mirror of
https://github.com/vacp2p/10ksim.git
synced 2025-02-22 11:08:28 +00:00
Cleaned plotter and modified yaml file. Also added np to requirements
This commit is contained in:
Alberto Soutullo
parent
8c88ded854
commit
12094c9482
BIN
requirements.txt
BIN
requirements.txt
Binary file not shown.
@ -10,4 +10,11 @@ metrics_to_scrape:
|
||||
- "rate(container_network_receive_bytes_total{namespace='zerotesting'}[$__rate_interval])": "pod-node"
|
||||
- "rate(container_network_transmit_bytes_total{namespace='zerotesting'}[$__rate_interval])": "pod-node"
|
||||
- "sum by(job) (libp2p_gossipsub_low_peers_topics)" : "job"
|
||||
- "sum by(job) (libp2p_gossipsub_healthy_peers_topics)": "job"
|
||||
- "sum by(job) (libp2p_gossipsub_healthy_peers_topics)": "job"
|
||||
plotting:
|
||||
"bandwidth":
|
||||
"folder":
|
||||
- "test/nwaku/1000-1KB-1msgs/"
|
||||
"data":
|
||||
- "libp2p-rx"
|
||||
- "libp2p-tx"
|
||||
@ -1,38 +1,17 @@
|
||||
# Python Imports
|
||||
import logging
|
||||
import os
|
||||
|
||||
import numpy as np
|
||||
|
||||
import logger
|
||||
import pandas as pd
|
||||
import seaborn as sns
|
||||
from typing import List, Dict
|
||||
import matplotlib.pyplot as plt
|
||||
from matplotlib import ticker
|
||||
from matplotlib.axis import Axis
|
||||
from typing import List, Dict
|
||||
|
||||
# Project Imports
|
||||
from utils.file_utils import get_files_from_folder_path, get_file_name_from_path
|
||||
|
||||
# logger = logging.getLogger(__name__)
|
||||
logger = logging.getLogger(__name__)
|
||||
sns.set_theme()
|
||||
# En carpeta: metrica1, metrica2, metrica3...
|
||||
# o bien, metrica1_1, metrica1_2...
|
||||
# La carpeta padre, es carpetas de carpetas de metricas ()
|
||||
|
||||
|
||||
# Cuando se descarga, se crea la carpeta del experimento
|
||||
# Del experimento, se pueden crear plots (metrica),
|
||||
# Del experimento, se pueden meter varios plots juntos (metricas para comparar, RX-TX)
|
||||
# Pensar que del mismo experimento, se puedan meter varios plots por fila (compactación)
|
||||
# De varios experimentos, se pueden meter en el mismo plot (metrica)
|
||||
|
||||
|
||||
# test = {"bandwidth": {"folder": ["waku"], "data": ["RX", "TX"], "include": ["experiment1", "experiment2"]}}
|
||||
# Del tipo de programa "FOLDER", quiero que me plotees rx contra tx, y que me incluyas los experimentos x y z
|
||||
# en este caso se plotearán tantos plots como "data" haya. con INCLUDE boxplots, y FOLDER tipos de boxplots.
|
||||
# test = {"bandwidth": {"folder": ["../../data/plotter_data_test/rust/", "../../data/plotter_data_test/go/"], "data": ["rx", "tx"], "include": ["0.5KB-1.csv", "0.5KB-1.csv"]}}
|
||||
test = {"bandwidth": {"folder": ["../../test/nwaku/1000-1KB-1msgs/"], "data": ["libp2p-rx", "libp2p-tx"], "include": ["0.5KB-1.csv", "0.5KB-1.csv"]}}
|
||||
|
||||
|
||||
class Plotter:
|
||||
@ -66,9 +45,9 @@ class Plotter:
|
||||
data_files_path = get_files_from_folder_path(subplot_path)
|
||||
|
||||
for file_path in data_files_path:
|
||||
group_df = self._dump_file_mean_into_df(file_path, group_df)
|
||||
group_df = self._dump_file_mean_into_df(subplot_path+"/"+file_path, group_df)
|
||||
|
||||
subplot_df["class"] = subplot_path.split("/")[-2]
|
||||
group_df["class"] = subplot_path.split("/")[-2]
|
||||
subplot_df = pd.concat([subplot_df, group_df])
|
||||
|
||||
subplot_df = pd.melt(subplot_df, id_vars=["class"])
|
||||
@ -101,103 +80,3 @@ class Plotter:
|
||||
]
|
||||
|
||||
return subplot_path
|
||||
|
||||
|
||||
def add_data_to_df(main_df, df: pd.DataFrame, name):
|
||||
df_avg = df.mean()
|
||||
df_avg = pd.DataFrame(df_avg, columns=[name])
