bitcoin/test/functional/feature_fee_estimation.py

369 lines
15 KiB
Python
Raw Normal View History

2016-03-19 20:58:06 +01:00
#!/usr/bin/env python3
# Copyright (c) 2014-2021 The Bitcoin Core developers
# Distributed under the MIT software license, see the accompanying
# file COPYING or http://www.opensource.org/licenses/mit-license.php.
"""Test fee estimation code."""
from decimal import Decimal
import os
import random
estimatefee / estimatepriority RPC methods New RPC methods: return an estimate of the fee (or priority) a transaction needs to be likely to confirm in a given number of blocks. Mike Hearn created the first version of this method for estimating fees. It works as follows: For transactions that took 1 to N (I picked N=25) blocks to confirm, keep N buckets with at most 100 entries in each recording the fees-per-kilobyte paid by those transactions. (separate buckets are kept for transactions that confirmed because they are high-priority) The buckets are filled as blocks are found, and are saved/restored in a new fee_estiamtes.dat file in the data directory. A few variations on Mike's initial scheme: To estimate the fee needed for a transaction to confirm in X buckets, all of the samples in all of the buckets are used and a median of all of the data is used to make the estimate. For example, imagine 25 buckets each containing the full 100 entries. Those 2,500 samples are sorted, and the estimate of the fee needed to confirm in the very next block is the 50'th-highest-fee-entry in that sorted list; the estimate of the fee needed to confirm in the next two blocks is the 150'th-highest-fee-entry, etc. That algorithm has the nice property that estimates of how much fee you need to pay to get confirmed in block N will always be greater than or equal to the estimate for block N+1. It would clearly be wrong to say "pay 11 uBTC and you'll get confirmed in 3 blocks, but pay 12 uBTC and it will take LONGER". A single block will not contribute more than 10 entries to any one bucket, so a single miner and a large block cannot overwhelm the estimates.
2014-03-17 13:19:54 +01:00
from test_framework.messages import (
COIN,
COutPoint,
CTransaction,
CTxIn,
CTxOut,
)
from test_framework.script import (
CScript,
OP_1,
OP_DROP,
OP_TRUE,
)
from test_framework.script_util import (
script_to_p2sh_script,
)
from test_framework.test_framework import BitcoinTestFramework
from test_framework.util import (
assert_equal,
assert_greater_than,
assert_greater_than_or_equal,
assert_raises_rpc_error,
satoshi_round,
)
estimatefee / estimatepriority RPC methods New RPC methods: return an estimate of the fee (or priority) a transaction needs to be likely to confirm in a given number of blocks. Mike Hearn created the first version of this method for estimating fees. It works as follows: For transactions that took 1 to N (I picked N=25) blocks to confirm, keep N buckets with at most 100 entries in each recording the fees-per-kilobyte paid by those transactions. (separate buckets are kept for transactions that confirmed because they are high-priority) The buckets are filled as blocks are found, and are saved/restored in a new fee_estiamtes.dat file in the data directory. A few variations on Mike's initial scheme: To estimate the fee needed for a transaction to confirm in X buckets, all of the samples in all of the buckets are used and a median of all of the data is used to make the estimate. For example, imagine 25 buckets each containing the full 100 entries. Those 2,500 samples are sorted, and the estimate of the fee needed to confirm in the very next block is the 50'th-highest-fee-entry in that sorted list; the estimate of the fee needed to confirm in the next two blocks is the 150'th-highest-fee-entry, etc. That algorithm has the nice property that estimates of how much fee you need to pay to get confirmed in block N will always be greater than or equal to the estimate for block N+1. It would clearly be wrong to say "pay 11 uBTC and you'll get confirmed in 3 blocks, but pay 12 uBTC and it will take LONGER". A single block will not contribute more than 10 entries to any one bucket, so a single miner and a large block cannot overwhelm the estimates.
2014-03-17 13:19:54 +01:00
# Construct 2 trivial P2SH's and the ScriptSigs that spend them
# So we can create many transactions without needing to spend
# time signing.
SCRIPT = CScript([OP_1, OP_DROP])
P2SH = script_to_p2sh_script(SCRIPT)
REDEEM_SCRIPT = CScript([OP_TRUE, SCRIPT])
def small_txpuzzle_randfee(
from_node, conflist, unconflist, amount, min_fee, fee_increment
):
"""Create and send a transaction with a random fee.
2016-05-02 22:23:21 +02:00
The transaction pays to a trivial P2SH script, and assumes that its inputs
are of the same form.
The function takes a list of confirmed outputs and unconfirmed outputs
and attempts to use the confirmed list first for its inputs.
It adds the newly created outputs to the unconfirmed list.
Returns (raw transaction, fee)."""
