Utils.py

#!/usr/bin/env python
# coding: utf-8

import pandas as pd
import numpy as np
import copy
import math

from functools import reduce
from decimal import Decimal
import itertools

from graphviz import Digraph


def graph(df, df_lag, eng = 'dot'):
    edge_style = ""
    g = Digraph(engine=eng)
    g.attr(ratio='fill', size='3,3')
   
    for k, row in enumerate(df.index):
        if any(df.iloc[k]) or any(df[row]):
            g.node(str(k),row, shape='oval', fontsize='10', width='0', style='filled', fillcolor='#c9c9c9', color="white") 

    for j, col in enumerate(df.columns):
        for i, row in enumerate(df[col]):
            if(row):
                te_val  = str(np.round(row, 6))
                g.edge(str(i), str(j), label=str(df_lag[df_lag.columns[j]][i]),style= edge_style, color='black')  
    return g 

def graph_simple(df, eng = 'dot'):
    edge_style = ""
    g = Digraph(engine=eng)
    in_graph = []
    for k, row in enumerate(df.index):
        if any(df.loc[row]):
            g.node(str(row),row, shape='oval', fontsize='10', width='0', style='filled', fillcolor='#c9c9c9', color="gray")
            in_graph.append(row)

              
    for c, col in enumerate(df.columns):
        if any(df[col]):
            if col not in in_graph:
                g.node(str(col), col, shape='oval', fontsize='10', width='0', style='filled', fillcolor='#c9c9c9', color="gray") 

    for j, col in enumerate(df.columns):
        for i, row in enumerate(df.index):
            if(df[col][i]):
                g.edge(str(row), str(col), label=str(df.at[row,col]), style= edge_style, color='black')  
    return g

def graph_from_dict(dictionary, eng = 'dot'):
    edge_style = ""
    g = Digraph(engine=eng)
   
    for k in dictionary.keys():
        if any([k in sub for sub in dictionary.values() for key in dictionary.keys()]) or dictionary[k]:
            g.node(str(k),k, shape='oval', fontsize='10', width='0', style='filled', fillcolor='#c9c9c9', color="gray") 

    for k, i in dictionary.items():
        for it in i:
            g.edge(str(it), str(k), label='',style= edge_style, color='black')  
    return g

def df_to_dictTree(df):
    dict_tree = {}
    for col in df.columns:
        non_zero = df[col].nonzero()
        dict_tree[col] = df[col].index[non_zero].values.tolist()
    return dict_tree


def app_roll_mean(df, window):
    roll  = df.copy().rolling(window).mean() 
    roll.dropna(inplace=True)
    roll = roll.round(decimals=0).copy()
    roll.reset_index(drop=True, inplace=True)
    return roll