Skip to content
Snippets Groups Projects
Commit 01ebe6fb authored by Danilo Ferreira de Lima's avatar Danilo Ferreira de Lima
Browse files

Removed plotting from offline analysis and moved all plots to prepare_plots.

parent 4be78de5
No related branches found
No related tags found
1 merge request!14Corrected bugs in the BNN and added many plotting scripts adapted for the paper
Expand all Hide whitespace changes
Inline Side-by-side
pes_to_spec/test/offline_analysis.py
+ 79
− 393
View file @ 01ebe6fb
This diff is collapsed.
pes_to_spec/test/prepare_plots.py
+ 151
− 23
View file @ 01ebe6fb
#!/usr/bin/env python
import os
import re
import matplotlib
matplotlib.use('Agg')
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
from matplotlib.gridspec import GridSpec
import seaborn as sns
from scipy.interpolate import make_interp_spline, BSpline
SMALL_SIZE = 12
MEDIUM_SIZE = 18
BIGGER_SIZE = 22
BIGGER_SIZE = 24
plt.rc('font', size=BIGGER_SIZE) # controls default text sizes
plt.rc('axes', titlesize=BIGGER_SIZE) # fontsize of the axes title
......@@ -24,8 +30,8 @@ def plot_final(df: pd.DataFrame, filename: str):
ax = fig.add_subplot(gs[0, 0])
ax.plot(df.energy, df.spec, c='b', lw=3, label="Grating spectrometer")
ax.plot(df.energy, df.prediction, c='r', ls='--', lw=3, label="Prediction")
ax.fill_between(df.energy, df.prediction - df.unc, df.prediction + df.unc, facecolor='gold', alpha=0.5, label="68% unc. (total)")
ax.fill_between(df.energy, df.prediction - df.unc_pca, df.prediction + df.unc_pca, facecolor='magenta', alpha=0.5, label="68% unc. (PCA only)")
ax.fill_between(df.energy, df.prediction - 2*df.unc, df.prediction + 2*df.unc, facecolor='gold', alpha=0.5, label="95% unc. (total)")
ax.fill_between(df.energy, df.prediction - 2*df.unc_pca, df.prediction + 2*df.unc_pca, facecolor='magenta', alpha=0.5, label="95% unc. (PCA only)")
Y = np.amax(df.spec)
ax.legend(frameon=False, borderaxespad=0, loc='upper left')
ax.set_title(f"Beam intensity: {df.beam_intensity.iloc[0]:.1f} mJ", loc="left")
......@@ -35,6 +41,7 @@ def plot_final(df: pd.DataFrame, filename: str):
xlabel="Photon energy [eV]",
ylabel="Intensity [a.u.]",
ylim=(0, 1.3*Y))
plt.tight_layout()
fig.savefig(filename)
plt.close(fig)
......@@ -76,22 +83,52 @@ def plot_rmse_intensity(df: pd.DataFrame, filename: str):
fig.savefig(filename)
plt.close(fig)
def plot_residue(df: pd.DataFrame, filename: str):
cols = [k for k in df.columns if "res_prepca" in k]
df_res = df.loc[:, cols]
n_plots = len(df_res.columns)//10
fig = plt.figure(figsize=(8*n_plots, 8))
gs = GridSpec(1, n_plots)
for i_plot in range(n_plots):
ax = fig.add_subplot(gs[0, i_plot])
sns.kdeplot(data={f"Dim. {k+1}": df_res.loc[:, cols[k]] for k in range(i_plot*10, i_plot*10 + 10)},
linewidth=3, ax=ax)
ax.set(title=f"",
xlabel=r"residue/uncertainty [a.u.]",
ylabel="Counts [a.u.]",
xlim=(-3, 3),
)
ax.legend(frameon=False)
fig.savefig(filename)
plt.close(fig)
def plot_chi2_intensity(df: pd.DataFrame, filename: str):
fig = plt.figure(figsize=(12, 8))
gs = GridSpec(1, 1)
ax = fig.add_subplot(gs[0, 0])
ax.scatter(df.chi2_prepca/df.ndof.iloc[0], df.xgm_flux_t, c='r', s=30)
sns.kdeplot(x=df.chi2_prepca/df.ndof.iloc[0], y=df.xgm_flux_t*1e-3,
fill=True,
ax=ax)
sns.scatterplot(x=df.chi2_prepca/df.ndof.iloc[0], y=df.xgm_flux_t*1e-3,
s=200,
alpha=0.1,
#size=df.root_mean_squared_pca_unc,
#sizes=(20, 200),
ax=ax)
ax = plt.gca()
ax.set(title=f"",
xlabel=r"$\chi^2/$ndof",
ylabel="Beam intensity [uJ]",
ylabel="Beam intensity [mJ]",
xlim=(0, 5),
ylim=(0, df.xgm_flux_t.mean() + 3*df.xgm_flux_t.std())
ylim=(0, df.xgm_flux_t.mean()*1e-3 + 3*df.xgm_flux_t.std()*1e-3)
)
ax.spines['top'].set_visible(False)
ax.spines['right'].set_visible(False)
plt.tight_layout()
fig.savefig(filename)
plt.close(fig)
def pca_variance_plot(df: pd.DataFrame, filename: str):
def pca_variance_plot(df: pd.DataFrame, filename: str, max_comp_frac: float=0.99):
"""
Plot variance contribution.