|
||||
|
||||
concat_df = pd.concat([main_df, df_avg], axis=1)
|
||||
|
||||
return concat_df
|
||||
|
||||
|
||||
def create_plot():
|
||||
fig, axs = plt.subplots(nrows=1, ncols=2, sharey='row')
|
||||
|
||||
folders_grouped = [
|
||||
[f"{folder}{data}" for folder in test["bandwidth"]["folder"]]
|
||||
for data in test["bandwidth"]["data"]
|
||||
]
|
||||
|
||||
for k, group in enumerate(folders_grouped):
|
||||
main_df = pd.DataFrame()
|
||||
for folder_path in group:
|
||||
folder_df = pd.DataFrame()
|
||||
files = [f for f in os.listdir(folder_path) if
|
||||
os.path.isfile(os.path.join(folder_path, f))]
|
||||
for file in files:
|
||||
if not os.path.exists(folder_path+"/"+file): continue # todo delete this line
|
||||
df = pd.read_csv(folder_path+"/"+file, parse_dates=['Time'],
|
||||
index_col='Time')
|
||||
folder_df = add_data_to_df(folder_df, df, file)
|
||||
|
||||
folder_df["class"] = folder_path.split("/")[-2]
|
||||
main_df = pd.concat([main_df, folder_df])
|
||||
|
||||
main_df = pd.melt(main_df, id_vars=["class"])
|
||||
|
||||
box_plot = sns.boxplot(data=main_df, x="variable", y="value", hue="class", ax=axs[k])
|
||||
box_plot.set(xlabel='Payload size (KB)', ylabel=f"Kbytes")
|
||||
box_plot.tick_params(labelbottom=True)
|
||||
box_plot.xaxis.set_tick_params(rotation=45)
|
||||
|
||||
# formatter = ticker.FuncFormatter(lambda x, pos: '{:.0f}'.format(x / 1000))
|
||||
# Apply the custom formatter to the x-axis ticks
|
||||
# box_plot.yaxis.set_major_formatter(formatter)
|
||||
|
||||
plt.tight_layout()
|
||||
plt.savefig(f"teestt.png")
|
||||
plt.show()
|
||||
|
||||
|
||||
create_plot()
|
||||
|
||||
|
||||
def plot_merged(folders_grouped: List, metrics: List):
|
||||
fig, axs = plt.subplots(nrows=3, ncols=2, figsize=(14, 16), sharex=True, sharey='row')
|
||||
|
||||
for j, group in enumerate(folders_grouped):
|
||||
final_df = pd.DataFrame()
|
||||
for i, folder_path in enumerate(group):
|
||||
if not os.path.exists(folder_path): continue
|
||||
files = [f for f in os.listdir(folder_path) if
|
||||
os.path.isfile(os.path.join(folder_path, f))]
|
||||
files = sorted(files, key=lambda x: float(x.split("-")[1].split("KB")[0]))
|
||||
folder_df = pd.DataFrame()
|
||||
|
||||
for file in files:
|
||||
df = pd.read_csv(folder_path + file, parse_dates=['Time'], index_col='Time')
|
||||
|
||||
column_name = file.split("-")[1]
|
||||
|
||||
df_avg = df.mean()
|
||||
|
||||
folder_df = pd.concat([folder_df, df_avg.rename(column_name)], axis=1)
|
||||
folder_df["node"] = language[i]
|
||||
final_df = pd.concat([final_df, folder_df])
|
||||
|
||||
final_df = pd.melt(final_df, id_vars=["node"])
|
||||
|
||||
box_plot = sns.boxplot(data=final_df, x="variable", y="value", hue="node",
|
||||
ax=axs[j // 2, j % 2])
|
||||
box_plot.set_title(f'{data_to_plot[j]} (N=300)')
|
||||
|
||||
box_plot.set(xlabel='Payload size (KB)', ylabel=f"{y_label[j]}")
|
||||
box_plot.tick_params(labelbottom=True)
|
||||
# plt.ylabel(f"{y_label[j]}")
|
||||
# plt.xlabel('Payload size (KB)')
|
||||
|
||||
# sns.move_legend(box_plot, "upper left", bbox_to_anchor=(1, 1))
|
||||
# plt.tight_layout()
|
||||
|
||||
if scale[j]:
|
||||
# Create a custom formatter to divide x-axis ticks by 1000
|
||||
formatter = ticker.FuncFormatter(lambda x, pos: '{:.0f}'.format(x / 1000))
|
||||
# Apply the custom formatter to the x-axis ticks
|
||||
box_plot.yaxis.set_major_formatter(formatter)
|
||||
|
||||
plt.tight_layout()
|
||||
plt.savefig(f"all.png")
|
||||
plt.show()
|
||||
# box_plot.figure.savefig(f"{data_to_plot[j]}-melted.png")
|
||||
|
||||
Loading…
x
Reference in New Issue
Block a user