This commit is contained in:
Brendon J. Brewer 2021-08-16 09:52:40 +12:00 committed by Jack Robison
parent 234c03db09
commit 0c0e36b6f8
No known key found for this signature in database
GPG key ID: DF25C68FE0239BB2
6 changed files with 316 additions and 878 deletions

View file

@ -30,6 +30,7 @@ from lbry.wallet.server.db.claimtrie import get_takeover_name_ops, StagedActivat
from lbry.wallet.server.db.claimtrie import get_remove_name_ops, get_remove_effective_amount_ops
from lbry.wallet.server.db.prefixes import ACTIVATED_SUPPORT_TXO_TYPE, ACTIVATED_CLAIM_TXO_TYPE
from lbry.wallet.server.db.prefixes import PendingActivationKey, PendingActivationValue, Prefixes, ClaimToTXOValue
from lbry.wallet.server.db.trending import TrendingDB
from lbry.wallet.server.udp import StatusServer
from lbry.wallet.server.db.revertable import RevertableOp, RevertablePut, RevertableDelete, RevertableOpStack
if typing.TYPE_CHECKING:
@ -263,6 +264,8 @@ class BlockProcessor:
self.claim_channels: Dict[bytes, bytes] = {}
self.hashXs_by_tx: DefaultDict[bytes, List[int]] = defaultdict(list)
self.trending_db = TrendingDB(env.db_dir)
async def claim_producer(self):
if self.db.db_height <= 1:
return
@ -310,6 +313,7 @@ class BlockProcessor:
start = time.perf_counter()
await self.run_in_thread(self.advance_block, block)
await self.flush()
self.trending_db.process_block(self.height, self.daemon.cached_height())
self.logger.info("advanced to %i in %0.3fs", self.height, time.perf_counter() - start)
if self.height == self.coin.nExtendedClaimExpirationForkHeight:
self.logger.warning(
@ -514,6 +518,9 @@ class BlockProcessor:
self.txo_to_claim[(tx_num, nout)] = pending
self.claim_hash_to_txo[claim_hash] = (tx_num, nout)
self.db_op_stack.extend_ops(pending.get_add_claim_utxo_ops())
self.trending_db.add_event({"claim_hash": claim_hash,
"event": "upsert",
"lbc": 1E-8*txo.amount})
def _add_support(self, txo: 'Output', tx_num: int, nout: int):
supported_claim_hash = txo.claim_hash[::-1]
@ -523,6 +530,9 @@ class BlockProcessor:
self.db_op_stack.extend_ops(StagedClaimtrieSupport(
supported_claim_hash, tx_num, nout, txo.amount
).get_add_support_utxo_ops())
self.trending_db.add_event({"claim_hash": supported_claim_hash,
"event": "support",
"lbc": 1E-8*txo.amount})
def _add_claim_or_support(self, height: int, tx_hash: bytes, tx_num: int, nout: int, txo: 'Output',
spent_claims: typing.Dict[bytes, Tuple[int, int, str]]):
@ -542,6 +552,10 @@ class BlockProcessor:
self.db_op_stack.extend_ops(StagedClaimtrieSupport(
spent_support, txin_num, txin.prev_idx, support_amount
).get_spend_support_txo_ops())
self.trending_db.add_event({"claim_hash": spent_support,
"event": "support",
"lbc": -1E-8*support_amount})
spent_support, support_amount = self.db.get_supported_claim_from_txo(txin_num, txin.prev_idx)
if spent_support:
supported_name = self._get_pending_claim_name(spent_support)
@ -619,6 +633,9 @@ class BlockProcessor:
if normalized_name.startswith('@'): # abandon a channel, invalidate signatures
self._invalidate_channel_signatures(claim_hash)
self.trending_db.add_event({"claim_hash": claim_hash,
"event": "delete"})
def _invalidate_channel_signatures(self, claim_hash: bytes):
for k, signed_claim_hash in self.db.db.iterator(
prefix=Prefixes.channel_to_claim.pack_partial_key(claim_hash)):

View file

@ -0,0 +1,299 @@
import math
import os
import sqlite3
import time
HALF_LIFE = 400
RENORM_INTERVAL = 1000
WHALE_THRESHOLD = 10000.0
def whale_decay_factor(lbc):
"""
An additional decay factor applied to whale claims.
