import streamlit as st import matplotlib from functions import * from dataset import get_dataset, get_columns, get_optional_params from format import validate_and_formate_columns, filter_selected_options, valid_rows_number from data_prep import get_frequency, tmp_renaming_columns, get_dates, split_dataset, get_forecast_dates from forecast import forecast_models from visualise import DataVisualizer from theme import set_theme from export import Exporter font = {'weight' : 'normal', 'size' : 18} matplotlib.rc('font', **font) def main(): set_theme() st.title("FoodQast") st.sidebar.title("1. Data") # Load data from uploaded dataset with st.sidebar.expander("Dataset", expanded=True): files, datasets, load_options = get_dataset() check_upload_status(files["file"], files["recipe_file"]) # sidebar setting if files are uploaded if files["file"] and files["recipe_file"]: with st.sidebar.expander("Columns", expanded=True): columns = get_columns(datasets, load_options) validate_and_formate_columns(datasets, columns, load_options) with st.sidebar.expander("Optional columns"): datasets, columns, load_options = get_optional_params(datasets, columns, load_options) datasets = filter_selected_options(datasets, columns, load_options) with st.sidebar.expander("Frequency"): load_options = get_frequency(datasets, columns, load_options) tmp_renaming_columns(datasets, columns, load_options) datasets["food_df"] = frequency_grouping(datasets["data"], load_options["freq"]) st.sidebar.title("2. Modelling") with st.sidebar.expander("Algorithm"): load_options["algorithm"] = st.selectbox("Select an algorithm:", ["Algorithm 1", "Algorithm 2", "Algorithm 3"]) st.sidebar.title("3. Forecast") load_options["evaluate"] = st.sidebar.checkbox("Evaluate model", value=True) if load_options["evaluate"]: with st.sidebar.expander("Split dates", expanded=True): dates = get_dates(datasets["food_df"]) datasets["train_df"], datasets["test_df"] = split_dataset(datasets["food_df"], dates) load_options["forecast"] = st.sidebar.checkbox("Forecast model") if load_options["forecast"]: with st.sidebar.expander("Duration", expanded= True): datasets["forecast_df"], load_options = get_forecast_dates(datasets["food_df"], load_options) # check enough rows valid_rows_number(datasets, load_options) st.sidebar.button("Run", on_click=click_button) # run forecast if button is clicked if st.session_state.clicked: results = forecast_models(datasets, load_options) visualiser = DataVisualizer(datasets, load_options) #visualizing the dataset st.header("1. Visualizing data") visualiser.visualise_data() #training the data if load_options["evaluate"]: st.header("2. Evaluation") st.subheader("Overview") visualiser.visualise_evaluation(results) st.subheader("Global performance") visualiser.display_evaluation(results) if load_options["forecast"]: st.header("3. Forecast" if load_options['evaluate'] else "2. Forecast") visualiser.visualise_forecast(results) # Export file as pdf exporter = Exporter(visualiser.plots, visualiser.tables, load_options) exporter.export() if __name__ == '__main__': main()