getTokenDistributionETHMonthly.py

#import libaries
import pandas as pd
import numpy as np
from web3 import Web3, HTTPProvider
import json 
from enum import Enum
from multiprocessing import Process
import numpy as np
from os.path import exists

#url of Ethereum node
urlETH = ""
#connect to Ethereum client 
w3 = Web3(HTTPProvider(urlETH))

tokens =pd.read_csv('data/general/govInfoETH.csv',sep='\t')[-8:]
tokens =tokens[tokens['delegation']=='R']
tokens =tokens[tokens['name']=='Gas']

class Decimals(Enum): 
    ETH = 18

def getDataFromNodeHolder(holders,w3,directory,token,address,blocks): 
    abi = json.load(open('abi/COMP.abi'))
    contract = w3.eth.contract(address=w3.to_checksum_address(address), abi=abi)
    tokenDistribution=[]    
    decimals = contract.functions.decimals().call()
    for holder in holders:
        holder0 = "0x"+ holder[3:]
        for block in blocks: 
            try:
                data = contract.functions.balanceOf(w3.to_checksum_address(holder0)).call({}, block)
                tokenDistribution.append({'holder':holder0,'block':block,'balance':data})
            except:
                pass
    tokenDistributionDF = pd.DataFrame.from_records(tokenDistribution).reindex()
    tokenDistributionDF['balance'] = tokenDistributionDF['balance'].apply(lambda x: x/(10**decimals))
    tokenDistributionDF=tokenDistributionDF[tokenDistributionDF['balance']>0]
    tokenDistributionDF.to_pickle(directory+token.lower()+'.pkl')

directory ="data/tokenDistMontly/" 
processes = []
    
for idx,row in tokens.iterrows():
    blocksMonthly =pd.read_pickle("data/time/blocksMonthly.p")
    blocks = blocksMonthly.number.tolist()[1:]
    tokenDistribution={}       
    holderstemp= pd.read_csv("data/holders/"+row['token'].lower()+".csv",delimiter ="\t",names=['holder'])
    holders = holderstemp['holder'].values.tolist()
    p = Process(target=getDataFromNodeHolder, args=(holders,w3,directory,row['token'],row['address'],blocks), )
    processes.append(p)
    # start process
    p.start()
    # join processes
for process in processes:
    process.join()
    processes.remove(process)