optimizer = units.getOptimizer(name=params.optimizer,lr=params.lr), metrics=['accuracy']) model.summary() # generators = [reader.getTrain(iterable=False) for i in range(params.epochs)] # q,a,score = reader.getPointWiseSamples() # model.fit(x = [q,a],y = score,epochs = 1,batch_size =params.batch_size) def gen(): while True: for sample in reader.getPointWiseSamples(iterable = True): yield sample model.fit_generator(gen(),epochs = 2,steps_per_epoch=1000)
def run_evaluation(model_type): params = Params('params/' + model_type + '.json') score, scores = evaluate_actor_critic(params, 'models/' + model_type + '.pt') print('Average reward after 100 episodes: {0:.2f}'.format(score)) plot_performance(scores)
def __init__(self): self.model = Model() self.params = Params() self.screen = Screen() self.fzf = FzfPrompt() self.info = ''
def test_eq_ne(self): """Tests equality methods.""" a = Params() b = Params() self.assertEqual(a, b) a.define("x", 0, "") self.assertNotEqual(a, b) b.define("x", 0, "") self.assertEqual(a, b) a.x = 1 self.assertNotEqual(a, b) a.x = 0 self.assertEqual(a, b) c = Params() c.define("y", 0, "") self.assertNotEqual(b, c)
def __init__(self, answers, APP, nodeps = False, update = False, target_path = None, dryrun = False, **kwargs): run_path = os.path.dirname(os.path.realpath(__file__)) self.dryrun = dryrun recursive = not nodeps app = APP #FIXME self.params = Params(recursive, update, target_path) self.utils = utils.Utils(self.params) if os.path.exists(app): logger.info("App path is %s, will be populated to %s" % (app, target_path)) app = self.utils.loadApp(app) else: logger.info("App name is %s, will be populated to %s" % (app, target_path)) if not target_path: if self.params.app_path: self.params.target_path = self.params.app_path else: self.params.target_path = os.getcwd() self.params.app = app self.answers_file = answers
def __init__(self, address, protocol, vehicle): self.address = address self.vehicle = vehicle self.protocol = protocol # Queue is filled with incoming messages (receive_thread) and emptied by receive worker (receive_task_thread) self.msg_queue = Queue.Queue() # Loads the initial vehicle's values according to class Params self.vehicle_params = Params(network=self, vehicle=vehicle) # For collision avoidance purposes self.drones = [] self.context = {'mode': None, 'mission': None, 'next_wp': None} self.priority = None # First entry in drones list will be our own vehicle self.drones.append(self.vehicle_params) # Normally it is kept as is until the first receive_thread # self.populate_drones('two_dummies') self.sock_send = None self.sock_receive = None # Create transceiver and worker threads self.t_send = SendThread(self, self.address) self.t_receive = ReceiveThread(self, self.msg_queue) self.t_task = ReceiveTaskThread(self, self.msg_queue) self.receive_count = self.t_receive.count self.task_count = self.t_task.count
def test_roi(self): """Upload an roi and test it's fields""" # Make a skeleton makeAnno(p, 9) # test the parent parent = random.randint(0, 65535) f = setField(p, 'parent', parent) f = getField(p, 'parent') assert parent == int(f.content) # make a bunch of children ROIs q = Params() q.token = 'unittest' q.resolution = 0 q.channels = ['unit_anno'] childids = [] for i in range(0, 4): makeAnno(q, 9) f = setField(q, 'parent', p.annoid) childids.append(q.annoid) # Test children f = getField(p, 'children') rchildids = f.content.split(',') for cid in rchildids: assert int(cid) in childids assert len(rchildids) == 4
def __init__(self): self.params = Params() if self.params.seed: seed(self.params.seed) self.action_space = [i for i in range(self.params.action_size)] self.state_size = (self.params.vision_grid, self.params.vision_grid, self.params.state_features) self.experience_replay = deque(maxlen=self.params.memory_size) self.run_counter = 1 # creature counter self.align_counter = 1 # batch retrain counter self.q_eval = DeepQNetwork() self.q_next = DeepQNetwork() # tensorboard if self.params.tensorboard: self.writer = self.board() self.cum_reward = 0 self.align_target() self.random_action = True self.loss = 0 self.agent_hash = randint(1, 10000000)
