app.py

from wordcloud import WordCloud, STOPWORDS, ImageColorGenerator  # Generate word clouds
from PIL import Image  # Load images
import pandas as pd 
import numpy as np  # Convert images to numbers
import re  # clean the user_text_put and remove none words char
import streamlit as st
import matplotlib.pyplot as plt
import plotly.express as px
from collections import Counter  # Count the frequency of distinct strings
from gensim.parsing.preprocessing import remove_stopwords

# Streamline syntax ro create title, subtitle and text_input
st.sidebar.image('logo1.png', width=300)
# Add the expander to provide some information about the app
with st.sidebar.expander("About the App"):
    st.write("""
        Word clouds play a major role in demonstrating what the text has to say. This application will help anyone to generate a word cloud of set dimensions and resolution.""")
st.title("Wordcloud Generator")  # set the title

text_input = st.text_input(" ").lower()  # get user input as string and turn it all lowercase

# create text input field and display it on sidebar
background_color = st.sidebar.text_input('Choose background color', value="white")

# Create number input field and display it on sidebar
height = int(st.sidebar.number_input('Choose height between 500 - 1000 pixels',
                                     value=640))
width = int(st.sidebar.number_input('Choose width between 500 - 1000 pixels',
                                    value=800))


# st.markdown(text_input)

# Create a function to remove none words in the job description
@st.cache
def text():
    pattern = r'\W+'  # return only words with in text
    # Split the text into a list of individual words
    word = re.split(pattern, text_input)
    return word


# join text with space
words = " ".join(text())
words_NST = remove_stopwords(words)

im_shape = st.sidebar.file_uploader("Upload your shape as an jpg, image must have white backround (optional).", type='png')
# Create a wordcloud generator
try:
    if im_shape is not None:
        image = Image.open(im_shape)  # Load the image from a file

        mask = np.array(image)  # Convert the image to a numeric representation

        # Create a wordcloud generator using Heart shape
        wordcloud = WordCloud(mask=mask, background_color=background_color).generate(words)
        fig, ax = plt.subplots()
        ax.imshow(wordcloud)
        plt.xticks([])  # hide y tick marks from the graph
        plt.yticks([])  # hide y tick marks from the graph
        # plt.axis('off')
        st.pyplot(fig)
    else:
        # Create a wordcloud generator with defaults default arguments
        wordcloud = WordCloud(stopwords=STOPWORDS, background_color=background_color,
                              height=height, width=width).generate(words)
        fig, ax = plt.subplots()
        ax.imshow(wordcloud)
        plt.xticks([])
        plt.yticks([])
        # plt.axis('off')
        st.pyplot(fig)
except ValueError:
    st.markdown("##### Type your text in the box above & hit Enter!")  # set the sub header

check = st.checkbox('Graph Bar Chart')

if check and text_input:
    list_of_word = list((words_NST.split(" ")))
    #st.write(list_of_word)
    list_of_words = [word for word in list_of_word]
   # st.write(list_of_word)
    count = Counter(list_of_words)

    #st.write(count)

    list_of_words_df = pd.DataFrame(count.most_common(100), columns=["words", "word_count"])
    # st.write(list_of_words_df)
    # Plot horizontal bar graph
    list_of_words_df.sort_values(by='word_count')
    fig = px.bar(list_of_words_df, x="word_count", y="words",
        labels={'word_count':'Word Frequences'}, height=640,  width=1000)
    st.plotly_chart(fig)