1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454
|
#!/usr/bin/env python3
# -*- coding: utf-8 -*-
# Copyright (C) 2009-2020 Authors of CryptoMiniSat, see AUTHORS file
#
# This program is free software; you can redistribute it and/or
# modify it under the terms of the GNU General Public License
# as published by the Free Software Foundation; version 2
# of the License.
#
# This program is distributed in the hope that it will be useful,
# but WITHOUT ANY WARRANTY; without even the implied warranty of
# MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
# GNU General Public License for more details.
#
# You should have received a copy of the GNU General Public License
# along with this program; if not, write to the Free Software
# Foundation, Inc., 51 Franklin Street, Fifth Floor, Boston, MA
# 02110-1301, USA.
from __future__ import print_function
import sqlite3
import optparse
import time
import pickle
import re
import pandas as pd
import numpy as np
import os.path
import sys
import helper
class QueryAddIdxes (helper.QueryHelper):
def __init__(self, dbfname):
super(QueryAddIdxes, self).__init__(dbfname)
def measure_size(self):
t = time.time()
ret = self.c.execute("select count() from reduceDB")
rows = self.c.fetchall()
rdb_rows = rows[0][0]
print("We have %d lines of RDB" % (rdb_rows))
t = time.time()
ret = self.c.execute("select count() from clause_stats")
rows = self.c.fetchall()
clss_rows = rows[0][0]
print("We have %d lines of clause_stats" % (clss_rows))
def create_indexes(self):
t = time.time()
print("Recreating indexes...")
queries = """
create index `idxclid33` on `sum_cl_use` (`clauseID`, `last_confl_used`);
---
create index `idxclid1` on `clause_stats` (`clauseID`, conflicts, latest_satzilla_feature_calc);
create index `idxclid1-2` on `clause_stats` (`clauseID`);
create index `idxclid2` on `clause_stats` (clauseID, conflicts, latest_satzilla_feature_calc);
create index `idxclid5` on `tags` ( `name`);
---
create index `idxclid6` on `reduceDB` (`clauseID`, conflicts);
create index `idxclid6-9` on `reduceDB` (`conflicts`);
create index `idxclid9` on `reduceDB_common` (`conflicts`, `latest_satzilla_feature_calc`);
create index `idxclid9-2` on `reduceDB_common` (`conflicts`, `latest_satzilla_feature_calc`);
create index `idxclid9-3` on `reduceDB_common` (`conflicts`);
create index `idxclid9-4` on `reduceDB_common` (`latest_satzilla_feature_calc`);
create index `idxclid6-2` on `reduceDB` (`clauseID`, `dump_no`);
create index `idxclid6-3` on `reduceDB` (`clauseID`, `conflicts`, `dump_no`);
create index `idxclid6-4` on `reduceDB` (`clauseID`, `conflicts`)
---
create index `idxclid7` on `satzilla_features` (`latest_satzilla_feature_calc`);
---
create index `idxclidUCLS-1` on `used_clauses` ( `clauseID`, `used_at`);
create index `idxclidUCLS-2` on `used_clauses` ( `used_at`);
---
create index `idxcl_last_in_solver-1` on `cl_last_in_solver` ( `clauseID`, `conflicts`);
---
create index `used_later_percentiles_idx3` on `used_later_percentiles` (`type_of_dat`, `percentile_descr`, `percentile`, `val`);
create index `used_later_percentiles_idx2` on `used_later_percentiles` (`type_of_dat`, `percentile_descr`, `val`);
"""
for l in queries.split('\n'):
t2 = time.time()
if options.verbose:
print("Creating index: ", l)
self.c.execute(l)
if options.verbose:
print("Index creation T: %-3.2f s" % (time.time() - t2))