"""
if lbc <= WHALE_THRESHOLD:
return 1.0
adjusted_half_life = HALF_LIFE/(math.log10(lbc/WHALE_THRESHOLD) + 1.0)
return 2.0**(1.0/HALF_LIFE - 1.0/adjusted_half_life)
def soften(lbc):
mag = abs(lbc) + 1E-8
sign = 1.0 if lbc >= 0.0 else -1.0
return sign*mag**0.25
def delay(lbc: int):
if lbc <= 0:
return 0
elif lbc < 1000000:
return int(lbc**0.5)
else:
return 1000
def inflate_units(height):
blocks = height % RENORM_INTERVAL
return 2.0 ** (blocks/HALF_LIFE)
PRAGMAS = ["PRAGMA FOREIGN_KEYS = OFF;",
"PRAGMA JOURNAL_MODE = WAL;",
"PRAGMA SYNCHRONOUS = 0;"]
class TrendingDB:
def __init__(self, data_dir):
"""
Opens the trending database in the directory data_dir.
For testing, pass data_dir=":memory:"
"""
if data_dir == ":memory:":
path = ":memory:"
else:
path = os.path.join(data_dir, "trending.db")
self.db = sqlite3.connect(path, check_same_thread=False)
for pragma in PRAGMAS:
self.execute(pragma)
self.execute("BEGIN;")
self._create_tables()
self._create_indices()
self.execute("COMMIT;")
self.pending_events = []
def execute(self, *args, **kwargs):
return self.db.execute(*args, **kwargs)
def add_event(self, event):
self.pending_events.append(event)
# print(f"Added event: {event}.", flush=True)
def _create_tables(self):
self.execute("""CREATE TABLE IF NOT EXISTS claims
(claim_hash BYTES NOT NULL PRIMARY KEY,
bid_lbc REAL NOT NULL,
support_lbc REAL NOT NULL,
trending_score REAL NOT NULL,
needs_write BOOLEAN NOT NULL)
WITHOUT ROWID;""")
self.execute("""CREATE TABLE IF NOT EXISTS spikes
(claim_hash BYTES NOT NULL REFERENCES claims (claim_hash),
activation_height INTEGER NOT NULL,
mass REAL NOT NULL);""")
def _create_indices(self):
self.execute("CREATE INDEX IF NOT EXISTS idx1 ON spikes\
(activation_height, claim_hash, mass);")
self.execute("CREATE INDEX IF NOT EXISTS idx2 ON spikes\
(claim_hash);")
self.execute("CREATE INDEX IF NOT EXISTS idx3 ON claims (trending_score);")
self.execute("CREATE INDEX IF NOT EXISTS idx4 ON claims (needs_write, claim_hash);")
self.execute("CREATE INDEX IF NOT EXISTS idx5 ON claims (bid_lbc + support_lbc);")
def get_trending_score(self, claim_hash):
result = self.execute("SELECT trending_score FROM claims\
WHERE claim_hash = ?;", (claim_hash, ))\
.fetchall()
if len(result) == 0:
return 0.0
else:
return result[0]
def _upsert_claim(self, height, event):
claim_hash = event["claim_hash"]
# Get old total lbc value of claim
old_lbc_pair = self.execute("SELECT bid_lbc, support_lbc FROM claims\
WHERE claim_hash = ?;",
(claim_hash, )).fetchone()
if old_lbc_pair is None:
old_lbc_pair = (0.0, 0.0)
if event["event"] == "upsert":
new_lbc_pair = (event["lbc"], old_lbc_pair[1])
elif event["event"] == "support":
new_lbc_pair = (old_lbc_pair[0], old_lbc_pair[1] + event["lbc"])
# Upsert the claim
self.execute("INSERT INTO claims VALUES (?, ?, ?, ?, 1)\
ON CONFLICT (claim_hash) DO UPDATE\
SET bid_lbc = excluded.bid_lbc,\
support_lbc = excluded.support_lbc;",
(claim_hash, new_lbc_pair[0], new_lbc_pair[1], 0.0))
if self.active:
old_lbc, lbc = sum(old_lbc_pair), sum(new_lbc_pair)
# Add the spike
softened_change = soften(lbc - old_lbc)
change_in_softened = soften(lbc) - soften(old_lbc)
spike_mass = (softened_change**0.25*change_in_softened**0.75).real
activation_height = height + delay(lbc)
if spike_mass != 0.0:
self.execute("INSERT INTO spikes VALUES (?, ?, ?);",
(claim_hash, activation_height, spike_mass))
def _delete_claim(self, claim_hash):
self.execute("DELETE FROM spikes WHERE claim_hash = ?;", (claim_hash, ))
self.execute("DELETE FROM claims WHERE claim_hash = ?;", (claim_hash, ))
def _apply_spikes(self, height):
spikes = self.execute("SELECT claim_hash, mass FROM spikes\
WHERE activation_height = ?;",
(height, )).fetchall()
for claim_hash, mass in spikes: # TODO: executemany for efficiency
self.execute("UPDATE claims SET trending_score = trending_score + ?,\
needs_write = 1\
WHERE claim_hash = ?;",
(mass, claim_hash))
self.execute("DELETE FROM spikes WHERE activation_height = ?;",
(height, ))
def _decay_whales(self):
whales = self.execute("SELECT claim_hash, bid_lbc + support_lbc FROM claims\