# It's best to exponentially distribute our random fees
# because the buckets are exponentially spaced.
# Exponentially distributed from 1-128 * fee_increment
rand_fee = float(fee_increment) * (1.1892 ** random.randint(0, 28))
# Total fee ranges from min_fee to min_fee + 127*fee_increment
fee = min_fee - fee_increment + satoshi_round(rand_fee)
tx = CTransaction()
total_in = Decimal("0.00000000")
while total_in <= (amount + fee) and len(conflist) > 0:
t = conflist.pop(0)
total_in += t["amount"]
tx.vin.append(CTxIn(COutPoint(int(t["txid"], 16), t["vout"]), b""))
while total_in <= (amount + fee) and len(unconflist) > 0:
t = unconflist.pop(0)
total_in += t["amount"]
tx.vin.append(CTxIn(COutPoint(int(t["txid"], 16), t["vout"]), b""))
if total_in <= amount + fee:
raise RuntimeError(f"Insufficient funds: need {amount + fee}, have {total_in}")
tx.vout.append(CTxOut(int((total_in - amount - fee) * COIN), P2SH))
tx.vout.append(CTxOut(int(amount * COIN), P2SH))
# These transactions don't need to be signed, but we still have to insert
# the ScriptSig that will satisfy the ScriptPubKey.
for inp in tx.vin:
inp.scriptSig = REDEEM_SCRIPT
txid = from_node.sendrawtransaction(hexstring=tx.serialize().hex(), maxfeerate=0)
unconflist.append({"txid": txid, "vout": 0, "amount": total_in - amount - fee})
unconflist.append({"txid": txid, "vout": 1, "amount": amount})
return (tx.serialize().hex(), fee)
def check_raw_estimates(node, fees_seen):
"""Call estimaterawfee and verify that the estimates meet certain invariants."""
delta = 1.0e-6 # account for rounding error
for i in range(1, 26):
for _, e in node.estimaterawfee(i).items():
feerate = float(e["feerate"])
assert_greater_than(feerate, 0)
if feerate + delta < min(fees_seen) or feerate - delta > max(fees_seen):
raise AssertionError(
f"Estimated fee ({feerate}) out of range ({min(fees_seen)},{max(fees_seen)})"
)
def check_smart_estimates(node, fees_seen):
"""Call estimatesmartfee and verify that the estimates meet certain invariants."""
delta = 1.0e-6 # account for rounding error
last_feerate = float(max(fees_seen))
all_smart_estimates = [node.estimatesmartfee(i) for i in range(1, 26)]
mempoolMinFee = node.getmempoolinfo()["mempoolminfee"]
minRelaytxFee = node.getmempoolinfo()["minrelaytxfee"]
for i, e in enumerate(all_smart_estimates): # estimate is for i+1
feerate = float(e["feerate"])
assert_greater_than(feerate, 0)
assert_greater_than_or_equal(feerate, float(mempoolMinFee))
assert_greater_than_or_equal(feerate, float(minRelaytxFee))
if feerate + delta < min(fees_seen) or feerate - delta > max(fees_seen):
raise AssertionError(
f"Estimated fee ({feerate}) out of range ({min(fees_seen)},{max(fees_seen)})"
)
if feerate - delta > last_feerate:
raise AssertionError(
f"Estimated fee ({feerate}) larger than last fee ({last_feerate}) for lower number of confirms"
)
last_feerate = feerate
if i == 0:
assert_equal(e["blocks"], 2)
else:
assert_greater_than_or_equal(i + 1, e["blocks"])
def check_estimates(node, fees_seen):
check_raw_estimates(node, fees_seen)
check_smart_estimates(node, fees_seen)
def send_tx(node, utxo, feerate):
"""Broadcast a 1in-1out transaction with a specific input and feerate (sat/vb)."""
overhead, op, scriptsig, nseq, value, spk = 10, 36, 5, 4, 8, 24
tx_size = overhead + op + scriptsig + nseq + value + spk
fee = tx_size * feerate
tx = CTransaction()
tx.vin = [CTxIn(COutPoint(int(utxo["txid"], 16), utxo["vout"]), REDEEM_SCRIPT)]
tx.vout = [CTxOut(int(utxo["amount"] * COIN) - fee, P2SH)]
txid = node.sendrawtransaction(tx.serialize().hex())
return txid
class EstimateFeeTest(BitcoinTestFramework):
def set_test_params(self):
self.num_nodes = 3
# mine non-standard txs (e.g. txs with "dust" outputs)
# Force fSendTrickle to true (via whitelist.noban)
self.extra_args = [
["-acceptnonstdtxn", "-whitelist=noban@127.0.0.1"],
["-acceptnonstdtxn", "-whitelist=noban@127.0.0.1", "-blockmaxweight=68000"],
["-acceptnonstdtxn", "-whitelist=noban@127.0.0.1", "-blockmaxweight=32000"],
]
def skip_test_if_missing_module(self):
self.skip_if_no_wallet()
def setup_network(self):
"""
We'll setup the network to have 3 nodes that all mine with different parameters.