print("indexes created T: %-3.2f s" % (time.time() - t))
class QueryCls (helper.QueryHelper):
def __init__(self, dbfname, tier, table):
super(QueryCls, self).__init__(dbfname)
self.fill_sql_query(tier, table=table)
def fill_sql_query(self, tier, table):
# sum_cl_use
self.sum_cl_use = helper.query_fragment(
"sum_cl_use", [], "sum_cl_use", options.verbose, self.c)
# RDB data
not_cols = [
"reduceDB_called"
, "clauseID"
, "in_xor"
, "locked"
, "conflicts"
, "activity_rel"]
self.rdb0_dat = helper.query_fragment(
"reduceDB", not_cols, "rdb0", options.verbose, self.c)
# reduceDB_common data
not_cols = [
"reduceDB_called"
, "simplifications"
, "restarts"
#, "conflicts"
, "latest_satzilla_feature_calc"
, "runtime"
]
self.rdb0_common_dat = helper.query_fragment(
"reduceDB_common", not_cols, "rdb0_common", options.verbose, self.c)
# clause data
not_cols = [
"simplifications"
, "restarts"
, "prev_restart"
, "antecedents_long_red_age_max"
, "antecedents_long_red_age_min"
, "latest_satzilla_feature_calc"
, "clauseID"]
self.clause_dat = helper.query_fragment(
"clause_stats", not_cols, "cl", options.verbose, self.c)
# satzilla data
not_cols = [
"simplifications"
, "restarts"
, "conflicts"
, "latest_satzilla_feature_calc"
, "irred_glue_distr_mean"
, "irred_glue_distr_var"]
self.satzfeat_dat = helper.query_fragment(
"satzilla_features", not_cols, "szfeat", options.verbose, self.c)
self.common_limits = """
order by random()
limit {limit}
"""
q_time_base="""
join {table}_{tier} on
{table}_{tier}.clauseID = rdb0.clauseID
and {table}_{tier}.rdb0conflicts = rdb0.conflicts
"""
q_columns_base="""
, {table}_{tier}.used_later as `x.{table}_{tier}`
, {table}_{tier}.percentile_fit as `x.{table}_{tier}_topperc`
"""
q_time = q_time_base.format(tier=tier, table=table)
q_columns = q_columns_base.format(tier=tier, table=table)
# final big query
self.q_select = """
SELECT
tags.val as `fname`
{clause_dat}
{rdb0_dat}
{rdb0_common_dat}
{sum_cl_use}
, (rdb0.conflicts - rdb0.introduced_at_conflict) as `cl.time_inside_solver`
, (sum_cl_use.last_confl_used - rdb0.introduced_at_conflict) as `x.a_lifetime`
{q_columns}
, sum_cl_use.num_used as `x.sum_cl_use`
FROM
reduceDB as rdb0
-- this is DELIBERATEY left-join: this way, clauses that as ternary
-- resolvents or otherwise generated during in-processing
-- can still be used
left join clause_stats as cl on
cl.clauseID = rdb0.clauseID
join reduceDB_common as rdb0_common on
rdb0_common.conflicts = rdb0.conflicts
join sum_cl_use on
sum_cl_use.clauseID = rdb0.clauseID
{q_time}
join cl_last_in_solver on
cl_last_in_solver.clauseID = rdb0.clauseID
, tags
WHERE
(cl.clauseID != 0 OR cl.clauseID is NULL)
and tags.name = "filename"
-- to avoid missing clauses and their missing data to affect results
and rdb0.conflicts + {del_at_least} <= cl_last_in_solver.conflicts
"""
self.myformat = {
"limit": 1000*1000*1000,
"clause_dat": self.clause_dat,
"satzfeat_dat_cur": self.satzfeat_dat.replace("szfeat.", "szfeat_cur."),
"rdb0_dat": self.rdb0_dat,
"sum_cl_use": self.sum_cl_use,
"rdb0_common_dat": self.rdb0_common_dat,
"q_time": q_time,
"q_columns": q_columns
}
def get_used_later_percentiles(self, name, table):
cur = self.conn.cursor()
q = """
select
`type_of_dat`,
`percentile_descr`,
`percentile`,
`val`
from {table}_percentiles
where `type_of_dat` = '{name}'
""".format(name=name, table=table)
cur.execute(q)
rows = cur.fetchall()