WHERE bid_lbc + support_lbc >= ?;", (WHALE_THRESHOLD, ))\
.fetchall()
for claim_hash, lbc in whales:
factor = whale_decay_factor(lbc)
self.execute("UPDATE claims SET trending_score = trending_score*?, needs_write = 1\
WHERE claim_hash = ?;", (factor, claim_hash))
def _renorm(self):
factor = 2.0**(-RENORM_INTERVAL/HALF_LIFE)
# Zero small values
self.execute("UPDATE claims SET trending_score = 0.0, needs_write = 1\
WHERE trending_score <> 0.0 AND ABS(?*trending_score) < 1E-6;",
(factor, ))
# Normalise other values
self.execute("UPDATE claims SET trending_score = ?*trending_score, needs_write = 1\
WHERE trending_score <> 0.0;", (factor, ))
def process_block(self, height, daemon_height):
self.active = daemon_height - height <= 10*HALF_LIFE
self.execute("BEGIN;")
if self.active:
# Check for a unit change
if height % RENORM_INTERVAL == 0:
self._renorm()
# Apply extra whale decay
self._decay_whales()
# Upsert claims
for event in self.pending_events:
if event["event"] == "upsert":
self._upsert_claim(height, event)
# Process supports
for event in self.pending_events:
if event["event"] == "support":
self._upsert_claim(height, event)
# Delete claims
for event in self.pending_events:
if event["event"] == "delete":
self._delete_claim(event["claim_hash"])
if self.active:
# Apply spikes
self._apply_spikes(height)
# Get set of claims that need writing to ES
claims_to_write = set()
for row in self.db.execute("SELECT claim_hash FROM claims WHERE\
needs_write = 1;"):
claims_to_write.add(row[0])
self.db.execute("UPDATE claims SET needs_write = 0\
WHERE needs_write = 1;")
self.execute("COMMIT;")
self.pending_events.clear()
return claims_to_write
if __name__ == "__main__":
import matplotlib.pyplot as plt
import numpy as np
import numpy.random as rng
import os
trending_db = TrendingDB(":memory:")
heights = list(range(1, 1000))
heights = heights + heights[::-1] + heights
events = [{"height": 45,
"what": dict(claim_hash="a", event="upsert", lbc=1.0)},
{"height": 100,
"what": dict(claim_hash="a", event="support", lbc=3.0)},
{"height": 150,
"what": dict(claim_hash="a", event="support", lbc=-3.0)},
{"height": 170,
"what": dict(claim_hash="a", event="upsert", lbc=100000.0)},
{"height": 730,
"what": dict(claim_hash="a", event="delete")}]
inverse_events = [{"height": 730,
"what": dict(claim_hash="a", event="upsert", lbc=100000.0)},
{"height": 170,
"what": dict(claim_hash="a", event="upsert", lbc=1.0)},
{"height": 150,
"what": dict(claim_hash="a", event="support", lbc=3.0)},
{"height": 100,
"what": dict(claim_hash="a", event="support", lbc=-3.0)},
{"height": 45,
"what": dict(claim_hash="a", event="delete")}]
xs, ys = [], []
last_height = 0
for height in heights:
# Prepare the changes
if height > last_height:
es = events
else:
es = inverse_events
for event in es:
if event["height"] == height:
trending_db.add_event(event["what"])
# Process the block
trending_db.process_block(height, height)
if height > last_height: # Only plot when moving forward
xs.append(height)
y = trending_db.execute("SELECT trending_score FROM claims;").fetchone()
y = 0.0 if y is None else y[0]
ys.append(y/inflate_units(height))
last_height = height
xs = np.array(xs)
ys = np.array(ys)
plt.figure(1)
plt.plot(xs, ys, "o-", alpha=0.2)
plt.figure(2)
plt.plot(xs)
plt.show()

View file

@ -1,9 +0,0 @@
from . import zscore
from . import ar
from . import variable_decay
TRENDING_ALGORITHMS = {
'zscore': zscore,
'ar': ar,
'variable_decay': variable_decay
}

View file

@ -1,265 +0,0 @@
import copy
import math
import time
# Half life in blocks
HALF_LIFE = 134
# Decay coefficient per block
DECAY = 0.5**(1.0/HALF_LIFE)
# How frequently to write trending values to the db
SAVE_INTERVAL = 10
# Renormalisation interval
RENORM_INTERVAL = 1000
# Assertion
assert RENORM_INTERVAL % SAVE_INTERVAL == 0
# Decay coefficient per renormalisation interval
DECAY_PER_RENORM = DECAY**(RENORM_INTERVAL)
# Log trending calculations?
TRENDING_LOG = True
def install(connection):
"""
Install the AR trending algorithm.
