Nothing is interesting if you're not interested

Helen Clark MacInnes

Python Stuff

I tend to spend too much time mucking around with Python code - as you can see from the bits of code scattered around this site. So, I've decided to add a page specifically for stuff related to just Python. It'll include snippets of code, interesting applications, algorithms or anything else Python related. Code that is purely there as a means to demonstrate or calculate something specific will remain as it is, in the most relevant post. Most of this stuff will be migrating to my sister pages -

- in the near future.

Sorry if you were hoping to see something on the Pythonidae family of snakes - interesting as they are.

Matplotlib - subplots

19.01.2020 Filed in: Python |Matplotlib |Code python, matplotlib, glob, subplots, sub-plots

The code below says it all really. Oh, and it's using Pandas to read in Compton scattering experiment data. The glob thing is a nice way of geting all the CSV files in a directory.

import glob
import math as mt
import numpy as np
import pandas as pd
import matplotlib.pyplot as plt

# Set the number of subplot columns
NUMBER_OF_COLUMNS = 2

# Find all CVS files in this directory
files = glob.glob('ComptonData/*.csv')

# For testing this removes the last file leaving an odd number
files=files[:-1]

# Read them into a list of DataFrames
df_list=[]
for file in files:
    df_ = pd.read_csv(file)
    df_list.append(df_)

# Do the plots with n_cols columns
n_cols = NUMBER_OF_COLUMNS
n_rows = mt.ceil(len(files)/n_cols)

# We'll have a common x scale but differet y scales
fig, axs = plt.subplots(nrows=n_rows, ncols=n_cols, sharex=True, sharey=False, figsize=(10,7.5))

# Iterate through the file list we've just generated and append the sub-plots
# I've chosen to just include the 'interesting' part of the data that includes Ka and Kb peaks.
for k, df in enumerate(df_list): 
    x = df['Channel'][200:350]
    y =  df['Counts'][200:350]
    # Generate some label text for later
    angle = df.at[0,'Value']
    label = 'Scattering angle = '+ str(int(angle))
    
    # These next 2 lines generate row, col pairs from the index, k
    col = k%n_cols
    row = (int(k/n_cols)%max(n_rows,n_cols))
    
    axs[row,col].plot(x,y)
    axs[row,col].grid()
    # Put a label in each subplot near top left
    axs[row,col].text(0.05, 0.85, label, transform=axs[row,col].transAxes, fontsize=7)

# This removes the last 'blank' sub-plots caused by odd number of plots
# AND makes sure you've got an x -axis label on all bottom plots    
if len(df_list)%n_cols != 0: #An 'odd' number of plots 
    blank_plots = n_cols-(len(df_list)%n_cols)
    for i in range(blank_plots):
        axs[n_rows-2,n_cols-(i+1)].xaxis.set_tick_params(which='both', labelbottom=True, labeltop=False)
        # Remove the redundant sub-plot
        fig.delaxes(axs[n_rows-1,n_cols-(i+1)])

    
# Now generate an overall, encompassing 'axes' plot for aggregate labels
ax = fig.add_subplot(111, frameon=False)
# Give this some common x and y labels. Use the argument `labelpad` to move label downwards.
ax.set_xlabel('Channel number', fontsize=14, fontweight='bold', fontstyle='italic', labelpad=25) 
ax.set_ylabel('Counts', fontsize=14, fontweight='bold', labelpad=35)
# But we don't want it to have any axis 'ticks'/values
ax.set_yticklabels([])
ax.set_xticklabels([])

# Show them
plt.tight_layout()
plt.show()