justinhchae

# how to pool processes for prophet

# given chunked data from data_chunker.py
# https://gist.github.com/justinhchae/13d246e8e2e2d521a8d2cce20eb09a09

# given prophet function wrapper
# https://gist.github.com/justinhchae/8ef78743f13f50051ad1aca2106eaa1a

# dependencies
from tqdm import tqdm

justinhchae

from fbprophet import Prophet
import pandas as pd
import os

# https://stackoverflow.com/questions/2125702/how-to-suppress-console-output-in-python
# https://medium.com/spikelab/forecasting-multiples-time-series-using-prophet-in-parallel-2515abd1a245
# https://facebook.github.io/prophet/docs/quick_start.html#python-api


class suppress_stdout_stderr(object):

justinhchae

# how to pool processes for arima

# given chunked data from data_chunker.py
# https://gist.github.com/justinhchae/13d246e8e2e2d521a8d2cce20eb09a09

# given arima function from run_arima.py
# https://gist.github.com/justinhchae/d2a2dc8b71b5f5fbbb0f7eabf68b6850

# dependencies
from tqdm import tqdm

justinhchae

def run_arima(chunked_data, price_col='y', n_prediction_units=1):
  # consume chunked data from https://gist.github.com/justinhchae/13d246e8e2e2d521a8d2cce20eb09a09
    # supress trivial warnings from ARIMA
    warnings.simplefilter('ignore', ConvergenceWarning)

    # initialize a list to hold results (a list of dataframes)
    results = []

    # numerate through a list of chunked tuples, each having a pair of dataframes
    for idx, (x_i, y_i) in enumerate(chunked_data):

justinhchae

import numpy as np
import math

# helper functions to chunk data for time series forecasting

def index_marks(nrows, chunk_size):
    """
    a helper function for split()
    return an index of chunk size
    https://yaoyao.codes/pandas/2018/01/23/pandas-split-a-dataframe-into-chunks

justinhchae

import pandas as pd
import streamlit as st

from collections import defaultdict
from streamlit.report_thread import get_report_ctx

# implementation
# https://share.streamlit.io/justinhchae/app_helper/main/app.py

class SomeClass():

justinhchae

# given a dataframe 

def reduce_precision(df):
        import numpy as np
        """
        usage: give a dataframe, this fuction returns an optimized dataframe
        
        df = reduce_precision(df)
        
        reference: https://gist.github.com/enamoria/fa9baa906f23d1636c002e7186516a7b

justinhchae

import pandas as pd

gitcsv = 'https://raw.githubusercontent.com/justinhchae/medium/main/bools.csv'

df = pd.read_csv(gitcsv)

# some columns that are supposed to be bool
cols = ['flag1', 'flag2', 'flag3']

# use np.where to find and match, then replace

justinhchae

import pandas as pd

gitcsv = 'https://raw.githubusercontent.com/justinhchae/medium/main/bools.csv'

df = pd.read_csv(gitcsv)

print(df.head()
"""
results of printing dataframe head
  category  flag1 flag2  flag3

justinhchae

import pandas as pd
import plotly.graph_objects as go
import plotly_express as px

gitcsv = 'https://raw.githubusercontent.com/justinhchae/medium/main/sample.csv'
df = pd.read_csv(gitcsv, index_col=0)

df['dates'] = pd.to_datetime(df['dates'])

freq='M' # or D or Y