def test_node(self): """Upload a skeleton node and test it's fields""" # Make a node makeAnno(p, 7) # test the nodetype nodetype = random.randint(0, 100) f = setField(p, 'nodetype', nodetype) f = getField(p, 'nodetype') assert nodetype == int(f.content) # test the skeletonid skeletonid = random.randint(0, 65535) f = setField(p, 'skeletonid', skeletonid) f = getField(p, 'skeletonid') assert skeletonid == int(f.content) # test the pointid pointid = random.randint(0, 65535) f = setField(p, 'pointid', pointid) f = getField(p, 'pointid') assert pointid == int(f.content) # test the parentid parentid = random.randint(0, 65535) f = setField(p, 'parentid', parentid) f = getField(p, 'parentid') assert parentid == int(f.content) # test the radius radius = random.random() f = setField(p, 'radius', radius) f = getField(p, 'radius') assert abs(radius - float(f.content)) < 0.001 # test the location location = [random.random(), random.random(), random.random()] f = setField(p, 'location', ','.join([str(i) for i in location])) f = getField(p, 'location') assert ','.join([str(i) for i in location]) == f.content # make a bunch of children q = Params() q.token = 'unittest' q.resolution = 0 q.channels = ['unit_anno'] childids = [] for i in range(0, 4): makeAnno(q, 9) f = setField(q, 'parent', p.annoid) childids.append(q.annoid) # Test children f = getField(p, 'children') rchildids = f.content.split(',') for cid in rchildids: assert int(cid) in childids assert len(rchildids) == 4
def make_sorted_gt(): # only keep video snapshots that contain at least 10 face crops from each video. # the sorted gt will only include ground truth from videos that satistfies the minimal requirement. # im_list is a variable that saves a list of images, sorted by the ground truth's video sequence. # you may edit im_list to make a image list for caffe, or other frameworks. p = Params(config_path='params.cfg') cha_raw_gt = pd.read_pickle(p.cha_gt_file) new_gt = pd.DataFrame(columns=list(cha_raw_gt.columns.values)) video_list = cha_raw_gt.index.tolist() face_img_names = listdir(p.crop_face_im_dir) im_list, im_count_list = [], [] for cur_ind, cur_video in enumerate(video_list): if cur_ind % 100 == 0: print '{}-th video out of {}...'.format(cur_ind, len(video_list)) pattern = cur_video[:-4] subset = fnmatch.filter(face_img_names, pattern+'*.jpg') if len(subset) > p.lower_bound_frame_num: im_list.extend(subset) im_count_list.append(len(subset)) new_gt.loc[cha_raw_gt.index[cur_ind]] = cha_raw_gt.iloc[cur_ind] else: print 'video {} only contains {} image snapshots. Not included in the selection'.format(cur_video, len(subset)) new_gt['img_frame_num'] = pd.Series(im_count_list, index=new_gt.index) new_gt.to_pickle(p.mapped_gt_file) print 'New gt length = {}'.format(len(new_gt)) # how many videos are kept. return
def main(args): if args.debug: import pdb; pdb.set_trace() # objects radius = 5.0 manip_obj = Circle(radius, Position(5.0,radius)) finger0 = Circle(1.0, Position(-5.0, -5.0)) finger1 = Circle(1.0, Position(15.0, -5.0)) # initial contact information contact_state = OrderedDict([(finger0, Contact(f=(0.0, 0.0), ro=(-7., -7.), c=.5)), (finger1, Contact(f=(0.0, 0.0), ro=(7., -7.), c=.5))]) goals = [Position(5.0, 20.0)] world = World(manip_obj, [finger0, finger1], contact_state) stage_weights=[StageWeights(w_CI=0.1, w_physics=0.1, w_kinematics=0.0, w_task=1.0), StageWeights(w_CI=10.**1, w_physics=10.**0, w_kinematics=0., w_task=10.**1)] p = Params(world, K=10, delT=0.05, delT_phase=0.5, mass=1.0, mu=0.9, lamb=10**-3, stage_weights=stage_weights) if args.single: stage_info = CIO(goals, world, p, single=True) else: stage_info = CIO(goals, world, p) if args.save: save_run(args.save, p, world, stage_info)
def menu(): url = load_wsdl_url() while True: print("Выберите операцию:") print("1. Сложение") print("2. Вычитание") print("3. Умножение") print("4. Деление") print("0. Выход") option = int(input()) if option not in range(5): print("Неверный ввод!") continue if option == 0: break operation = '' if option == 1: operation = 'add' if option == 2: operation = 'sub' if option == 3: operation = 'mul' if option == 4: operation = 'div' print("Введите 1 операнд:") arg1 = input() print("Введите 2 операнд:") arg2 = input() params = Params(arg1, arg2, operation) print(f"Результат: {soap_request(url, params)}") print('---------------------------------')
def check_yaml(path: str) -> Params: import yaml yml = yaml.safe_load(open(path, 'r')) if (Checker.url(yml['url']) and Checker.tag_type(yml['tag_type']) and Checker.tag_identifier(yml['tag_identifier']) and Checker.tag_value(yml['tag_value']) and Checker.freq_val(yml['freq_val']) and Checker.freq_type(yml['freq_type'])): return Params(yml) else: if not Checker.url(yml['url']): print(f"URL '{yml['url']}' was not valid.") if not Checker.tag_type(yml['tag_type']): print(f"Tag type '{yml['tag_type']}' was not valid.") if not Checker.tag_identifier(yml['tag_identifier']): print( f"Tag identifier '{yml['tag_identifier']}' was not valid.") if not Checker.tag_value(yml['tag_value']): print(f"'{yml['tag_value']}' was not valid.") if not Checker.freq_val(yml['freq_val']): print(f"Frequency '{yml['freq_val']}' was not valid.") if not Checker.freq_type(yml['freq_type']): print(f"Time measure '{yml['freq_type']}' was not valid.")
