Renamed utility.simulations to utility.simulation; Implemented first version of the SimulationManager class

This commit is contained in:
Andreas Tsouchlos 2022-11-13 23:14:47 +01:00
parent 3e02dcf17c
commit ffcce7b3f2
3 changed files with 346 additions and 119 deletions

View File

@ -1,7 +1,7 @@
import unittest
import numpy as np
from utility import simulations, noise, codes
from utility import simulation, noise, codes
class CountBitErrorsTestCase(unittest.TestCase):
@ -17,9 +17,9 @@ class CountBitErrorsTestCase(unittest.TestCase):
d3 = np.array([0, 0, 0, 0])
y_hat3 = np.array([1, 1, 1, 1])
self.assertEqual(simulations.count_bit_errors(d1, y_hat1), 2)
self.assertEqual(simulations.count_bit_errors(d2, y_hat2), 0)
self.assertEqual(simulations.count_bit_errors(d3, y_hat3), 4)
self.assertEqual(simulation.count_bit_errors(d1, y_hat1), 2)
self.assertEqual(simulation.count_bit_errors(d2, y_hat2), 0)
self.assertEqual(simulation.count_bit_errors(d3, y_hat3), 4)
# TODO: Rewrite tests for new SNR calculation

342
sw/utility/simulation.py Normal file
View File

@ -0,0 +1,342 @@
"""This file contains utility functions relating to tests and simulations of the decoders."""
import time
import numpy as np
import typing
from tqdm import tqdm
from timeit import default_timer
import signal
from dataclasses import dataclass
import pickle
import os.path
from pathlib import Path
import spdlog as spd
from utility import noise
def count_bit_errors(d: np.array, d_hat: np.array) -> int:
"""Count the number of wrong bits in a decoded codeword.
:param d: Originally sent data
:param d_hat: Received data
:return: Number of bit errors
"""
return np.sum(d != d_hat)
# def test_decoder(n: int,
# k: int,
# decoder: typing.Any,
# SNRs: typing.Sequence[float] = np.linspace(1, 7, 7),
# target_frame_errors: int = 100) \
# -> typing.Tuple[np.array, np.array]:
# """Calculate the Bit Error Rate (BER) for a given decoder for a number of SNRs.
#
# This function assumes the all-zeros assumption holds. Progress is printed to stdout.
#
# :param n: Length of a codeword of the used code
# :param k: Length of a dataword of the used code
# :param decoder: Instance of the decoder to be tested
# :param SNRs: List of SNRs for which the BER should be calculated
# :param target_frame_errors: Number of frame errors after which to stop the simulation
# :param N_max: Maximum number of iterations to perform for each SNR
# :return: Tuple of numpy arrays of the form (SNRs, BERs)
# """
#
# x = np.zeros(n)
# x_bpsk = 1 - 2 * x # Map x from [0, 1]^n to [-1, 1]^n
#
# BERs = []
# for SNR in tqdm(SNRs,
# desc=f"Calculating BERs for {decoder.__class__.__name__}",
# position=1,
# leave=False,
# bar_format="{l_bar}{bar}| {n_fmt}/{total_fmt}"):
#
# total_bit_errors = 0
# total_bits = 0
# total_frame_errors = 0
#
# pbar = tqdm(total=target_frame_errors,
# desc=f"Simulating for SNR = {SNR} dB",
# position=2,
# leave=False,
# bar_format="{l_bar}{bar}| {n_fmt}/{total_fmt} [{elapsed}<{remaining}]")
#
# while total_frame_errors < target_frame_errors:
# # Simulate channel
# y = noise.add_awgn(x_bpsk, SNR, n, k)
#
# # Decode received frame
# x_hat = decoder.decode(y)
#
# # Calculate statistics
# bit_errors = count_bit_errors(x, x_hat)
# total_bits += x.size
#
# if bit_errors > 0:
# total_frame_errors += 1
# total_bit_errors += bit_errors
# pbar.update(1)
#
# pbar.close()
#
# BERs.append(total_bit_errors / total_bits)
#
# return np.array(SNRs), np.array(BERs)
#
#
# def test_decoders(n: int,
# k: int,
# decoders: typing.List,
# SNRs: typing.Sequence[float] = np.linspace(1, 7, 7),
# target_frame_errors: int = 100) \
# -> typing.Tuple[np.array, np.array]:
# """Calculate the Bit Error Rate (BER) for a number of given decoders for a number of SNRs.
