codex-storage/das-research
1
0
Fork 0
mirror of https://github.com/codex-storage/das-research.git synced 2025-02-21 15:28:09 +00:00
Code Issues Packages Projects Releases Wiki Activity

fixed bug and generalized averaging

Browse Source
This commit is contained in:
HajarZaiz 2023-03-30 11:24:30 +00:00
parent aa99601826
commit 89368d5fbe
2 changed files with 62 additions and 35 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

79
DAS/visualizer.py
View File

@ -6,19 +6,21 @@ import matplotlib.pyplot as plt
import numpy as np import numpy as np
import seaborn as sns import seaborn as sns
from itertools import combinations from itertools import combinations
from config_example import runs
class Visualizer: class Visualizer:
def __init__(self, execID): def __init__(self, execID):
self.execID = execID self.execID = execID
self.folderPath = "results/"+self.execID self.folderPath = "results/"+self.execID
self.parameters = ['run', 'blockSize', 'failureRate', 'numberNodes', 'netDegree', self.parameters = ['run', 'blockSize', 'failureRate', 'numberNodes', 'netDegree', 'chi', 'vpn1', 'vpn2', 'class1ratio', 'bwUplinkProd', 'bwUplink1', 'bwUplink2']
'chi', 'vpn1', 'vpn2', 'bwUplinkProd', 'bwUplink1', 'bwUplink2']
self.minimumDataPoints = 2 self.minimumDataPoints = 2
self.maxTTA = 999
def plottingData(self): def plottingData(self):
"""Store data with a unique key for each params combination""" """Store data with a unique key for each params combination"""
data = {} data = {}
bw = []
"""Loop over the xml files in the folder""" """Loop over the xml files in the folder"""
for filename in os.listdir(self.folderPath): for filename in os.listdir(self.folderPath):
"""Loop over the xmls and store the data in variables""" """Loop over the xmls and store the data in variables"""
@ -29,6 +31,7 @@ class Visualizer:
blockSize = int(root.find('blockSize').text) blockSize = int(root.find('blockSize').text)
failureRate = int(root.find('failureRate').text) failureRate = int(root.find('failureRate').text)
numberNodes = int(root.find('numberNodes').text) numberNodes = int(root.find('numberNodes').text)
class1ratio = float(root.find('class1ratio').text)
netDegree = int(root.find('netDegree').text) netDegree = int(root.find('netDegree').text)
chi = int(root.find('chi').text) chi = int(root.find('chi').text)
vpn1 = int(root.find('vpn1').text) vpn1 = int(root.find('vpn1').text)
@ -38,25 +41,31 @@ class Visualizer:
bwUplink2 = int(root.find('bwUplink2').text) bwUplink2 = int(root.find('bwUplink2').text)
tta = int(root.find('tta').text) tta = int(root.find('tta').text)
"""Loop over all possible combinations of length 4 of the parameters""" # if tta == -1:
# tta = self.maxTTA
"""Store BW"""
bw.append(bwUplinkProd)
"""Loop over all possible combinations of length of the parameters minus x, y params"""
for combination in combinations(self.parameters, len(self.parameters)-2): for combination in combinations(self.parameters, len(self.parameters)-2):
# Get the indices and values of the parameters in the combination # Get the indices and values of the parameters in the combination
indices = [self.parameters.index(element) for element in combination] indices = [self.parameters.index(element) for element in combination]
selectedValues = [run, blockSize, failureRate, numberNodes, netDegree, chi, vpn1, vpn2, bwUplinkProd, bwUplink1, bwUplink2] selectedValues = [run, blockSize, failureRate, numberNodes, netDegree, chi, vpn1, vpn2, class1ratio, bwUplinkProd, bwUplink1, bwUplink2]
values = [selectedValues[index] for index in indices] values = [selectedValues[index] for index in indices]
names = [self.parameters[i] for i in indices] names = [self.parameters[i] for i in indices]
keyComponents = [f"{name}_{value}" for name, value in zip(names, values)] keyComponents = [f"{name}_{value}" for name, value in zip(names, values)]
key = tuple(keyComponents[:len(self.parameters)-2]) key = tuple(keyComponents[:len(self.parameters)-2])
"""Get the names of the other 2 parameters that are not included in the key""" """Get the names of the other parameters that are not included in the key"""
otherParams = [self.parameters[i] for i in range(len(self.parameters)) if i not in indices] otherParams = [self.parameters[i] for i in range(len(self.parameters)) if i not in indices]
"""Append the values of the other 2 parameters and the ttas to the lists for the key""" """Append the values of the other parameters and the ttas to the lists for the key"""
otherIndices = [i for i in range(len(self.parameters)) if i not in indices] otherIndices = [i for i in range(len(self.parameters)) if i not in indices]
"""Initialize the dictionary for the key if it doesn't exist yet""" """Initialize the dictionary for the key if it doesn't exist yet"""
if key not in data: if key not in data:
data[key] = {} data[key] = {}