"""
check_trending_values(connection)
if TRENDING_LOG:
f = open("trending_ar.log", "a")
f.close()
# Stub
CREATE_TREND_TABLE = ""
def check_trending_values(connection):
"""
If the trending values appear to be based on the zscore algorithm,
reset them. This will allow resyncing from a standard snapshot.
"""
c = connection.cursor()
needs_reset = False
for row in c.execute("SELECT COUNT(*) num FROM claim WHERE trending_global <> 0;"):
if row[0] != 0:
needs_reset = True
break
if needs_reset:
print("Resetting some columns. This might take a while...", flush=True, end="")
c.execute(""" BEGIN;
UPDATE claim SET trending_group = 0;
UPDATE claim SET trending_mixed = 0;
UPDATE claim SET trending_global = 0;
UPDATE claim SET trending_local = 0;
COMMIT;""")
print("done.")
def spike_height(trending_score, x, x_old, time_boost=1.0):
"""
Compute the size of a trending spike.
"""
# Change in softened amount
change_in_softened_amount = x**0.25 - x_old**0.25
# Softened change in amount
delta = x - x_old
softened_change_in_amount = abs(delta)**0.25
# Softened change in amount counts more for minnows
if delta > 0.0:
if trending_score >= 0.0:
multiplier = 0.1/((trending_score/time_boost + softened_change_in_amount) + 1.0)
softened_change_in_amount *= multiplier
else:
softened_change_in_amount *= -1.0
return time_boost*(softened_change_in_amount + change_in_softened_amount)
def get_time_boost(height):
"""
Return the time boost at a given height.
"""
return 1.0/DECAY**(height % RENORM_INTERVAL)
def trending_log(s):
"""
Log a string.
"""
if TRENDING_LOG:
fout = open("trending_ar.log", "a")
fout.write(s)
fout.flush()
fout.close()
class TrendingData:
"""
An object of this class holds trending data
"""
def __init__(self):
self.claims = {}
# Have all claims been read from db yet?
self.initialised = False
def insert_claim_from_load(self, claim_hash, trending_score, total_amount):
assert not self.initialised
self.claims[claim_hash] = {"trending_score": trending_score,
"total_amount": total_amount,
"changed": False}
def update_claim(self, claim_hash, total_amount, time_boost=1.0):
"""
Update trending data for a claim, given its new total amount.
"""
assert self.initialised
# Extract existing total amount and trending score
# or use starting values if the claim is new
if claim_hash in self.claims:
old_state = copy.deepcopy(self.claims[claim_hash])
else:
old_state = {"trending_score": 0.0,
"total_amount": 0.0,
"changed": False}
# Calculate LBC change
change = total_amount - old_state["total_amount"]
# Modify data if there was an LBC change
if change != 0.0:
spike = spike_height(old_state["trending_score"],
total_amount,
old_state["total_amount"],
time_boost)
trending_score = old_state["trending_score"] + spike
self.claims[claim_hash] = {"total_amount": total_amount,
"trending_score": trending_score,
"changed": True}
def test_trending():
"""
Quick trending test for something receiving 10 LBC per block
"""
data = TrendingData()
data.insert_claim_from_load("abc", 10.0, 1.0)
data.initialised = True
for height in range(1, 5000):
if height % RENORM_INTERVAL == 0:
data.claims["abc"]["trending_score"] *= DECAY_PER_RENORM
time_boost = get_time_boost(height)
data.update_claim("abc", data.claims["abc"]["total_amount"] + 10.0,
time_boost=time_boost)
print(str(height) + " " + str(time_boost) + " " \
+ str(data.claims["abc"]["trending_score"]))
# One global instance
# pylint: disable=C0103
trending_data = TrendingData()
def run(db, height, final_height, recalculate_claim_hashes):
if height < final_height - 5*HALF_LIFE:
trending_log("Skipping AR trending at block {h}.\n".format(h=height))
return
start = time.time()
trending_log("Calculating AR trending at block {h}.\n".format(h=height))
trending_log(" Length of trending data = {l}.\n"\
.format(l=len(trending_data.claims)))
# Renormalise trending scores and mark all as having changed
if height % RENORM_INTERVAL == 0:
trending_log(" Renormalising trending scores...")
keys = trending_data.claims.keys()
for key in keys:
if trending_data.claims[key]["trending_score"] != 0.0:
trending_data.claims[key]["trending_score"] *= DECAY_PER_RENORM
trending_data.claims[key]["changed"] = True
# Tiny becomes zero
if abs(trending_data.claims[key]["trending_score"]) < 1E-9:
trending_data.claims[key]["trending_score"] = 0.0
trending_log("done.\n")
# Regular message.
trending_log(" Reading total_amounts from db and updating"\
+ " trending scores in RAM...")