#
# This function assumes the all-zeros assumption holds. Progress is printed to stdout.
#
# :param n: Length of a codeword of the used code
# :param k: Length of a dataword of the used code
# :param decoders: List of decoder objects to be tested
# :param SNRs: List of SNRs for which the BER should be calculated
# :param target_frame_errors: Number of frame errors after which to stop the simulation
# :return: Tuple of the form (SNRs, [BERs_1, BERs_2, ...]) where SNR and BERs_x are numpy arrays
# """
# result_BERs = []
#
# start_time = default_timer()
#
# for decoder in tqdm(decoders,
# desc="Calculating the answer to life, the universe and everything",
# position=0,
# leave=False,
# bar_format="{l_bar}{bar}| {n_fmt}/{total_fmt}"):
# _, BERs = test_decoder(n, k, decoder, SNRs, target_frame_errors)
# result_BERs.append(BERs)
#
# end_time = default_timer()
# print(f"Elapsed time: {end_time - start_time:.2f}s")
#
# return SNRs, result_BERs
#
@dataclass
class SimulationParameters:
n: int
k: int
decoders: typing.Sequence[typing.Any]
SNRs: typing.Sequence[float]
target_frame_errors: int
@dataclass
class SimulationState:
"""Data structure storing the state of the simulation."""
num_frame_errors: int = 0
num_bit_errors: int = 0
num_total_bits: int = 0
# simulation_time: float = 0
current_decoder_index = 0
current_SNRs_index: int = 0
# TODO: Make more generic
# TODO: Remove save data after successful execution
class SimulationManager:
def __init__(self, save_dir: str, results_dir: str):
self._save_dir = save_dir
self._sim_parameters_filepath = f"{self._save_dir}/sim_parameters.pickle"
self._sim_state_filepath = f"{self._save_dir}/sim_state.pickle"
self._logs_filepath = f"{self._save_dir}/logs.txt"
self._results_dir = results_dir
# TODO: Should the be none or SimulationParameters() and SimulationState() respectively?
self._sim_params = None
self._sim_state = None
self._sim_running = False
Path(self._save_dir).mkdir(parents=True, exist_ok=True)
self._logger = spd.FileLogger("SimulationManager", self._logs_filepath)
self._logger.set_level(spd.LogLevel.DEBUG)
signal.signal(signal.SIGINT, self._exit_gracefully)
signal.signal(signal.SIGTERM, self._exit_gracefully)
#
# Functions relating to the pausing and restarting of simulations
#
def unfinished_simulation_present(self) -> bool:
return os.path.isfile(self._sim_parameters_filepath) \
and os.path.isfile(self._sim_state_filepath)
def continue_unfinished(self):
assert self.unfinished_simulation_present()
with open(self._sim_parameters_filepath, "rb") as file:
self._sim_params = pickle.load(file)
with open(self._sim_state_filepath, "rb") as file:
self._sim_state = pickle.load(file)
self._logger.info("Loaded saved simulation state")
self.start()
# TODO: Make sure old state is overwritten
def _save_state(self):
with open(self._sim_parameters_filepath, "wb") as file:
pickle.dump(self._sim_params, file)
with open(self._sim_state_filepath, "wb") as file:
pickle.dump(self._sim_state, file)
self._logger.info("Saved simulation state")
def _exit_gracefully(self, *args):
self._logger.debug("Intercepted signal SIGINT/SIGTERM")
self._sim_running = False
if (self._sim_params is not None) and (self._sim_state is not None):
self._save_state()