But first we need to use one node to create a lot of outputs
which we will use to generate our transactions.
"""
self.add_nodes(3, extra_args=self.extra_args)
# Use node0 to mine blocks for input splitting
# Node1 mines small blocks but that are bigger than the expected transaction rate.
# NOTE: the CreateNewBlock code starts counting block weight at 4,000 weight,
# (68k weight is room enough for 120 or so transactions)
# Node2 is a stingy miner, that
# produces too small blocks (room for only 55 or so transactions)
self.start_nodes()
self.import_deterministic_coinbase_privkeys()
self.stop_nodes()
def transact_and_mine(self, numblocks, mining_node):
min_fee = Decimal("0.00001")
# We will now mine numblocks blocks generating on average 100 transactions between each block
# We shuffle our confirmed txout set before each set of transactions
# small_txpuzzle_randfee will use the transactions that have inputs already in the chain when possible
# resorting to tx's that depend on the mempool when those run out
for _ in range(numblocks):
random.shuffle(self.confutxo)
for _ in range(random.randrange(100 - 50, 100 + 50)):
from_index = random.randint(1, 2)
(txhex, fee) = small_txpuzzle_randfee(
self.nodes[from_index],
self.confutxo,
self.memutxo,
Decimal("0.005"),
min_fee,
min_fee,
)
tx_kbytes = (len(txhex) // 2) / 1000.0
self.fees_per_kb.append(float(fee) / tx_kbytes)
self.sync_mempools(wait=0.1)
mined = mining_node.getblock(self.generate(mining_node, 1)[0], True)["tx"]
2016-05-02 22:23:21 +02:00
# update which txouts are confirmed
newmem = []
for utx in self.memutxo:
if utx["txid"] in mined:
self.confutxo.append(utx)
else:
newmem.append(utx)
self.memutxo = newmem
def initial_split(self, node):
"""Split two coinbase UTxOs into many small coins"""
utxo_count = 2048
self.confutxo = []
splitted_amount = Decimal("0.04")
fee = Decimal("0.1")
change = Decimal("100") - splitted_amount * utxo_count - fee
tx = CTransaction()
tx.vin = [
CTxIn(COutPoint(int(cb["txid"], 16), cb["vout"]), b"")
for cb in node.listunspent()[:2]
]
tx.vout = [CTxOut(int(splitted_amount * COIN), P2SH) for _ in range(utxo_count)]
tx.vout.append(CTxOut(int(change * COIN), P2SH))
txhex = node.signrawtransactionwithwallet(tx.serialize().hex())["hex"]
txid = node.sendrawtransaction(txhex)
self.confutxo = [
{"txid": txid, "vout": i, "amount": splitted_amount}
for i in range(utxo_count)
]
while len(node.getrawmempool()) > 0:
self.generate(node, 1, sync_fun=self.no_op)
def sanity_check_estimates_range(self):
"""Populate estimation buckets, assert estimates are in a sane range and
are strictly increasing as the target decreases."""
self.fees_per_kb = []
self.memutxo = []
2017-03-08 00:46:17 +01:00
self.log.info("Will output estimates for 1/2/3/6/15/25 blocks")
for _ in range(2):
self.log.info(
"Creating transactions and mining them with a block size that can't keep up"
)
# Create transactions and mine 10 small blocks with node 2, but create txs faster than we can mine
self.transact_and_mine(10, self.nodes[2])
check_estimates(self.nodes[1], self.fees_per_kb)
self.log.info(
"Creating transactions and mining them at a block size that is just big enough"
)
# Generate transactions while mining 10 more blocks, this time with node1
# which mines blocks with capacity just above the rate that transactions are being created
self.transact_and_mine(10, self.nodes[1])
check_estimates(self.nodes[1], self.fees_per_kb)
# Finish by mining a normal-sized block:
while len(self.nodes[1].getrawmempool()) > 0:
self.generate(self.nodes[1], 1)
2017-03-08 00:46:17 +01:00
self.log.info("Final estimates after emptying mempools")
check_estimates(self.nodes[1], self.fees_per_kb)
estimatefee / estimatepriority RPC methods New RPC methods: return an estimate of the fee (or priority) a transaction needs to be likely to confirm in a given number of blocks. Mike Hearn created the first version of this method for estimating fees. It works as follows: For transactions that took 1 to N (I picked N=25) blocks to confirm, keep N buckets with at most 100 entries in each recording the fees-per-kilobyte paid by those transactions. (separate buckets are kept for transactions that confirmed because they are high-priority) The buckets are filled as blocks are found, and are saved/restored in a new fee_estiamtes.dat file in the data directory. A few variations on Mike's initial scheme: To estimate the fee needed for a transaction to confirm in X buckets, all of the samples in all of the buckets are used and a median of all of the data is used to make the estimate. For example, imagine 25 buckets each containing the full 100 entries. Those 2,500 samples are sorted, and the estimate of the fee needed to confirm in the very next block is the 50'th-highest-fee-entry in that sorted list; the estimate of the fee needed to confirm in the next two blocks is the 150'th-highest-fee-entry, etc. That algorithm has the nice property that estimates of how much fee you need to pay to get confirmed in block N will always be greater than or equal to the estimate for block N+1. It would clearly be wrong to say "pay 11 uBTC and you'll get confirmed in 3 blocks, but pay 12 uBTC and it will take LONGER". A single block will not contribute more than 10 entries to any one bucket, so a single miner and a large block cannot overwhelm the estimates.