lookup = {}
for row in rows:
mystr = "%s_%s_perc" % (row[1], row[2])
print("table: {table} type: {t}, perc_desc: {perc_desc}, perc: {perc}, val: {val}".format(
table=table,
t=row[0],
perc_desc=row[1],
perc=row[2],
val=row[3]))
lookup[mystr] = row[3]
#print("perc lookup:", lookup)
return lookup
def get_one_data(self, tier, table):
perc = self.get_used_later_percentiles(tier, table)
self.myformat["del_at_least"] = getattr(options, tier)
# when del_at_least is over 2 million, then we need to make this smaller
# or we will delete all data
if self.myformat["del_at_least"] > 2*1000*1000:
self.myformat["del_at_least"] = 100*1000
# Make sure these stratas are equally represented
t = time.time()
dfs = self.run_stratified_queries(
limit=options.limit, perc=perc, tier=tier, table=table)
data = pd.concat(dfs)
print("** Queries finished. Total size: %s -- T: %-3.2f" % (
data.shape, time.time() - t))
return data
# perc == percentile distribution
# limit == MAX in total
# tier == forever/long/short
def run_stratified_queries(self, limit, perc, tier, table):
dfs = []
final_limit = limit
# NOTE: these are NON-ZERO percentages, but we replace 100 with "0", so the LAST chunk contains ALL, including 0, which is a large part of the data
for beg_perc, end_perc in [(0.0, options.cut1), (options.cut1, options.cut2), (options.cut2, 100.0)]:
beg = perc["top_non_zero_{perc}_perc".format(perc=beg_perc)]
if end_perc == 100.0:
end = 0.0
else:
end = perc["top_non_zero_{perc}_perc".format(perc=end_perc)]
what_to_strata = "`x.{table}_{tier}`".format(tier=tier, table=table)
print("Limit is {limit} value strata perc: ({a}, {b}) translates to value strata ({beg}, {end})".format(
limit=limit, a=beg_perc, b=end_perc, beg=beg, end=end))
# create one query for (beg,end) with different dump numbers
df, weighted_sizes = self.query_strata_per_dumpno(
str(self.q_select),
limit,
what_to_strata=what_to_strata,
strata=(beg,end))
dfs.append(df)
return dfs
def query_strata_per_dumpno(self, q, limit, what_to_strata, strata):
print("* Getting one set of data with limit %s" % limit)
weighted_size = []
df_parts = []
def one_part(mult, dump_no_filter=""):
self.myformat["limit"] = int(limit*mult)
df_parts.append(self.one_query(q + dump_no_filter, what_to_strata, strata))
print("--> Num rows for strata %s -- '%s': %s" % (strata, dump_no_filter, df_parts[-1].shape[0]))
ws = df_parts[-1].shape[0]/mult
print("--> The weight was %f so weighted size is: %d" % (mult, int(ws)))
weighted_size.append(ws)
one_part(1/4.0, dump_no_filter=" and rdb0.dump_no = 1 ")
one_part(1/4.0, dump_no_filter=" and rdb0.dump_no = 2 ")
one_part(1/4.0, dump_no_filter=" and rdb0.dump_no > 2 ")
one_part(1/4.0, dump_no_filter=" and rdb0.dump_no > 20 ")
df = pd.concat(df_parts)
print("-> size of all dump_no-s, strata {strata} data: {size}".format(
strata=strata, size=df.shape))
return df, weighted_size
def one_query(self, q, what_to_strata, strata):
q = q.format(**self.myformat)
if strata[1] == 0.0:
my_less_equal = ">="
else:
my_less_equal = ">"
q = """
select * from ( {q} )
where
{what_to_strata} <= {beg}
and {what_to_strata} {my_less_equal} {end}""".format(
q=q,
what_to_strata=what_to_strata,
beg=strata[0],
end=strata[1],
my_less_equal=my_less_equal)
q += self.common_limits
q = q.format(**self.myformat)
t = time.time()
sys.stdout.write("-> Running query for {} stratas {}...".format(what_to_strata, strata))