#
# Functions responsible for the actual simulation
#
# def test_decoders(self,
# n: int,
# k: int,
# decoders: typing.Sequence[typing.Any],
# SNRs: typing.Sequence[float] = np.linspace(1, 7, 7),
# target_frame_errors: int = 100):
# """Calculate the Bit Error Rate (BER) for a number of given decoders for a number of SNRs.
#
# This function assumes the all-zeros assumption holds. Progress is printed to stdout.
#
# :param n: Length of a codeword of the used code
# :param k: Length of a dataword of the used code
# :param decoders: List of decoder objects to be tested
# :param SNRs: List of SNRs for which the BER should be calculated
# :param target_frame_errors: Number of frame errors after which to stop the simulation
# :return: Tuple of the form (SNRs, [BERs_1, BERs_2, ...]) where SNR and BERs_x are numpy arrays
# """
# # TODO
#
# # Save simulation
# self._sim_parameters = SimulationMetaData(n, k, decoders, SNRs, target_frame_errors)
# self._sim_state = SimulationState()
#
# self._logger.info("Initialized new simulation state")
#
# # Simulation
#
# result_BERs = []
#
# start_time = default_timer()
#
# for decoder in tqdm(decoders,
# desc="Calculating the answer to life, the universe and everything",
# position=0,
# leave=False,
# bar_format="{l_bar}{bar}| {n_fmt}/{total_fmt}"):
# _, BERs = self.test_decoder(n, k, decoder, SNRs, target_frame_errors)
# result_BERs.append(BERs)
#
# end_time = default_timer()
# print(f"Elapsed time: {end_time - start_time:.2f}s")
#
# return SNRs, result_BERs
# def test_decoder(self,
# n: int,
# k: int,
# decoder: typing.Any,
# SNRs: typing.Sequence[float] = np.linspace(1, 7, 7),
# target_frame_errors: int = 100) \
# -> typing.Tuple[np.array, np.array]:
def start(self):
self._sim_running = True # TODO: Move this somewhere else
decoder = self._sim_params.decoders[self._sim_state.current_decoder_index]
x = np.zeros(self._sim_params.n)
x_bpsk = 1 - 2 * x # Map x from [0, 1]^n to [-1, 1]^n
BERs = []
for SNR in tqdm(self._sim_params.SNRs[self._sim_state.current_SNRs_index:],
desc=f"Calculating BERs for {decoder.__class__.__name__}",
position=1,
leave=False,
bar_format="{l_bar}{bar}| {n_fmt}/{total_fmt}"):
pbar = tqdm(total=self._sim_params.target_frame_errors,
desc=f"Simulating for SNR = {SNR} dB",
position=2,
leave=False,
bar_format="{l_bar}{bar}| {n_fmt}/{total_fmt} [{elapsed}<{remaining}]")
pbar.update(self._sim_state.num_frame_errors)
while self._sim_state.num_frame_errors < self._sim_params.target_frame_errors:
if not self._sim_running:
return
# Simulate channel
y = noise.add_awgn(x_bpsk, SNR, self._sim_params.n, self._sim_params.k)
# Decode received frame
x_hat = decoder.decode(y)
# Calculate statistics
bit_errors = count_bit_errors(x, x_hat)
self._sim_state.num_total_bits += x.size
if bit_errors > 0:
self._sim_state.num_frame_errors += 1
self._sim_state.num_bit_errors += bit_errors
pbar.update(1)
# TODO: Load BERs from file as well
BERs.append(self._sim_state.num_bit_errors / self._sim_state.num_total_bits)
pbar.close()
self._sim_state.current_SNRs_index += 1
self._sim_state.num_frame_errors = 0
self._sim_state.num_bit_errors = 0
self._sim_state.num_total_bits = 0
# return np.array(self._sim_params.SNRs), np.array(BERs)
class DecoderTester:
"""Class used to test decoders simulating BPSK modulation and an AWGN channel.
Allows for recovering a stopped simulation if its previous state is known.
"""
def __init__(self, initial_sim_state: SimulationState = SimulationState()):
"""Construct a DecoderTester object.
:param initial_sim_state: State the simulation should start from
"""
self._state = initial_sim_state
def get_state(self) -> SimulationState:
return self._state
def configure(self, n: int,
k: int,
SNRs: typing.Sequence[float] = np.linspace(1, 7, 7),
target_frame_errors: int = 100):
pass
def start_test(self, decoders: typing.List):
pass
def stop(self):
pass

