app.py

import gradio as gr
import pandas as pd
from prophet import Prophet
import plotly.graph_objs as go
import re
import logging
import os
import torch
from chronos import ChronosPipeline
import numpy as np
import requests
import tempfile

try:
    from google.colab import userdata
    PG_PASSWORD = userdata.get('FASHION_PG_PASS')
    CH_PASSWORD = userdata.get('FASHION_CH_PASS')
except:
    PG_PASSWORD = os.environ['FASHION_PG_PASS']
    CH_PASSWORD = os.environ['FASHION_CH_PASS']

logging.getLogger("prophet").setLevel(logging.WARNING)
logging.getLogger("cmdstanpy").setLevel(logging.WARNING)

# Dictionary to map Russian month names to month numbers
russian_months = {
    "январь": "01", "февраль": "02", "март": "03", "апрель": "04",
    "май": "05", "июнь": "06", "июль": "07", "август": "08",
    "сентябрь": "09", "октябрь": "10", "ноябрь": "11", "декабрь": "12"
}

def read_and_process_file(file):
    # Read the first three lines as a single text string
    with open(file.name, 'r') as f:
        first_three_lines = ''.join([next(f) for _ in range(3)])

    # Check for "Неделя" or "Week" (case-insensitive)
    if not any(word in first_three_lines.lower() for word in ["неделя", "week"]):
        period_type = "Month"
    else:
        period_type = "Week"

    # Read the file again to process it
    with open(file.name, 'r') as f:
        lines = f.readlines()

    # Check if the second line is empty
    if lines[1].strip() == '':
        source = 'Google'
        data = pd.read_csv(file.name, skiprows=2)
        # Replace any occurrences of "<1" with 0
    else:
        source = 'Yandex'
        data = pd.read_csv(file.name, sep=';', skiprows=0, usecols=[0, 2])
        if period_type == "Month":
            # Replace Russian months with yyyy-MM format
            data.iloc[:, 0] = data.iloc[:, 0].apply(lambda x: re.sub(r'(\w+)\s(\d{4})', lambda m: f'{m.group(2)}-{russian_months[m.group(1).lower()]}', x) + '-01')
        if period_type == "Week":
            data.iloc[:, 0] = pd.to_datetime(data.iloc[:, 0], format="%d.%m.%Y")
        # Replace any occurrences of "<1" with 0
    data.iloc[:, 1] = data.iloc[:, 1].apply(str).str.replace('<1', '0').str.replace(' ', '').str.replace(',', '.').astype(float)

    # Process the date column and set it as the index
    period_col = data.columns[0]
    data[period_col] = pd.to_datetime(data[period_col])
    data.set_index(period_col, inplace=True)

    return data, period_type, period_col

def get_data_from_db(query):
    # conn = psycopg2.connect(
    #     dbname="kroyscappingdb",
    #     user="read_only",
    #     password=PG_PASSWORD,
    #     host="rc1d-vbh2dw5ha0gpsazk.mdb.yandexcloud.net",
    #     port="6432",
    #     sslmode="require"
    # )
    cert_data = requests.get('https://storage.yandexcloud.net/cloud-certs/RootCA.pem').text

    with tempfile.NamedTemporaryFile(delete=False) as temp_cert_file:
        temp_cert_file.write(cert_data.encode())
        cert_file_path = temp_cert_file.name

    client = Client(host='rc1d-a93v7vf0pjfr6e2o.mdb.yandexcloud.net',
                    port = 9440,
                    user='user1',
                    password=CH_PASSWORD,
                    database='db1',
                    secure=True,
                    ca_certs=cert_file_path)

    # data = pd.read_sql_query(query, conn)
    result, columns = client.execute(query, with_column_types=True)
    column_names = [col[0] for col in columns]
    data = pd.DataFrame(result, columns=column_names)
    # conn.close()
    return data

def forecast_time_series(file, product_name, wb, ozon, model_choice):
    if file is None:
        # Construct the query
        marketplaces = []
        if wb:
            marketplaces.append('wildberries')
        if ozon:
            marketplaces.append('ozon')
        mp_filter = "', '".join(marketplaces)
        # query = f"""
        # select
        #     to_char(dm.end_date, 'yyyy-mm-dd') as ds,
        #     1.0*sum(turnover) / (max(sum(turnover)) over ()) as y
        # from v_datamart dm
        # where {product_name}
        #   and mp in ('{mp_filter}')
        # group by ds
        # order by ds
        # """
        query = f"""
        select
            cast(start_date as date) as ds,
            1.0*sum(turnover) / (max(sum(turnover)) over ()) as y
        from datamart_all_1
            join week_data 
            using (id_week)
        where {product_name}
          and mp in ('{mp_filter}')
        group by ds
        order by ds
        """
        print(query)
        data = get_data_from_db(query)
        print(data)
        period_type = "Week"
        period_col = "ds"

        if len(data)==0:
            raise gr.Error("No data found in database. Please adjust filters")

        data.iloc[:, 0] = pd.to_datetime(data.iloc[:, 0], format='%Y-%m-%d')
        data.set_index('ds', inplace=True)
    else:
        data, period_type, period_col = read_and_process_file(file)

    if period_type == "Month":
        year = 12
        n_periods = 24
        freq = "MS"
    else:
        year = 52
        n_periods = year * 2
        freq = "W"

    df = data.reset_index().rename(columns={period_col: 'ds', data.columns[0]: 'y'})

    if model_choice == "Prophet":
        forecast, yoy_change = forecast_prophet(df, n_periods, freq, year)
    elif model_choice == "Chronos":
        forecast, yoy_change = forecast_chronos(df, n_periods, freq, year)
    else:
        raise ValueError("Invalid model choice")