"""Initialize lists for the other 2 parameters and the ttas with the key""" """Initialize lists for the other parameters and the ttas with the key"""
data[key][otherParams[0]] = [] data[key][otherParams[0]] = []
data[key][otherParams[1]] = [] data[key][otherParams[1]] = []
data[key]['ttas'] = [] data[key]['ttas'] = []
@ -73,14 +82,14 @@ class Visualizer:
print("Getting data from the folder...") print("Getting data from the folder...")
return data return data
def averageRuns(self, data): def averageRuns(self, data, runs):
"""Get the average of run 0 and run 1 for each key""" """Get the average of run 0 and run 1 for each key"""
newData = {} newData = {}
for key, value in data.items(): for key, value in data.items():
runExists = False runExists = False
"""Check if the key contains 'run_' with a numerical value""" """Check if the key contains 'run_' with a numerical value"""
for item in key: for item in key:
if item.startswith('run_0'): if item.startswith('run_'):
runExists = True runExists = True
break break
if runExists: if runExists:
@ -88,20 +97,20 @@ class Visualizer:
"""Create a new key with the other items in the tuple""" """Create a new key with the other items in the tuple"""
if item.startswith('run_'): if item.startswith('run_'):
newKey = tuple([x for x in key if x != item]) newKey = tuple([x for x in key if x != item])
key0 = ('run_0',) + newKey """Average the similar key values"""
data0 = data[key0] total = [0] * len(data[key]['ttas'])
key1 = ('run_1',) + newKey for i in range(runs):
data1 = data[key1] key0 = (f'run_{i}',) + newKey
for cnt, tta in enumerate(data[key0]['ttas']):
total[cnt] += tta
for i in range(len(total)):
total[i] = total[i]/runs
averages = {} averages = {}
for key in data0.keys(): for subkey in data[key].keys():
if key in data1: if subkey == 'ttas':
values = [] averages[subkey] = total
for i in range(len(data0[key])): else:
value = (data0[key][i] + data1[key][i]) / 2.0 averages[subkey] = data[key][subkey]
if isinstance(data0[key][i], int) and isinstance(data1[key][i], int) and key != 'ttas':
value = int(value)
values.append(value)
averages[key] = values
newData[newKey] = averages newData[newKey] = averages
print("Getting the average of the runs...") print("Getting the average of the runs...")
return newData return newData
@ -135,9 +144,12 @@ class Visualizer:
def plotHeatmaps(self): def plotHeatmaps(self):
"""Plot and store the 2D heatmaps in subfolders""" """Plot and store the 2D heatmaps in subfolders"""
data = self.plottingData() data= self.plottingData()
data = self.averageRuns(data) """Average the runs if needed"""
if(len(runs) > 1):
data = self.averageRuns(data, len(runs))
filteredKeys = self.similarKeys(data) filteredKeys = self.similarKeys(data)
vmin, vmax = 0, self.maxTTA
print("Plotting heatmaps...") print("Plotting heatmaps...")
"""Create the directory if it doesn't exist already""" """Create the directory if it doesn't exist already"""
@ -152,10 +164,10 @@ class Visualizer:
ylabels = np.sort(np.unique(data[key][labels[1]])) ylabels = np.sort(np.unique(data[key][labels[1]]))
if len(xlabels) < self.minimumDataPoints or len(ylabels) < self.minimumDataPoints: if len(xlabels) < self.minimumDataPoints or len(ylabels) < self.minimumDataPoints:
continue continue
hist, xedges, yedges = np.histogram2d(data[key][labels[0]], data[key][labels[1]], bins=(len(xlabels), len(ylabels)), weights=data[key]['ttas']) hist, xedges, yedges = np.histogram2d(data[key][labels[0]], data[key][labels[1]], bins=(len(xlabels), len(ylabels)), weights=data[key]['ttas'], normed=False)
hist = hist.T hist = hist.T
fig, ax = plt.subplots(figsize=(10, 6)) fig, ax = plt.subplots(figsize=(10, 6))
sns.heatmap(hist, xticklabels=xlabels, yticklabels=ylabels, cmap='Purples', cbar_kws={'label': 'Time to block availability'}, linecolor='black', linewidths=0.3, annot=True, fmt=".2f", ax=ax) sns.heatmap(hist, xticklabels=xlabels, yticklabels=ylabels, cmap='Purples', cbar_kws={'label': 'Time to block availability'}, linecolor='black', linewidths=0.3, annot=True, fmt=".2f", ax=ax, vmin=vmin, vmax=vmax)
plt.xlabel(self.formatLabel(labels[0])) plt.xlabel(self.formatLabel(labels[0]))
plt.ylabel(self.formatLabel(labels[1])) plt.ylabel(self.formatLabel(labels[1]))
filename = "" filename = ""
@ -177,3 +189,18 @@ class Visualizer:
plt.savefig(os.path.join(targetFolder, filename)) plt.savefig(os.path.join(targetFolder, filename))
plt.close() plt.close()
plt.clf() plt.clf()
def plotHist(self, bandwidth):
"""Plot Bandwidth Frequency Histogram"""
plt.hist(bandwidth, bins=5)
plt.xlabel('Bandwidth')
plt.ylabel('Frequency')
plt.title('Bandwidth Histogram')
"""Create the directory if it doesn't exist already"""
histogramFolder = self.folderPath + '/histogram'
if not os.path.exists(histogramFolder):
os.makedirs(histogramFolder)
filename = os.path.join(histogramFolder, 'histogram.png')
plt.savefig(filename)
plt.clf()

