How to obtain inputs and outputs of a bitcoin block more quickly, using JSON RPC or similar
Currently, it takes me almost 1.5 hours to obtain all inputs and outputs of a bitcoin block from January using JSON RPC. At this pace, it will take a significant amount of time to go all the way back to genesis. I am starting to think that I am using the wrong API call, or doing something else wrong. I would be greatly thankful if any of you could review the calls that I am currently using (pasted below) and perhaps tell me what I am doing wrong, or if bitcoind is the appropriate client for this.
from bitcoinrpc.authproxy import AuthServiceProxy, JSONRPCException
import json
import requests
import pprint
import time
import os
objectivestart_time=time.time()
pp= pprint.PrettyPrinter(indent=4)
rpc_user='n'
rpc_password='n'
url=("http://%s:%s@1"%(rpc_user, rpc_password))
def req(payload):
return requests.post(url,data=json.dumps(payload)).json()
def BlockHash_JsonRPCpayload(number):
dic={
"jsonrpc":"2.0",
"method": "getblockhash",
"params": [number] ,
"id": 1
}
return (dic)
def Block_JsonRPCpayload(h):
dic={
"jsonrpc":"2.0",
"method": "getblock",
"params": [str(h)] ,
"id": 1
}
return (dic)
def rawtx_JsonRPCpayload(tx):
dic={
"jsonrpc":"2.0",
"method": "getrawtransaction",
"params": [tx, True] ,
"id": 1
}
return (dic)
def rawvin_JsonRPCpayload(tx,vin):
block_payload={
"jsonrpc":"2.0",
"method": "getrawtransaction",
"params": [tx,1] ,
"id": 1
}
request=req(block_payload)
value=request['result']['vout'][int(vin)]['value']
return float(value)
BTC_dic={}
hash_payload=[]
start_time=time.time()
block_start=481101
block_end=490000
start=block_start
end= block_start
while end < block_end:
block_tries=0
while True:
try:
end=start+100
hash_payload=[]
print ( 'start', 'end', start, end )
for i in range(start,end):
hash_payload.append(BlockHash_JsonRPCpayload(i))
start_time=time.time()
print ( 'mark 1 ')
hash_response=requests.post(url, data=json.dumps(hash_payload)).json()
print('time to get hash batch',start, "_", end, time.time()-start_time)
print ( 'success in getting hashes for ranges', start, "_ ", end )
time.sleep(0.4)
block_payload=[]
for i in hash_response:
h=i['result']
block_payload.append(Block_JsonRPCpayload(h))
start_time=time.time()
block_response=requests.post(url,data=json.dumps(block_payload)).json()
print ( 'time to get block batch ',start, "_", end, time.time()-start_time)
print ( 'success in getting blocks for ranges', start, "_ ", end )
time.sleep(0.4)
for i in range(len(block_response)):
block=block_response[i]['result']
height=block['height']
print('getting', len(block['tx']),'transactions for block ', height)
BTC_dic[height]=block
tx_payload=[]
for tx in block['tx']:
tx_payload.append(rawtx_JsonRPCpayload(tx))
try_errors=0
while True:
try:
#print ( len(tx_payload), 'number of transactions in block ', height)
start_time=time.time()
transaction=requests.post(url,data=json.dumps(tx_payload)).json()
time.sleep(0.5)
#print ('time to get tx', time.time()-start_time)
print ( 'success in getting transactions for block ', height )
except:
try_errors+=1
print ( '!! Try error !!')
print ( try_errors, ' transaction tries for block height', height)
time.sleep(3)
continue
break
for i in range(len(transaction)):
tx_result=transaction[i]['result']
vin=tx_result['vin']
vout=tx_result['vout']
try:
fees=0
for j in range(len(vout)):
fees -= tx_result['vout'][j]['value']
for j in range(len(vin)):
vout=vin[j]['vout']
txid=vin[j]['txid']
vin_error=0
while True:
try:
value=rawvin_JsonRPCpayload(txid,vout)
except:
print ( '!! vin error !!')
print ( 'attempt', vin_error )
vin_error+=1
time.sleep(0.4)
continue
break
fees+=value
tx_result['vin'][j]['value']=value
tx_result['fees']=fees
except:
tx_result['vin'][0]['value']='No Inputs (Newly Generated Coins)'
tx_result['fees']=0
BTC_dic[height]['tx'][i]=tx_result
if len(BTC_dic) >= 100:
print ( 'len BTC_dic =', len(BTC_dic))
for block in BTC_dic:
block_fee=0
transactions= BTC_dic[block]['tx']
#pp.pprint(transactions)
for j in transactions:
fee=j['fees']
block_fee+=fee
BTC_dic[block]['block fees']=block_fee
block_array=[]
for block in BTC_dic:
block_array.append(block)
file_name='jcampbell_btc_txdata'+str(min(block_array))+'_'+str(max(block_array))+'.json'
print ('saving', file_name, 'time', time.time()-objectivestart_time)
with open(file_name, 'w') as outfile:
json.dump(BTC_dic,outfile)
BTC_dic={}
start=end
except:
block_tries+=1
print(block_tries, 'tries to get data for block range ', start,"_",end)
time.sleep(5)
continue
break
for block in BTC_dic:
block_fee=0
transactions= BTC_dic[block]['tx']
#pp.pprint(transactions)
for j in transactions:
fee=j['fees']
block_fee+=fee
BTC_dic[block]['block fees']=block_fee
block_array=[]
for block in BTC_dic:
block_array.append(block)
file_name='jcampbell_btc_txdata'+str(min(block_array))+'_'+str(max(block_array))+'.json'
print ('saving', file_name, 'time', time.time()-objectivestart_time)
with open(file_name, 'w') as outfile:
json.dump(BTC_dic,outfile)
Background
If we want BTC to go mainstream, we need a better way of showing Economic Activity. I need you to help me prove that Bitcoin is more than a speculative instrument.
I come from academia (Economics, specifically), and I often find it difficult to convincingly describe to my peers the growth of the Bitcoin ecosystem in numbers that translate to economic value.
Given the rapid growth this market has experienced, when people think about Bitcoin, price is often the first thing the comes to mind. For those of us that for some time have been researching Bitcoin, six thousand US dollars per bitcoin is, irrefutably, an incredible indicator of popularity. However, we need a better metric to show actual economic activity, which has also increased substantially in the past couple of years. If Bitcoin is a peer-to-peer electronic cash system, we need a better indicator of BTC volume between network peers. I have started doing research on Bitcoin Days Destroyed, but I often find this metric difficult to be explained as it does not accurately answers the question: how much value is being exchanged within the network?
I understand that a completely accurate answer to this question is nearly impossible due to Bitcoin's UTXO model, but we need more accurate estimations. Blockchain.info provides estimated transaction volume on a per day basis, but random samples I have processed indicate that their methodology significantly underestimates volume. After months of testing random samples and manually looking at addresses, I believe I have a developed a more accurate methodology to estimate on-chain volume. I plan to share this methodology with the community once I can test my hypothesis using fuller blocks, but the challenge now is to obtain all inputs and outputs from more recent blocks.
I believe a more realistic volume estimation backed by a peer reviewed methodology would increase interest from academia and add credibility to the Bitcoin network. If you are not familiar with JSON RPC, I would appreciate if you could please upvote this post so that more people see it.
Thank you.
-JC
http://ift.tt/2ygKFFt
Comments
Post a Comment