# Get the value of the time boost
time_boost = get_time_boost(height)
# Update claims from db
if not trending_data.initialised:
# On fresh launch
for row in db.execute("""
SELECT claim_hash, trending_mixed,
(amount + support_amount)
AS total_amount
FROM claim;
"""):
trending_data.insert_claim_from_load(row[0], row[1], 1E-8*row[2])
trending_data.initialised = True
else:
for row in db.execute(f"""
SELECT claim_hash,
(amount + support_amount)
AS total_amount
FROM claim
WHERE claim_hash IN
({','.join('?' for _ in recalculate_claim_hashes)});
""", list(recalculate_claim_hashes)):
trending_data.update_claim(row[0], 1E-8*row[1], time_boost)
trending_log("done.\n")
# Write trending scores to DB
if height % SAVE_INTERVAL == 0:
trending_log(" Writing trending scores to db...")
the_list = []
keys = trending_data.claims.keys()
for key in keys:
if trending_data.claims[key]["changed"]:
the_list.append((trending_data.claims[key]["trending_score"],
key))
trending_data.claims[key]["changed"] = False
trending_log("{n} scores to write...".format(n=len(the_list)))
db.executemany("UPDATE claim SET trending_mixed=? WHERE claim_hash=?;",
the_list)
trending_log("done.\n")
trending_log("Trending operations took {time} seconds.\n\n"\
.format(time=time.time() - start))
if __name__ == "__main__":
test_trending()

View file

@ -1,485 +0,0 @@
"""
AR-like trending with a delayed effect and a faster
decay rate for high valued claims.
"""
import math
import time
import sqlite3
# Half life in blocks *for lower LBC claims* (it's shorter for whale claims)
HALF_LIFE = 200
# Whale threshold, in LBC (higher -> less DB writing)
WHALE_THRESHOLD = 10000.0
# Decay coefficient per block
DECAY = 0.5**(1.0/HALF_LIFE)
# How frequently to write trending values to the db
SAVE_INTERVAL = 10
# Renormalisation interval
RENORM_INTERVAL = 1000
# Assertion
assert RENORM_INTERVAL % SAVE_INTERVAL == 0
# Decay coefficient per renormalisation interval
DECAY_PER_RENORM = DECAY**(RENORM_INTERVAL)
# Log trending calculations?
TRENDING_LOG = True
def install(connection):
"""
Install the trending algorithm.
"""
check_trending_values(connection)
trending_data.initialise(connection.cursor())
if TRENDING_LOG:
f = open("trending_variable_decay.log", "a")
f.close()
# Stub
CREATE_TREND_TABLE = ""
def check_trending_values(connection):
"""
If the trending values appear to be based on the zscore algorithm,
reset them. This will allow resyncing from a standard snapshot.
"""
c = connection.cursor()
needs_reset = False
for row in c.execute("SELECT COUNT(*) num FROM claim WHERE trending_global <> 0;"):
if row[0] != 0:
needs_reset = True
break
if needs_reset:
print("Resetting some columns. This might take a while...", flush=True,
end="")
c.execute(""" BEGIN;
UPDATE claim SET trending_group = 0;
UPDATE claim SET trending_mixed = 0;
COMMIT;""")
print("done.")
def trending_log(s):
"""
Log a string to the log file
"""
if TRENDING_LOG:
fout = open("trending_variable_decay.log", "a")
fout.write(s)
fout.flush()
fout.close()
def trending_unit(height):
"""
Return the trending score unit at a given height.
"""
# Round to the beginning of a SAVE_INTERVAL batch of blocks.
_height = height - (height % SAVE_INTERVAL)
return 1.0/DECAY**(height % RENORM_INTERVAL)
class TrendingDB:
"""
An in-memory database of trending scores
"""
def __init__(self):
self.conn = sqlite3.connect(":memory:", check_same_thread=False)
self.cursor = self.conn.cursor()
self.initialised = False
self.write_needed = set()
def execute(self, query, *args, **kwargs):
return self.conn.execute(query, *args, **kwargs)
def executemany(self, query, *args, **kwargs):
return self.conn.executemany(query, *args, **kwargs)
def begin(self):
self.execute("BEGIN;")
def commit(self):
self.execute("COMMIT;")
def initialise(self, db):
"""
Pass in claims.db
"""
if self.initialised:
return
trending_log("Initialising trending database...")