def main(): logging.basicConfig(filename='load_balancer.log', level=logging.INFO) logging.info('Started') #This takes input and validates it params = Params() #This is how you access the required params print params.vip_interface print params.vip_port print params.vip_ip print params.target_ip #do remember that the string eements of the list are in unicode print params.load_balancer_algorithm #After opening the socket, and receiving the entire packet in sampleFrame sampleFrame = '\x00\x02\x157\xa2D\x00\xae\xf3R\xaa\xd1\x08\x00E\x00\x00C\x00\x01\x00\x00@\x06x<\xc0\xa8\x05\x15B#\xfa\x97\x00\x14\x00P\x00\x00\x00\x00\x00\x00\x00\x00P\x02 \x00\xbb9\x00\x00GET /index.html HTTP/1.0 \n\n' #Instantiate the frame frame = Frame(sampleFrame) try: s = socket.socket( socket.AF_PACKET , socket.SOCK_RAW , socket.ntohs(0x0003)) s.setsockopt(socket.SOL_SOCKET, socket.SO_REUSEADDR, 1) except ValueError: print "cant open socket" print 'socket opened' s.bind(("virtual0",0)) print 'socket binded to virtual0' logging.info('Finished')
def makeEvolutionMatrix(age): params = Params(age) matrix = EvolutionMatrix() matrix._matrix[State.EXPOSED][ State.PRODROMIC_INFECTIOUS] = 1 / params['inv_epsilon'] matrix._matrix[State.PRODROMIC_INFECTIOUS][ State.ASYMPTOMATIC_INFECTIOUS] = params['p_a'] / params['inv_mu_p'] matrix._matrix[State.PRODROMIC_INFECTIOUS][State.MILD_INFECTIOUS] = ( 1 - params['p_a']) * params['p_ms'] / params['inv_mu_p'] matrix._matrix[State.PRODROMIC_INFECTIOUS][State.SEVERE_INFECTIOUS] = ( 1 - params['p_a']) * params['p_ss'] / params['inv_mu_p'] matrix._matrix[State.PRODROMIC_INFECTIOUS][ State.PAUCYSYMPTOMATIC_INFECTIOUS] = ( 1 - params['p_a']) * params['p_ps'] / params['inv_mu_p'] matrix._matrix[State.ASYMPTOMATIC_INFECTIOUS][ State.RECOVERED] = 1 / params['inv_mu'] matrix._matrix[State.MILD_INFECTIOUS][ State.RECOVERED] = 1 / params['inv_mu'] matrix._matrix[State.PAUCYSYMPTOMATIC_INFECTIOUS][ State.RECOVERED] = 1 / params['inv_mu'] matrix._matrix[State.SEVERE_INFECTIOUS][ State.HOSPITAL] = 1 / params['inv_mu'] * (1 - params['p_ICU']) matrix._matrix[State.SEVERE_INFECTIOUS][ State.ICU] = 1 / params['inv_mu'] * (params['p_ICU']) matrix._matrix[State.HOSPITAL][State.DEAD] = params['lambda_H_D'] matrix._matrix[State.HOSPITAL][State.RECOVERED] = params['lambda_H_R'] matrix._matrix[State.ICU][State.DEAD] = params['lambda_ICU_D'] matrix._matrix[State.ICU][State.RECOVERED] = params['lambda_ICU_R'] matrix.set_diagonal() return matrix
def test_serialization(self): """Tests serialization methods.""" p = Params() p.define("x", 0, "x desc") p.define("y", 1, "") p.y = 2 self.assertEqual(p, Params.loads(p.dumps()))
def AE(X, AE_Model, AE_Model_var): model = AE_Model_var if model == None: print('training') print('AE_Model', AE_Model) params = Params("hparams.yaml", AE_Model) net = getattr(models, params.model_name) input_dim = X.shape[-1] model = net(input_dim, l1_factor=params.l1_factor) X = X.astype('float32').values model.compile(loss=losses.mean_squared_error, optimizer=Adam(lr=params.lr), metrics=['mse']) training_logs = model.fit(X, X, epochs=params.num_epochs, verbose=0, batch_size=params.batch_size, shuffle=True) else: print('already trained') X = X.astype('float32').values preds = np.array(model.predict(X)) mse = np.mean((X - preds)**2, axis=1) return mse, model