......@@ -100,25 +137,88 @@ def pca_variance_plot(df: pd.DataFrame, filename: str):
variance_ratio: Contribution of each component's variance.
"""
fig = plt.figure(figsize=(8, 8))
fig = plt.figure(figsize=(12, 8))
gs = GridSpec(1, 1)
ax = fig.add_subplot(gs[0, 0])
c = np.cumsum(df.variance_ratio)
n_comp = df.n_comp.iloc[0]
n_comp = int(df.n_comp.iloc[0])
ax.bar(1+np.arange(len(df.variance_ratio)), df.variance_ratio*100, color='tab:red', alpha=0.3, label="Per component")
ax.plot(1+np.arange(len(df.variance_ratio)), c*100, c='tab:blue', lw=5, label="Cumulative")
ax.plot([n_comp, n_comp], [0, c[n_comp]*100], lw=3, ls='--', c='m', label="Components kept")
ax.plot([0, n_comp], [c[n_comp]*100, c[n_comp]*100], lw=3, ls='--', c='m')
ax.legend(frameon=False)
print(f"PCA plot: total n. components: {len(df.variance_ratio)}")
x_max = np.where(c > 0.99)[0][0]
x_max = np.where(c > max_comp_frac)[0][0]
print(f"Fraction of variance: {c[n_comp]}")
ax.set_yscale('log')
ax.set(title=f"",
xlabel="Component",
ylabel="Variance [%]",
xlim=(1, x_max),
ylim=(0.1, 100))
ylim=(0.01, 100))
ax.spines['top'].set_visible(False)
ax.spines['right'].set_visible(False)
plt.tight_layout()
fig.savefig(filename)
plt.close(fig)
def moving_average(a, n=3):
ret = np.cumsum(a)
ret[n:] = ret[n:] - ret[:-n]
return ret[n - 1:] / n
def plot_impulse(df: pd.DataFrame, filename: str):
"""
Plot variance contribution.
Args:
filename: Output file name.
variance_ratio: Contribution of each component's variance.
"""
fig = plt.figure(figsize=(12, 8))
gs = GridSpec(1, 1)
ax = fig.add_subplot(gs[0, 0])
x = df.wiener_energy.to_numpy()
y = np.absolute(df.impulse.to_numpy())
#x_new = np.linspace(-6, 6, 601)
#spl = make_interp_spline(x, np.log10(y), k=3)
#y_new = np.power(10, spl(x_new))
x_new = moving_average(x, n=10)
y_new = moving_average(y, n=10)
sel = (x_new >= -10) & (x_new <= 10)
ax.plot(x_new[sel], y_new[sel], c='tab:blue', lw=4)
ax.set_yscale('log')
ax.set(title=f"",
xlabel="Energy [eV]",
ylim=(1e-4, 0.5),
ylabel="Response [a.u.]",
)
ax.spines['top'].set_visible(False)
ax.spines['right'].set_visible(False)
plt.tight_layout()
fig.savefig(filename)
plt.close(fig)
def plot_wiener(df: pd.DataFrame, filename: str):
"""
Plot variance contribution.
Args:
filename: Output file name.
variance_ratio: Contribution of each component's variance.
"""
fig = plt.figure(figsize=(12, 8))
gs = GridSpec(1, 1)
ax = fig.add_subplot(gs[0, 0])
ax.plot(df.wiener_energy, np.absolute(df.wiener_filter), c='tab:blue', lw=3)
ax.set_yscale('log')
ax.set(title=f"",
xlabel="Energy [eV]",
ylim=(1e-3, 1),
ylabel="Response [a.u.]",
)
ax.spines['top'].set_visible(False)
ax.spines['right'].set_visible(False)
plt.tight_layout()
......@@ -134,36 +234,64 @@ def plot_pes(df: pd.DataFrame, channel:str, filename: str):
pes_raw_int: Low-resolution spectrum.
