app.py

from datetime import datetime
import numpy as np
import pandas as pd
from sklearn.ensemble import RandomForestRegressor
import gradio as gr
import plotly.graph_objects as go
from huggingface_hub import from_pretrained_keras
import os


def predictAirPassengers(df, split):
    ts= pd.read_csv('DemandaQuito.csv')
    df2 =ts.copy()
    ttSplit=split/100
    ts['Month']=pd.to_datetime(ts['Month'])
    ts.rename(columns={'#Valor':'Valor'},inplace=True)
    ts=ts.set_index(['Month'])
    ts['months'] = [x.month for x in ts.index]
    ts['years'] = [x.year for x in ts.index]
    ts.reset_index(drop=True, inplace=True)

    # Split Data
    X=ts.drop("Valor",axis=1)
    Y= ts["Valor"]
    X_train=X[:int (len(Y)*ttSplit)] 
    X_test=X[int(len(Y)*ttSplit):]
    Y_train=Y[:int (len(Y)*ttSplit)] 
    Y_test=Y[int(len(Y)*ttSplit):]

    # fit the model
    rf = RandomForestRegressor()
    rf.fit(X_train, Y_train)
    
    df1=df2.set_index(['Month'])
    df1.rename(columns={'#Valor':'Valor'},inplace=True)
    train=df1.Valor[:int (len(ts.Valor)*ttSplit)]
    test=df1.Valor[int(len(ts.Valor)*ttSplit):]
    preds=rf.predict(X_test).astype(int) 
    predictions=pd.DataFrame(preds,columns=['Valor'])
    predictions.index=test.index
    predictions.reset_index(inplace=True)
    predictions['Month']=pd.to_datetime(predictions['Month'])
    print(predictions)
    
    #combine all into one table
    ts_df=df.copy()
    ts_df.rename(columns={'#Valor':'Valor'},inplace=True)
    train= ts_df[:int (len(ts_df)*ttSplit)]
    test= ts_df[int(len(ts_df)*ttSplit):] 

    df2['Month']=pd.to_datetime(df2['Month'])
    df2.rename(columns={'#Valor':'Valor'},inplace=True)
    df3= predictions
    df2['origin']='ground truth'
    df3['origin']='prediction'
    df4=pd.concat([df2, df3])
    print(df4)
    return df4

demo = gr.Interface(
    fn =predictAirPassengers,
    inputs = [
        gr.Timeseries(label="Input for the timeseries", max_rows=1, interactive=False),
        gr.Slider(1, 100, value=75, step=1, label="Train test split percentage"),
    ],
    outputs= [
       gr.LinePlot(x='Month', y='Valor', color='origin')
       #gr.Timeseries(x='Month')

    ],
    examples=[
        [os.path.join(os.path.abspath(''), "DemandaQuito_dt.csv"), 75],
    ]
)

demo.launch()