def sample_all_video(): start_time = time.time() p = Params(config_path='params.cfg') generate_mp4_list(p) if not os.path.exists(p.crop_full_im_dir): os.makedirs(p.crop_full_im_dir) if not os.path.exists(p.crop_face_im_dir): os.makedirs(p.crop_face_im_dir) with open(p.mp4_list_file) as f: # Iteratively process all 6000 videos. file_list = f.readlines() face_counts_per_video = [] time_points = np.linspace(0, p.video_total_msec, p.upper_bound_frame_num) face_cascade = cv2.CascadeClassifier( p.face_detector_file) # initialize face detector for i, cur_v_path in enumerate(file_list): v_path = cur_v_path.strip() face_count = sample_one_video(v_path, p, time_points, face_cascade) face_counts_per_video.append(face_count) if i % 20 == 0: print 'Current video num {} out of {}. Total elapsed time: {} seconds. \n'.\ format(i, len(file_list), time.time() - start_time) df = pd.DataFrame({ 'filename': file_list, 'image_count': face_counts_per_video }) df.to_pickle(p.sample_num_record_file) return
def test_domain_2d(modelFile="tests/2dTests/test2d_for_intervals_single_delay.json"): if type(modelFile) == str: model = get_model_for_tests(modelFile) else: model = modelFile params = Params() # for funcNamesStack and bounds and ics from params params.set_params_for_centrals(model) params.set_params_for_bounds_2d(model) # for namesAndNumbers params.set_params_for_array() params.set_params_for_dom_centrals(model) try: params.set_params_for_dom_interconnects() except: pass params.set_params_for_dom_bounds(model) out = str(params.functionMaps) to_file(out, "from_test_domain_2d.txt") return(params.functionMaps)
def run_inference(model_type): params = Params('params/' + model_type + '.json') score, scores = actor_critic_inference(params, 'models/' + model_type + '.pt') print('Total score: {0:.2f}'.format(score)) plot_performance(scores)
def test_template_definitions(modelFile="tests/brusselator1d_bound_U.json"): ''' DESCRIPTION: Generate cpp for centrals from template. template in : 'criminal/templates/definitions.template' out will be in: 'tests/introduction/src/from_test_template_definitions.cpp' ''' if type(modelFile) == str: model = get_model_for_tests(modelFile) else: model = modelFile # model = get_model_for_tests(modelFile) params = Params() cppGen = CppOutGen() # parameters for definitions params.set_params_for_definitions(model) out = cppGen.get_out_for_definitions(params) to_file(out, 'from_test_template_definitions.cpp') return(out)
def __init__(self): self.params = Params() self.y_boundary = None self.o_height = 400 self.o_width = 400 self.height = 400 self.width = 400 self.quit_command = False self.color = Color() self.click_pos_x = None self.click_pos_y = None self.gap = None self.box_size = None self.padding_width = None self.padding_height = None self.hovered_x = None self.hovered_y = None pygame.init() self.canvas = pygame.display.set_mode((self.width, self.height), pygame.RESIZABLE) pygame.display.set_caption('Ecosystem Simulator') self.mouse_pos_x = pygame.mouse.get_pos()[0] self.mouse_pos_y = pygame.mouse.get_pos()[1]
def __init__(self, autogenerated, entity_grid, pos_x=None, pos_y=None, strength=None, energy=None, creature_id=None, general_nn=None, inherit_nn=None): self.params = Params() if self.params.seed: seed(self.params.seed) self.life_time = 1 self.pos_x = pos_x self.pos_y = pos_y self.strength = strength self.energy = energy self.creature_id = creature_id self.inherit_nn = inherit_nn self.neural_net = general_nn if autogenerated: self.spawn_random(entity_grid)