"""
fig = plt.figure(figsize=(16, 8))
fig = plt.figure(figsize=(12, 8))
gs = GridSpec(1, 1)
ax = fig.add_subplot(gs[0, 0])
first, last = df.loc[:, 'first'].iloc[0], df.loc[:, 'last'].iloc[0]
ax.plot(df.loc[(df.bin >= first) & (df.bin < last), "bin"], df.loc[(df.bin >= first) & (df.bin < last), channel], c='b', lw=3)
#ax.legend()
first = first+220
last = last-270
print("Range:", first, last)
sel = (df.bin >= first) & (df.bin < last)
x = df.loc[sel, "bin"]
if channel == "sum":
y = df.loc[sel, [k for k in df.columns if "channel_" in k]].sum(axis=1)
ax.plot(x, y, c='b', lw=5)
elif isinstance(channel, list):
for ch in channel:
sch = ch.replace('_', ' ')
y = df.loc[sel, ch]
ax.plot(x, y, lw=5, label=sch)
else:
y = df.loc[sel, channel]
ax.plot(x, y, c='b', lw=5)
ax.legend(frameon=False)
ax.set(title=f"",
xlabel="Time-of-flight index",
ylabel="Counts [a.u.]")
ax.spines['top'].set_visible(False)
ax.spines['right'].set_visible(False)
plt.tight_layout()
fig.savefig(filename)
plt.close(fig)
if __name__ == '__main__':
indir = 'p900331r69t70'
channel = 'channel_4_A'
fname = 'test_q100_1724098413'
plot_final(pd.read_csv(f'{indir}/{fname}.csv'), f'{fname}.pdf')
plot_pes(pd.read_csv(f'{indir}/{fname}_pes.csv'), channel, f'{fname}_{channel}.pdf')
channel = ['channel_1_A', 'channel_4_A', 'channel_3_B']
#channel = 'sum'
#for fname in os.listdir(indir):
# if re.match(r'test_q100_[0-9]*\.csv', fname):
# fname = fname[:-4]
# print(f"Plotting {fname}")
# plot_final(pd.read_csv(f'{indir}/{fname}.csv'), f'{fname}.pdf')
# plot_pes(pd.read_csv(f'{indir}/{fname}_pes.csv'), channel, f'{fname}_pes.pdf')
fname = 'test_q100_1724098596'
plot_final(pd.read_csv(f'{indir}/{fname}.csv'), f'{fname}.pdf')
plot_pes(pd.read_csv(f'{indir}/{fname}_pes.csv'), channel, f'{fname}_{channel}.pdf')
for fname in ('test_q100_1724098413', 'test_q100_1724098596', 'test_q50_1724099445'):
plot_final(pd.read_csv(f'{indir}/{fname}.csv'), f'{fname}.pdf')
plot_pes(pd.read_csv(f'{indir}/{fname}_pes.csv'), channel, f'{fname}_pes.pdf')
plot_chi2(pd.read_csv(f'{indir}/quality.csv'), f'chi2_prepca.pdf')
plot_chi2_intensity(pd.read_csv(f'{indir}/quality.csv'), f'intensity_vs_chi2_prepca.pdf')
plot_rmse(pd.read_csv(f'{indir}/quality.csv'), f'rmse.pdf')
plot_rmse_intensity(pd.read_csv(f'{indir}/quality.csv'), f'intensity_vs_rmse.pdf')
pca_variance_plot(pd.read_csv(f'{indir}/pca_spec.csv'), f'pca_spec.pdf')
pca_variance_plot(pd.read_csv(f'{indir}/pca_pes.csv'), f'pca_pes.pdf')
plot_residue(pd.read_csv(f'{indir}/quality.csv'), f'residue.pdf')
df_model = pd.read_csv(f'{indir}/model.csv')
df_model.impulse = df_model.impulse.str.replace('i','j').apply(lambda x: np.complex(x))
df_model.wiener_filter = df_model.wiener_filter.str.replace('i','j').apply(lambda x: np.complex(x))
plot_impulse(df_model, f'impulse.pdf')
plot_wiener(df_model, f'wiener.pdf')
pca_variance_plot(pd.read_csv(f'{indir}/pca_spec.csv'), f'pca_spec.pdf', max_comp_frac=0.99)
pca_variance_plot(pd.read_csv(f'{indir}/pca_pes.csv'), f'pca_pes.pdf', max_comp_frac=0.95)
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment

Legal Notice | Contacts