View File

@ -1,115 +0,0 @@
"""This file contains utility functions relating to tests and simulations of the decoders."""
import numpy as np
import typing
from tqdm import tqdm
from timeit import default_timer
from utility import noise
def count_bit_errors(d: np.array, d_hat: np.array) -> int:
"""Count the number of wrong bits in a decoded codeword.
:param d: Originally sent data
:param d_hat: Received data
:return: Number of bit errors
"""
return np.sum(d != d_hat)
def test_decoder(n: int,
k: int,
decoder: typing.Any,
SNRs: typing.Sequence[float] = np.linspace(1, 7, 7),
target_frame_errors: int = 100) \
-> typing.Tuple[np.array, np.array]:
"""Calculate the Bit Error Rate (BER) for a given decoder for a number of SNRs.
This function assumes the all-zeros assumption holds. Progress is printed to stdout.
:param n: Length of a codeword of the used code
:param k: Length of a dataword of the used code
:param decoder: Instance of the decoder to be tested
:param SNRs: List of SNRs for which the BER should be calculated
:param target_frame_errors: Number of frame errors after which to stop the simulation
:param N_max: Maximum number of iterations to perform for each SNR
:return: Tuple of numpy arrays of the form (SNRs, BERs)
"""
x = np.zeros(n)
x_bpsk = 1 - 2 * x # Map x from [0, 1]^n to [-1, 1]^n
BERs = []
for SNR in tqdm(SNRs,
desc=f"Calculating BERs for {decoder.__class__.__name__}",
position=1,
leave=False,
bar_format="{l_bar}{bar}| {n_fmt}/{total_fmt}"):
total_bit_errors = 0
total_bits = 0
total_frame_errors = 0
pbar = tqdm(total=target_frame_errors,
desc=f"Simulating for SNR = {SNR} dB",
position=2,
leave=False,
bar_format="{l_bar}{bar}| {n_fmt}/{total_fmt} [{elapsed}<{remaining}]")
while total_frame_errors < target_frame_errors:
# Simulate channel
y = noise.add_awgn(x_bpsk, SNR, n, k)
# Decode received frame
x_hat = decoder.decode(y)
# Calculate statistics
bit_errors = count_bit_errors(x, x_hat)
total_bits += x.size
if bit_errors > 0:
total_frame_errors += 1
total_bit_errors += bit_errors
pbar.update(1)
pbar.close()
BERs.append(total_bit_errors / total_bits)
return np.array(SNRs), np.array(BERs)
def test_decoders(n: int,
k: int,
decoders: typing.List,
SNRs: typing.Sequence[float] = np.linspace(1, 7, 7),
target_frame_errors: int = 100) \
-> typing.Tuple[np.array, np.array]:
"""Calculate the Bit Error Rate (BER) for a number of given decoders for a number of SNRs.
This function assumes the all-zeros assumption holds. Progress is printed to stdout.
:param n: Length of a codeword of the used code
:param k: Length of a dataword of the used code
:param decoders: List of decoder objects to be tested
:param SNRs: List of SNRs for which the BER should be calculated
:param target_frame_errors: Number of frame errors after which to stop the simulation
:return: Tuple of the form (SNRs, [BERs_1, BERs_2, ...]) where SNR and BERs_x are numpy arrays
"""
result_BERs = []
start_time = default_timer()
for decoder in tqdm(decoders,
desc="Calculating the answer to life, the universe and everything",
position=0,
leave=False,
bar_format="{l_bar}{bar}| {n_fmt}/{total_fmt}"):
_, BERs = test_decoder(n, k, decoder, SNRs, target_frame_errors)
result_BERs.append(BERs)
end_time = default_timer()
print(f"Elapsed time: {end_time - start_time:.2f}s")
return SNRs, result_BERs