eda.py

import streamlit as st
import os
from PIL import Image
import pandas as pd
import numpy as np


# Fill in the labels & img_list lists
@st.experimental_memo
def read_dataset(data_path):
    img_list = []
    labels = []
    for class_name in os.listdir(data_path):
        if class_name not in labels:
            labels.append(class_name)
        img_dir = data_path + class_name + "/"
        for img_filename in os.listdir(img_dir):
            img_path = img_dir + img_filename
            img_list.append([img_path, class_name])
    return img_list, labels


# Filter on the label part of the sublist (eg: [[img_path, label], ...]
@st.experimental_memo
def get_filtered_list(filter: str, list: list):
    return [x[0] for x in list if x[1] == filter]


# Calculate the average resolution for a given list of images
@st.experimental_memo
def get_average_img_resolution(images: list):
    widths = []
    heights = []

    for img in images:
        im = Image.open(img)
        widths.append(im.size[0])
        heights.append(im.size[1])

    avg_width = round(sum(widths) / len(widths))
    avg_height = round(sum(heights) / len(heights))
    return [avg_width, avg_height]


def eda():
    st.header('Thomas More - Big Data Team Assigment: Image Classification - EDA')

    data_path = 'data/'
    img_list, labels = read_dataset(data_path)
    total_imgs_overview = []

    # --- Tabs ---
    overview_tab, category_tab = st.tabs(["Overview", "Category details"])
    with category_tab:
        # --- Dataset category Tabs ---
        tab1, tab2, tab3, tab4, tab5 = st.tabs(labels)

        # Generate a tab for each of the labels
        for index, tab in enumerate([tab1, tab2, tab3, tab4, tab5]):
            with tab:
                images = get_filtered_list(labels[index], img_list)
                total_imgs = len(images)
                total_imgs_overview.append(total_imgs)
                avg_w, avg_h = get_average_img_resolution(images)
                st.header(labels[index])
                st.write(f'There are a total of {total_imgs} samples.')
                st.write(f'The average resolution is {avg_w}x{avg_h}')
                to_show = st.slider('How many samples to show? (A large number can take a while to load)', 0, total_imgs,
                                    30)
                st.image(images[:to_show], width=200)

    with overview_tab:
        st.header("Overview")
        avg_w_overview, avg_h_overview = get_average_img_resolution([x[0] for x in img_list])
        st.write(f'There are a total of {sum(total_imgs_overview)} samples.')
        st.write(f'The average resolution is {avg_w_overview}x{avg_h_overview}')

        chart_data = pd.DataFrame({'# of samples': total_imgs_overview}, index=labels)
        st.bar_chart(chart_data)