# The need for speed
self.execute("PRAGMA JOURNAL_MODE=OFF;")
self.execute("PRAGMA SYNCHRONOUS=0;")
self.begin()
# Create the tables
self.execute("""
CREATE TABLE IF NOT EXISTS claims
(claim_hash BYTES PRIMARY KEY,
lbc REAL NOT NULL DEFAULT 0.0,
trending_score REAL NOT NULL DEFAULT 0.0)
WITHOUT ROWID;""")
self.execute("""
CREATE TABLE IF NOT EXISTS spikes
(id INTEGER PRIMARY KEY,
claim_hash BYTES NOT NULL,
height INTEGER NOT NULL,
mass REAL NOT NULL,
FOREIGN KEY (claim_hash)
REFERENCES claims (claim_hash));""")
# Clear out any existing data
self.execute("DELETE FROM claims;")
self.execute("DELETE FROM spikes;")
# Create indexes
self.execute("CREATE INDEX idx1 ON spikes (claim_hash, height, mass);")
self.execute("CREATE INDEX idx2 ON spikes (claim_hash, height, mass DESC);")
self.execute("CREATE INDEX idx3 on claims (lbc DESC, claim_hash, trending_score);")
# Import data from claims.db
for row in db.execute("""
SELECT claim_hash,
1E-8*(amount + support_amount) AS lbc,
trending_mixed
FROM claim;
"""):
self.execute("INSERT INTO claims VALUES (?, ?, ?);", row)
self.commit()
self.initialised = True
trending_log("done.\n")
def apply_spikes(self, height):
"""
Apply spikes that are due. This occurs inside a transaction.
"""
spikes = []
unit = trending_unit(height)
for row in self.execute("""
SELECT SUM(mass), claim_hash FROM spikes
WHERE height = ?
GROUP BY claim_hash;
""", (height, )):
spikes.append((row[0]*unit, row[1]))
self.write_needed.add(row[1])
self.executemany("""
UPDATE claims
SET trending_score = (trending_score + ?)
WHERE claim_hash = ?;
""", spikes)
self.execute("DELETE FROM spikes WHERE height = ?;", (height, ))
def decay_whales(self, height):
"""
Occurs inside transaction.
"""
if height % SAVE_INTERVAL != 0:
return
whales = self.execute("""
SELECT trending_score, lbc, claim_hash
FROM claims
WHERE lbc >= ?;
""", (WHALE_THRESHOLD, )).fetchall()
whales2 = []
for whale in whales:
trending, lbc, claim_hash = whale
# Overall multiplication factor for decay rate
# At WHALE_THRESHOLD, this is 1
# At 10*WHALE_THRESHOLD, it is 3
decay_rate_factor = 1.0 + 2.0*math.log10(lbc/WHALE_THRESHOLD)
# The -1 is because this is just the *extra* part being applied
factor = (DECAY**SAVE_INTERVAL)**(decay_rate_factor - 1.0)
# Decay
trending *= factor
whales2.append((trending, claim_hash))
self.write_needed.add(claim_hash)
self.executemany("UPDATE claims SET trending_score=? WHERE claim_hash=?;",
whales2)
def renorm(self, height):
"""
Renormalise trending scores. Occurs inside a transaction.
"""
if height % RENORM_INTERVAL == 0:
threshold = 1.0E-3/DECAY_PER_RENORM
for row in self.execute("""SELECT claim_hash FROM claims
WHERE ABS(trending_score) >= ?;""",
(threshold, )):
self.write_needed.add(row[0])
self.execute("""UPDATE claims SET trending_score = ?*trending_score
WHERE ABS(trending_score) >= ?;""",
(DECAY_PER_RENORM, threshold))
def write_to_claims_db(self, db, height):
"""
Write changed trending scores to claims.db.
"""
if height % SAVE_INTERVAL != 0:
return
rows = self.execute(f"""
SELECT trending_score, claim_hash
FROM claims
WHERE claim_hash IN
({','.join('?' for _ in self.write_needed)});
""", list(self.write_needed)).fetchall()
db.executemany("""UPDATE claim SET trending_mixed = ?
WHERE claim_hash = ?;""", rows)
# Clear list of claims needing to be written to claims.db
self.write_needed = set()
def update(self, db, height, recalculate_claim_hashes):
"""
Update trending scores.
Input is a cursor to claims.db, the block height, and the list of
claims that changed.
"""
assert self.initialised
self.begin()
self.renorm(height)
# Fetch changed/new claims from claims.db
for row in db.execute(f"""
SELECT claim_hash,
1E-8*(amount + support_amount) AS lbc
FROM claim
WHERE claim_hash IN
({','.join('?' for _ in recalculate_claim_hashes)});
""", list(recalculate_claim_hashes)):
claim_hash, lbc = row
# Insert into trending db if it does not exist
self.execute("""
INSERT INTO claims (claim_hash)
VALUES (?)