    # Create Plotly figure (common for both models)
    fig = create_plot(data, forecast)

    # Combine original data and forecast
    combined_df = pd.concat([data, forecast.set_index('ds')], axis=1)

    # Save combined data
    combined_file = 'combined_data.csv'
    combined_df.to_csv(combined_file)

    return fig, f'Year-over-Year Change in Sum of Values: {yoy_change:.2%}', combined_file

def forecast_prophet(df, n_periods, freq, year):
    model = Prophet()
    model.fit(df)
    future = model.make_future_dataframe(periods=n_periods, freq=freq)
    forecast = model.predict(future)
    
    sum_last_year_original = df['y'].iloc[-year:].sum()
    sum_first_year_forecast = forecast['yhat'].iloc[-n_periods:-n_periods + year].sum()
    yoy_change = (sum_first_year_forecast - sum_last_year_original) / sum_last_year_original
    
    return forecast, yoy_change

def forecast_chronos(df, n_periods, freq, year):
    pipeline = ChronosPipeline.from_pretrained(
        "amazon/chronos-t5-mini",
        device_map="cpu",
        torch_dtype=torch.bfloat16,
    )
    
    # Check for non-numeric values
    if not pd.api.types.is_numeric_dtype(df['y']):
        non_numeric = df[pd.to_numeric(df['y'], errors='coerce').isna()]
        if not non_numeric.empty:
            error_message = f"Non-numeric values found in 'y' column. First few problematic rows:\n{non_numeric.head().to_string()}"
            raise ValueError(error_message)
    
    try:
        y_values = df['y'].values.astype(np.float32)
    except ValueError as e:
        raise ValueError(f"Unable to convert 'y' column to float32: {str(e)}")

    chronos_forecast = pipeline.predict(
        context=torch.tensor(y_values),
        prediction_length=n_periods,
        num_samples=20,
        limit_prediction_length=False
    )
    
    forecast_index = pd.date_range(start=df['ds'].iloc[-1], periods=n_periods+1, freq=freq)[1:]
    low, median, high = np.quantile(chronos_forecast[0].numpy(), [0.1, 0.5, 0.9], axis=0)
    
    forecast = pd.DataFrame({
        'ds': forecast_index,
        'yhat': median,
        'yhat_lower': low,
        'yhat_upper': high
    })
    
    sum_last_year_original = df['y'].iloc[-year:].sum()
    sum_first_year_forecast = median[:year].sum()
    yoy_change = (sum_first_year_forecast - sum_last_year_original) / sum_last_year_original
    
    return forecast, yoy_change

def create_plot(data, forecast):
    fig = go.Figure()
    fig.add_trace(go.Scatter(x=data.index, y=data.iloc[:, 0], mode='lines', name='Observed'))
    fig.add_trace(go.Scatter(x=forecast['ds'], y=forecast['yhat'], mode='lines', name='Forecast', line=dict(color='red')))
    fig.add_trace(go.Scatter(x=forecast['ds'], y=forecast['yhat_lower'], fill=None, mode='lines', line=dict(color='pink'), name='Lower CI'))
    fig.add_trace(go.Scatter(x=forecast['ds'], y=forecast['yhat_upper'], fill='tonexty', mode='lines', line=dict(color='pink'), name='Upper CI'))
    
    fig.update_layout(
        title='Observed Time Series and Forecast with Confidence Intervals',
        xaxis_title='Date',
        yaxis_title='Values',
        legend=dict(orientation='h', yanchor='bottom', y=1.02, xanchor='right', x=1),
        hovermode='x unified'
    )
    
    return fig

# Create Gradio interface using Blocks
with gr.Blocks(theme=gr.themes.Monochrome()) as interface:
    gr.Markdown("# Time Series Forecasting")
    gr.Markdown("Upload a CSV file with a time series to forecast the next 2 years and see the YoY % change. Download the combined original and forecast data.")

    with gr.Row():
        file_input = gr.File(label="Upload Time Series CSV")

    with gr.Row():
        wb_checkbox = gr.Checkbox(label="Wildberries", value=True)
        ozon_checkbox = gr.Checkbox(label="Ozon", value=True)

    with gr.Row():
        product_name_input = gr.Textbox(label="Product Name Filter", value="name like '%пуховик%'")

    with gr.Row():
        model_choice = gr.Radio(["Prophet", "Chronos"], label="Choose Model", value="Prophet")

    with gr.Row():
        compute_button = gr.Button("Compute")

    with gr.Row():
        plot_output = gr.Plot(label="Time Series + Forecast Chart")

    with gr.Row():
        yoy_output = gr.Text(label="YoY % Change")

    with gr.Row():
        csv_output = gr.File(label="Download Combined Data CSV")

    compute_button.click(
        forecast_time_series,
        inputs=[file_input, product_name_input, wb_checkbox, ozon_checkbox, model_choice],
        outputs=[plot_output, yoy_output, csv_output]
    )

# Launch the interface
interface.launch(debug=True)