2014-03-17 13:19:54 +01:00
def test_feerate_mempoolminfee(self):
high_val = 3 * self.nodes[1].estimatesmartfee(1)["feerate"]
self.restart_node(1, extra_args=[f"-minrelaytxfee={high_val}"])
check_estimates(self.nodes[1], self.fees_per_kb)
self.restart_node(1)
def sanity_check_rbf_estimates(self, utxos):
"""During 5 blocks, broadcast low fee transactions. Only 10% of them get
confirmed and the remaining ones get RBF'd with a high fee transaction at
the next block.
The block policy estimator should return the high feerate.
"""
# The broadcaster and block producer
node = self.nodes[0]
miner = self.nodes[1]
# In sat/vb
low_feerate = 1
high_feerate = 10
# Cache the utxos of which to replace the spender after it failed to get
# confirmed
utxos_to_respend = []
txids_to_replace = []
assert len(utxos) >= 250
for _ in range(5):
# Broadcast 45 low fee transactions that will need to be RBF'd
for _ in range(45):
u = utxos.pop(0)
txid = send_tx(node, u, low_feerate)
utxos_to_respend.append(u)
txids_to_replace.append(txid)
# Broadcast 5 low fee transaction which don't need to
for _ in range(5):
send_tx(node, utxos.pop(0), low_feerate)
# Mine the transactions on another node
self.sync_mempools(wait=0.1, nodes=[node, miner])
for txid in txids_to_replace:
miner.prioritisetransaction(txid=txid, fee_delta=-COIN)
self.generate(miner, 1)
# RBF the low-fee transactions
while True:
try:
u = utxos_to_respend.pop(0)
send_tx(node, u, high_feerate)
except IndexError:
break
# Mine the last replacement txs
self.sync_mempools(wait=0.1, nodes=[node, miner])
self.generate(miner, 1)
# Only 10% of the transactions were really confirmed with a low feerate,
# the rest needed to be RBF'd. We must return the 90% conf rate feerate.
high_feerate_kvb = Decimal(high_feerate) / COIN * 10 ** 3
est_feerate = node.estimatesmartfee(2)["feerate"]
assert est_feerate == high_feerate_kvb
def run_test(self):
self.log.info("This test is time consuming, please be patient")
self.log.info("Splitting inputs so we can generate tx's")
# Split two coinbases into many small utxos
self.start_node(0)
self.initial_split(self.nodes[0])
self.log.info("Finished splitting")
# Now we can connect the other nodes, didn't want to connect them earlier
# so the estimates would not be affected by the splitting transactions
self.start_node(1)
self.start_node(2)
self.connect_nodes(1, 0)
self.connect_nodes(0, 2)
self.connect_nodes(2, 1)
self.sync_all()
self.log.info("Testing estimates with single transactions.")
self.sanity_check_estimates_range()
# check that the effective feerate is greater than or equal to the mempoolminfee even for high mempoolminfee
self.log.info(
"Test fee rate estimation after restarting node with high MempoolMinFee"
)
self.test_feerate_mempoolminfee()
self.log.info("Restarting node with fresh estimation")
self.stop_node(0)
fee_dat = os.path.join(self.nodes[0].datadir, self.chain, "fee_estimates.dat")
os.remove(fee_dat)
self.start_node(0)
self.connect_nodes(0, 1)
self.connect_nodes(0, 2)
self.log.info("Testing estimates with RBF.")
self.sanity_check_rbf_estimates(self.confutxo + self.memutxo)
self.log.info("Testing that fee estimation is disabled in blocksonly.")
self.restart_node(0, ["-blocksonly"])
assert_raises_rpc_error(
-32603, "Fee estimation disabled", self.nodes[0].estimatesmartfee, 2
)
if __name__ == "__main__":
EstimateFeeTest().main()