ON CONFLICT (claim_hash) DO NOTHING;""",
(claim_hash, ))
# See if it was an LBC change
old = self.execute("SELECT * FROM claims WHERE claim_hash=?;",
(claim_hash, )).fetchone()
lbc_old = old[1]
# Save new LBC value into trending db
self.execute("UPDATE claims SET lbc = ? WHERE claim_hash = ?;",
(lbc, claim_hash))
if lbc > lbc_old:
# Schedule a future spike
delay = min(int((lbc + 1E-8)**0.4), HALF_LIFE)
spike = (claim_hash, height + delay, spike_mass(lbc, lbc_old))
self.execute("""INSERT INTO spikes
(claim_hash, height, mass)
VALUES (?, ?, ?);""", spike)
elif lbc < lbc_old:
# Subtract from future spikes
penalty = spike_mass(lbc_old, lbc)
spikes = self.execute("""
SELECT * FROM spikes
WHERE claim_hash = ?
ORDER BY height ASC, mass DESC;
""", (claim_hash, )).fetchall()
for spike in spikes:
spike_id, mass = spike[0], spike[3]
if mass > penalty:
# The entire penalty merely reduces this spike
self.execute("UPDATE spikes SET mass=? WHERE id=?;",
(mass - penalty, spike_id))
penalty = 0.0
else:
# Removing this spike entirely accounts for some (or
# all) of the penalty, then move on to other spikes
self.execute("DELETE FROM spikes WHERE id=?;",
(spike_id, ))
penalty -= mass
# If penalty remains, that's a negative spike to be applied
# immediately.
if penalty > 0.0:
self.execute("""
INSERT INTO spikes (claim_hash, height, mass)
VALUES (?, ?, ?);""",
(claim_hash, height, -penalty))
self.apply_spikes(height)
self.decay_whales(height)
self.commit()
self.write_to_claims_db(db, height)
# The "global" instance to work with
# pylint: disable=C0103
trending_data = TrendingDB()
def spike_mass(x, x_old):
"""
Compute the mass of a trending spike (normed - constant units).
x_old = old LBC value
x = new LBC value
"""
# Sign of trending spike
sign = 1.0
if x < x_old:
sign = -1.0
# Magnitude
mag = abs(x**0.25 - x_old**0.25)
# Minnow boost
mag *= 1.0 + 2E4/(x + 100.0)**2
return sign*mag
def run(db, height, final_height, recalculate_claim_hashes):
if height < final_height - 5*HALF_LIFE:
trending_log(f"Skipping trending calculations at block {height}.\n")
return
start = time.time()
trending_log(f"Calculating variable_decay trending at block {height}.\n")
trending_data.update(db, height, recalculate_claim_hashes)
end = time.time()
trending_log(f"Trending operations took {end - start} seconds.\n\n")
def test_trending():
"""
Quick trending test for claims with different support patterns.
Actually use the run() function.
"""
# Create a fake "claims.db" for testing
# pylint: disable=I1101
dbc = apsw.Connection(":memory:")
db = dbc.cursor()
# Create table
db.execute("""
BEGIN;
CREATE TABLE claim (claim_hash TEXT PRIMARY KEY,
amount REAL NOT NULL DEFAULT 0.0,
support_amount REAL NOT NULL DEFAULT 0.0,
trending_mixed REAL NOT NULL DEFAULT 0.0);
COMMIT;
""")
# Initialise trending data before anything happens with the claims
trending_data.initialise(db)
# Insert initial states of claims
everything = {"huge_whale": 0.01, "medium_whale": 0.01, "small_whale": 0.01,
"huge_whale_botted": 0.01, "minnow": 0.01}
def to_list_of_tuples(stuff):
l = []
for key in stuff:
l.append((key, stuff[key]))
return l
db.executemany("""
INSERT INTO claim (claim_hash, amount) VALUES (?, 1E8*?);
""", to_list_of_tuples(everything))
# Process block zero
height = 0
run(db, height, height, everything.keys())
# Save trajectories for plotting
trajectories = {}
for row in trending_data.execute("""
SELECT claim_hash, trending_score
FROM claims;
"""):
trajectories[row[0]] = [row[1]/trending_unit(height)]
# Main loop
for height in range(1, 1000):
# One-off supports
if height == 1:
everything["huge_whale"] += 5E5
everything["medium_whale"] += 5E4
everything["small_whale"] += 5E3
# Every block
if height < 500:
everything["huge_whale_botted"] += 5E5/500
everything["minnow"] += 1
# Remove supports
if height == 500:
for key in everything:
everything[key] = 0.01
# Whack into the db
db.executemany("""
UPDATE claim SET amount = 1E8*? WHERE claim_hash = ?;
""", [(y, x) for (x, y) in to_list_of_tuples(everything)])