sys.stdout.flush()
if options.dump_sql:
print("query:", q)
df = pd.read_sql_query(q, self.conn)
print("T: %-3.1f" % (time.time() - t))
return df
def dump_dataframe(df, name):
if options.dump_csv:
fname = "%s.csv" % name
print("Dumping CSV data to:", fname)
df.to_csv(fname, index=False, columns=sorted(list(df)))
fname = "%s.dat" % name
print("Dumping pandas data to:", fname)
with open(fname, "wb") as f:
pickle.dump(df, f)
def one_database(dbfname):
with QueryAddIdxes(dbfname) as q:
q.measure_size()
helper.drop_idxs(q.c)
q.create_indexes()
with helper.QueryFill(dbfname) as q:
q.delete_and_create_used_laters()
for tier in ["short", "long", "forever"]:
for table in ["used_later", "used_later_anc"]:
q.fill_used_later_X(tier, duration=getattr(options, tier), table=table)
q.fill_used_later_X_perc_fit(tier, table=table)
print("Using sqlite3 DB file %s" % dbfname)
for tier in ["short", "long", "forever"]:
for table in ["used_later", "used_later_anc"]:
print("------> Doing tier {tier} table {table}".format(
tier=tier,table=table))
with QueryCls(dbfname, tier, table) as q:
df = q.get_one_data(tier, table)
if options.verbose:
print("Describing----")
dat = df.describe()
print(dat)
print("Describe done.---")
print("Features: ", df.columns.values.flatten().tolist())
if options.verbose:
print("Describing post-transform ----")
print(df.describe())
print("Describe done.---")
print("Features: ", df.columns.values.flatten().tolist())
cleanname = re.sub(r'\.cnf.gz.sqlite$', '', dbfname)
cleanname = re.sub(r'\.db$', '', dbfname)
cleanname = re.sub(r'\.sqlitedb$', '', dbfname)
cleanname = "{cleanname}-cldata-{table}-{tier}-cut1-{cut1}-cut2-{cut2}-limit-{limit}".format(
cleanname=cleanname,
cut1=options.cut1,
cut2=options.cut2,
limit=options.limit,
tier=tier,
table=table)
dump_dataframe(df, cleanname)
if __name__ == "__main__":
usage = "usage: %prog [options] file1.sqlite [file2.sqlite ...]"
parser = optparse.OptionParser(usage=usage)
# verbosity level
parser.add_option("--verbose", "-v", action="store_true", default=False,
dest="verbose", help="Print more output")
parser.add_option("--sql", action="store_true", default=False,
dest="dump_sql", help="Dump SQL queries")
parser.add_option("--csv", action="store_true", default=False,
dest="dump_csv", help="Dump CSV (for weka)")
# limits
parser.add_option("--limit", default=20000, type=int,
dest="limit", help="Max number of samples to take from each strata (for each table/tier)")
parser.add_option("--cut1", default=5.0, type=float,
dest="cut1", help="Where to cut the distrib. Default: %default")
parser.add_option("--cut2", default=30.0, type=float,
dest="cut2", help="Where to cut the distrib. Default: %default")
# debugging is faster with this
parser.add_option("--noind", action="store_true", default=False,
dest="no_recreate_indexes",
help="Don't recreate indexes")
# lengths of short/long
parser.add_option("--short", default=10000, type=int,
dest="short", help="Short duration. Default: %default")
parser.add_option("--long", default=30*1000, type=int,
dest="long", help="Long duration. Default: %default")
parser.add_option("--forever", default=120*1000, type=int,
dest="forever", help="Long duration. Default: %default")
(options, args) = parser.parse_args()
if len(args) != 1:
print("ERROR: You must give exactly one file")
exit(-1)
np.random.seed(2097483)
one_database(args[0])
|