TomlDecodeError: Key name found without value. Reached end of line When deploying on Heroku

Question

Am getting this error message when deploying a streamlit app to Heroku and do not know how to troubleshoot it. I have deployed apps using this setup previously, and removing the config.toml file has always fixed it. But this time is different.

I went to Heroku's website to find help and no luck so far: https://discuss.streamlit.io/t/tomldecodeerror-key-name-found-without-value-reached-end-of-line-when-deploying-on-heroku/12285

It looks like this questions has left others hanging as well, since this question and this one have the same error message and there has not been an answer so far.

Here is my setup.sh file, which I believe is causing the problem:

mkdir -p ~/.streamlit/
echo "\
[general]\n\
email = \"myemailgmail.com\"\n\
" > ~/.streamlit/credentials.toml
echo "\
[server]\n\
headless = true\n\
enableCORS=false\n\
port = $PORT\n\
" > ~/.streamlit/config.toml

Here is my Procfile:

web: sh setup.sh && streamlit run app.py

Here is the application file itself, app.py:

import streamlit as st
import pandas as pd
import pickle
import plotly.express as px

@st.cache
def load_data(n_rows=3000):
    data = pd.read_csv('https://raw.githubusercontent.com/JonathanBechtel/dat-02-22/main/ClassMaterial/Unit3/data/ks2.csv', nrows=n_rows)
    return data

@st.cache
def group_data(x_axis, y_axis):
    result = data.groupby(x_axis)[y_axis].mean()
    return result

@st.cache
def load_model():
    with open('mod.pkl', 'rb') as mod:
        pipe = pickle.load(mod)
        
    return pipe

st.title("Understanding Kickstarter Applications -- See How Easy It Is Cindy?")

section = st.sidebar.radio('Section', ['Data Explorer', 'Model Predictions'])

n_rows = st.sidebar.number_input("Enter Number of Rows To Load", min_value=1000, max_value=100000, step=1000)
data = load_data(n_rows)
if section == 'Data Explorer':
    chart_type = st.sidebar.selectbox('Chart Type', ['Bar', 'Line', 'Strip'])
    
    st.write(data)
    
    x_axis = st.sidebar.selectbox('Choose Column for X-Axis', ['category',  'main_category', 'country'])
    y_axis = st.sidebar.selectbox('Choose Column for y-axis', ['state', 'goal'])
    
    st.header(f"Average value for {y_axis} for column {x_axis}")

    if chart_type == 'Bar':
        result = group_data(x_axis, y_axis)
        st.bar_chart(result)
    elif chart_type == 'Line':
        result = group_data(x_axis, y_axis)
        st.line_chart(result)
    else:
        result = data[[x_axis, y_axis]]
        st.plotly_chart(px.strip(result, x=x_axis, y=y_axis, color=x_axis))
        
elif section == 'Model Predictions':
    with open('mod.pkl', 'rb') as mod:
        pipe = pickle.load(mod)
    print(pipe)
    category       = st.sidebar.selectbox('Select A Category', data['category'].unique().tolist())
    main_category  = st.sidebar.selectbox('Select a Main Category', data['main_category'].unique().tolist())
    funding_amount = st.sidebar.number_input('Enter Your Funding Amount', min_value=0, value=1000, step=500) 
    
    sample = pd.DataFrame({
        'category': [category],
        'main_category': [main_category],
        'funding_amount': [funding_amount]
        })
    
    
    prediction = pipe.predict_proba(sample)
    
    st.header(f"Predicted Probability of Campaign Successs: {prediction[0][1]:.2%}")

My requirements.txt file reads like this:

streamlit == 0.67.0
scikit-learn
pandas
category_encoders == 2.*
xgboost == 1.3.*
protobuf == 3.13.0
plotly == 4.12.0

Any recommendations are welcome, because all wells have been run dry at this point.

Answer 1

You can try :

mkdir -p ~/.streamlit/

echo "[server]\nheadless = true\nport = $PORT\nenableCORS = false\n" > ~/.streamlit/config.toml