# Call run()
run(db, height, height, everything.keys())
# Append current trending scores to trajectories
for row in db.execute("""
SELECT claim_hash, trending_mixed
FROM claim;
"""):
trajectories[row[0]].append(row[1]/trending_unit(height))
dbc.close()
# pylint: disable=C0415
import matplotlib.pyplot as plt
for key in trajectories:
plt.plot(trajectories[key], label=key)
plt.legend()
plt.show()
if __name__ == "__main__":
test_trending()

View file

@ -1,119 +0,0 @@
from math import sqrt
# TRENDING_WINDOW is the number of blocks in ~6hr period (21600 seconds / 161 seconds per block)
TRENDING_WINDOW = 134
# TRENDING_DATA_POINTS says how many samples to use for the trending algorithm
# i.e. only consider claims from the most recent (TRENDING_WINDOW * TRENDING_DATA_POINTS) blocks
TRENDING_DATA_POINTS = 28
CREATE_TREND_TABLE = """
create table if not exists trend (
claim_hash bytes not null,
height integer not null,
amount integer not null,
primary key (claim_hash, height)
) without rowid;
"""
class ZScore:
__slots__ = 'count', 'total', 'power', 'last'
def __init__(self):
self.count = 0
self.total = 0
self.power = 0
self.last = None
def step(self, value):
if self.last is not None:
self.count += 1
self.total += self.last
self.power += self.last ** 2
self.last = value
@property
def mean(self):
return self.total / self.count
@property
def standard_deviation(self):
value = (self.power / self.count) - self.mean ** 2
return sqrt(value) if value > 0 else 0
def finalize(self):
if self.count == 0:
return self.last
return (self.last - self.mean) / (self.standard_deviation or 1)
def install(connection):
connection.create_aggregate("zscore", 1, ZScore)
connection.executescript(CREATE_TREND_TABLE)
def run(db, height, final_height, affected_claims):
# don't start tracking until we're at the end of initial sync
if height < (final_height - (TRENDING_WINDOW * TRENDING_DATA_POINTS)):
return
if height % TRENDING_WINDOW != 0:
return
db.execute(f"""
DELETE FROM trend WHERE height < {height - (TRENDING_WINDOW * TRENDING_DATA_POINTS)}
""")
start = (height - TRENDING_WINDOW) + 1
db.execute(f"""
INSERT OR IGNORE INTO trend (claim_hash, height, amount)
SELECT claim_hash, {start}, COALESCE(
(SELECT SUM(amount) FROM support WHERE claim_hash=claim.claim_hash
AND height >= {start}), 0
) AS support_sum
FROM claim WHERE support_sum > 0
""")
zscore = ZScore()
for global_sum in db.execute("SELECT AVG(amount) AS avg_amount FROM trend GROUP BY height"):
zscore.step(global_sum.avg_amount)
global_mean, global_deviation = 0, 1
if zscore.count > 0:
global_mean = zscore.mean
global_deviation = zscore.standard_deviation
db.execute(f"""
UPDATE claim SET
trending_local = COALESCE((
SELECT zscore(amount) FROM trend
WHERE claim_hash=claim.claim_hash ORDER BY height DESC
), 0),
trending_global = COALESCE((
SELECT (amount - {global_mean}) / {global_deviation} FROM trend
WHERE claim_hash=claim.claim_hash AND height = {start}
), 0),
trending_group = 0,
trending_mixed = 0
""")
# trending_group and trending_mixed determine how trending will show in query results
# normally the SQL will be: "ORDER BY trending_group, trending_mixed"
# changing the trending_group will have significant impact on trending results
# changing the value used for trending_mixed will only impact trending within a trending_group
db.execute(f"""
UPDATE claim SET
trending_group = CASE
WHEN trending_local > 0 AND trending_global > 0 THEN 4
WHEN trending_local <= 0 AND trending_global > 0 THEN 3
WHEN trending_local > 0 AND trending_global <= 0 THEN 2
WHEN trending_local <= 0 AND trending_global <= 0 THEN 1
END,
trending_mixed = CASE
WHEN trending_local > 0 AND trending_global > 0 THEN trending_global
WHEN trending_local <= 0 AND trending_global > 0 THEN trending_local
WHEN trending_local > 0 AND trending_global <= 0 THEN trending_local
WHEN trending_local <= 0 AND trending_global <= 0 THEN trending_global
END
WHERE trending_local <> 0 OR trending_global <> 0
""")