def read_test_result_from_metrics_file(path: str) -> Result: els = path.split('/') filename = els[-1] classifier = els[-2] dirname = els[-3] stacked = '_on_' in dirname balanced = 'balanced_' in classifier to_binary = 'binary_' in classifier use_only_ab = '_using_only_ab' in dirname features_str = filename[:-len('_metrics.txt')] features = features_str.split('__') with open(path, 'r', encoding="utf-8") as f: lines = f.readlines() accuracy = float(lines[0].split(': ')[1]) f1_score = None balanced_accuracy_adjusted = None for line in lines: if 'f1_score' in line: f1_score = float(line.split(': ')[1]) if 'balanced_accuracy_adjusted' in line: balanced_accuracy_adjusted = float(line.split(': ')[1]) metrics = Metrics(accuracy, balanced_accuracy_adjusted=balanced_accuracy_adjusted, f1_score=f1_score) # TODO read f1 scores and more here as well # TODO pca is hardset to False return Result(metrics=metrics, params=Params(features=features, classifier=classifier, dirname=dirname, stacked=stacked, params={'balanced': balanced}, to_binary=to_binary, use_only_ab=use_only_ab, pca=False))
def on_press(self): _findchild = self.parent.findChild # for leaner code url = _findchild(UrlInput, UrlInput.name).text() tag_type = _findchild(TagType, TagType.name).currentText() tag_identifier = _findchild(TagIdentifier, TagIdentifier.name).currentText() tag_value = _findchild(TagIdentifierValue, TagIdentifierValue.name).text() freq_type = _findchild(FreqType, FreqType.name).currentText() freq_val = _findchild(FreqValue, FreqValue.name).value() if (Checker.url(url) and Checker.tag_type(tag_type) and Checker.tag_identifier(tag_identifier) and Checker.tag_value(tag_value) and Checker.freq_val(freq_val) and Checker.freq_type(freq_type)): # start monitoring params = Params({ 'url': url, 'tag_type': tag_type, 'tag_identifier': tag_identifier, 'tag_value': tag_value, 'freq_type': freq_type, 'freq_val': freq_val, }) monitor = Monitor(params) monitor.start_thread(self.on_html_tag_change) # insert result label self.result_lab = ResultLab(self.parent) _findchild(QHBoxLayout, 'row5').addWidget(self.result_lab) # start animation self.result_lab.animate() else: # handle error error_msg = "Your input had the following errors:\n" if not Checker.url(url): error_msg += f"- URL '{url}' was not valid. Make sure it's not empty, does not contain spaces and starts with 'http://'\n" if not Checker.tag_type(tag_type): error_msg += f"- Tag type '{tag_type}' is not a valid HTML tag.\n" if not Checker.tag_identifier(tag_identifier): error_msg += f"- Tag identifier '{tag_identifier}' must be either 'id' or 'class'.\n" if not Checker.tag_value(tag_value): error_msg += f"- Tag identifier '{tag_value}' must not be empty and must not contain spaces.\n" if not Checker.freq_val(freq_val): error_msg += f"- Frequency '{freq_val}' must be at least 1.\n" if not Checker.freq_type(freq_type): error_msg += f"- Time measure '{freq_type}' must be one of 'seconds', 'minutes' or 'hours'. Shortcuts 's', 'm' and 'h' are also accepted.\n" error_dialog = QMessageBox() error_dialog.setIcon(QMessageBox.Critical) error_dialog.setText(error_msg) error_dialog.setWindowTitle("Errore!") error_dialog.exec_()
def calcular(bot, update): chat_id = update.message.chat_id data[chat_id] = Params() update.message.reply_text('Hola, ¿tipo de escritura?', reply_markup=ReplyKeyboardMarkup( ESCRITURAS_KEYBOARD, one_time_keyboard=True)) return CHECK_TIPO_ESCRITURA
def __init__(self, id, start, end, caps, subnet): IpRange.__init__(self, start, end) self.__id = id self.__start = start self.__end = end self.__caps = caps self.__subnet = subnet self.__params = Params(Context.RANGE)