18
config_example.py
View File

@ -24,36 +24,36 @@ logLevel = logging.INFO
# number of parallel workers. -1: all cores; 1: sequential # number of parallel workers. -1: all cores; 1: sequential
# for more details, see joblib.Parallel # for more details, see joblib.Parallel
numJobs = 3 numJobs = -1
# distribute rows/columns evenly between validators (True) # distribute rows/columns evenly between validators (True)
# or generate it using local randomness (False) # or generate it using local randomness (False)
evenLineDistribution = True evenLineDistribution = True
# Number of simulation runs with the same parameters for statistical relevance # Number of simulation runs with the same parameters for statistical relevance
runs = range(10) runs = range(3)
# Number of validators # Number of validators
numberNodes = range(256, 513, 128) numberNodes = range(256, 313, 128)
# Percentage of block not released by producer # Percentage of block not released by producer
failureRates = range(10, 91, 40) failureRates = range(10, 31, 40)
# Block size in one dimension in segments. Block is blockSizes * blockSizes segments. # Block size in one dimension in segments. Block is blockSizes * blockSizes segments.
blockSizes = range(32,65,16) blockSizes = range(32,35,16)
# Per-topic mesh neighborhood size # Per-topic mesh neighborhood size
netDegrees = range(6, 9, 2) netDegrees = range(6, 9, 2)
# number of rows and columns a validator is interested in # number of rows and columns a validator is interested in
chis = range(1, 5, 2) chis = range(2, 5, 2)
# ratio of class1 nodes (see below for parameters per class) # ratio of class1 nodes (see below for parameters per class)
class1ratios = np.arange(0, 1, .2) class1ratios = [0.8, 0.9]
# Number of validators per beacon node # Number of validators per beacon node
validatorsPerNode1 = [1] validatorsPerNode1 = [1]
validatorsPerNode2 = [2, 4, 8, 16, 32] validatorsPerNode2 = [50]
# Set uplink bandwidth. In segments (~560 bytes) per timestep (50ms?) # Set uplink bandwidth. In segments (~560 bytes) per timestep (50ms?)
# 1 Mbps ~= 1e6 / 20 / 8 / 560 ~= 11 # 1 Mbps ~= 1e6 / 20 / 8 / 560 ~= 11
@ -62,7 +62,7 @@ bwUplinks1 = [110]
bwUplinks2 = [2200] bwUplinks2 = [2200]
# Set to True if you want your run to be deterministic, False if not # Set to True if you want your run to be deterministic, False if not
deterministic = False deterministic = True
# If your run is deterministic you can decide the random seed. This is ignore otherwise. # If your run is deterministic you can decide the random seed. This is ignore otherwise.
randomSeed = "DAS" randomSeed = "DAS"